Suspended sediment budget and intra-event sediment dynamics of a small glaciated mountainous catchment in the Northern Caucasus

Abstract

PurposeThe sediment dynamics of (peri-)glacial catchments can be highly variable and complex. Understanding these dynamics and their underlying causes is not only of interest from a scientific perspective but also required to address the practical problems with which they are often associated. In order to better understand the sediment dynamics of glaciated mountainous catchments, suspended sediment fluxes in the 9.1 km2 Djankuat catchment (North Caucasus, Russia) were monitored intensively during the 2017 ablation season.Materials and methodsThe intra-event suspended sediment dynamics were studied using a newly proposed simple hysteresis index (SHI), quantifying to what extent evolutions in sediment concentration are characterized by a clockwise or anticlockwise hysteresis loop.Results and discussionThe resulting catchment suspended sediment yield was 1033 t km−2 year−1, with the glacier itself contributing 72% of the suspended sediment load. However, during rainfall events, also hillslope erosion in the proglacial area became a very significant sediment source. Clockwise hysteresis loops occurred in 61.8% of the events, while anticlockwise in 11.8%. On the other hand, only 47.8% of the total suspended sediment flux was transported during clockwise events. Our observations clearly indicate that events showing a stronger clockwise pattern (i.e., a higher SHI) are associated with a larger sediment input from the proglacial area.Conclusions
Overall, our results provide data and insights on sediment dynamics in an understudied environment. They illustrate that the type and characteristics of sediment concentration hysteresis loops are to some extent linked to the dominant sediment sources during the event. As such, the proposed methodology and SHI may also help with a better understanding of sediment dynamics in other environments.

Full text

SEDIMENTS, SEC 3 •HILLSLOPE AND RIVER BASIN SEDIMENT DYNAMICS •RESEARCH
ARTICLE
Suspended sediment budget and intra-event sediment dynamics
of a small glaciated mountainous catchment
in the Northern Caucasus
Anatoly Tsyplenkov
1,2
&Matthias Vanmaercke
3
&Valentin Golosov
1,2,4
&Sergey Chalov
1
Received: 13 January 2020 /Accepted: 19 April 2020
#Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract
Purpose The sediment dynamics of (peri-)glacial catchments can be highly variable and complex. Understanding these dynamics
and their underlying causes is not only of interest from a scientific perspective but also required to address the practical problems
with which they are often associated. In order to better understand the sediment dynamics of glaciated mountainous catchments,
suspended sediment fluxes in the 9.1 km
2
Djankuat catchment (North Caucasus, Russia) were monitored intensively during the
2017 ablation season.
Materials and methods The intra-event suspended sediment dynamics were studied using a newly proposed simple hysteresis
index (SHI), quantifying to what extent evolutions in sediment concentration are characterized by a clockwise or anticlockwise
hysteresis loop.
Results and discussion The resulting catchment suspended sediment yield was 1033 t km
−2
year
−1
, with the glacier itself
contributing 72% of the suspended sediment load. However, during rainfall events, also hillslope erosion in the proglacial area
became a very significant sediment source. Clockwise hysteresis loops occurred in 61.8% of the events, while anticlockwise in
11.8%. On the other hand, only 47.8% of the total suspended sediment flux was transported during clockwise events. Our
observations clearly indicate that events showing a stronger clockwise pattern (i.e., a higher SHI) are associated with a larger
sediment input from the proglacial area.
Conclusions Overall, our results provide data and insights on sediment dynamics in an understudied environment. They illustrate
that the type and characteristics of sediment concentration hysteresis loops are to some extent linked to the dominant sediment
sources during the event. As such, the proposed methodology and SHI may also help with a better understanding of sediment
dynamics in other environments.
Keywords Sedimentyield .Hysteresis .Mountainhydrology .
Suspended sediments .Sediment budget
1 Introduction
Over the past decades, there has been an increasing interest in
the response of fluvial sediment system to climate change
(e.g., Walling 1995; Jones 1999; Farnsworth and Milliman
2003; Koppes and Montgomery 2009; Glazirin and
Semakova 2019). One of the key elements in this response
are glaciers and their influence on the catchment sediment
yield (SY) of mountainous headwaters (Beniston 2003; Barry
2006). Glaciers are generally a much more efficient erosion
agent than rivers (e.g., Koppes and Montgomery 2009). They
are, therefore, a dominant sediment source in many mountain
Responsible editor: Hugh Smith
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s11368-020-02633-z) contains supplementary
material, which is available to authorized users.
*Anatoly Tsyplenkov
atsyplenkov@gmail.com
1
Faculty of Geography, Lomonosov Moscow State University,
Leninskiye Gory, 1, Moscow, Russian Federation 119991
2
Institute of Geography, Russian Academy of Sciences, Staromonetny
st. 29, Moscow, Russian Federation 119017
3
Département de Géographie, Université de Liège, U.R. SPHERES,
Clos Mercator 3, 4000 Liège, Belgium
4
Kazan Federal University, Kremlevskaya st., 18, Kazan, Russian
Federation 420008
Journal of Soils and Sediments
https://doi.org/10.1007/s11368-020-02633-z

catchments (e.g., Gurnell et al. 1996;Halletetal.1996)and
can exert a strong control on the SY of even larger river sys-
tems (e.g., Syvitski and Milliman 2007;Hinderer2012).
While the SY values from glaciated catchments often rank
amongst the highest in the world (e.g., Dedkov and Moszherin
1984;Halletetal.1996; Slaymaker 2018), it is also known
that the sediment dynamics of (peri-)glacial catchments can be
highly variable and complex (Warburton 1990; Gurnell et al.
1996;Hodgkinsetal.2003; Singh et al. 2005;Manoetal.
2009;O’Farrell et al. 2009;Iidaetal.2012; Mao and Carrillo
2017). Understanding these dynamics and their underlying
causes is not only of interest from a scientific perspective
but also required to address the practical problems with which
they are often associated. These include damage to hydro-
electric power plants, lake and reservoir sedimentation, irriga-
tion canal aggradation, and impacts on water quality (e.g.
Bogen 1989; Moore et al. 2009).
This need for a better understanding of sediment dynamics
from glaciated catchments becomes even more relevant in the
light of climate change (e.g., Morche et al. 2019). Ongoing
glacier retreat will likely increase suspended sediment concen-
tration in many proglacial streams due to an increase in sub-
glacial sources and the paraglacial conditions resulting from
glacial retreat (Moore et al. 2009). As such, not only the gla-
ciers itself but also the (typically barren) hillslopes and mo-
raine deposits in the direct vicinity of the glacier can be an
important sediment source (Carrivick et al. 2013;Geilhausen
et al. 2013;Morcheetal.2019). Expected changes in precip-
itation regime and intensity (e.g., Stott and Mount 2007;
Polade et al. 2015) can be expected to affect the significance
of such sediment sources, leading to further complexity.
Nonetheless, only a limited number of case studies exist on
the sediment budget of glaciated catchments (e.g., Warburton
1990; Otto et al. 2009;O’Farrell et al. 2009; Leggat et al.
2015; Mao and Carrillo 2017). This is especially the case for
the Caucasus. Although a significant number of works have
studied the denudation and river sediment yield in this moun-
tain range (Khmeleva et al. 2000;Jaoshvili2002;Khmeleva
and Shevchenko 2006; Vinogradova et al. 2007;Mozzherin
and Sharifullin 2015), only very few provide sediment bud-
gets for glaciated basins (e.g., Durgerov et al. 1972).
Nevertheless, also in the Caucasus, most glaciers retreat
(Solomina 2000;Stokesetal.2007; Kutuzov and
Shahgedanova 2009; Tielidze et al. 2015; Kutuzov et al.
2019). This happens at an increasing rate: from −0.44% year
−1
areal change in 1960–1986 to −0.69% year
−1
in 1986–2014
(Tielidze and Wheate 2018). As discussed above, this can be
expected to elevate the total sediment yield of the region in the
future.
Moreover, very few studies have focused on sediment dy-
namics at the intra-event scale. Nonetheless, sediment dis-
charge events are typically characterized by suspended sedi-
ment concentration (SSC,[gm
−3
])–runoff discharge (Q,
[m
3
s
−1
]) hysteresis loops. Such hysteresis loops can have a
significant impact on the accuracy oftotal suspended sediment
loads and should therefore be accounted for (e.g., Morehead
et al. 2003; Vanmaercke et al. 2010). However, it is also ex-
pected that they can provide insight into the dominant sedi-
ment sources during the event (e.g., Morehead et al. 2003;
Lloyd et al. 2016).
Also, in the case of glaciated catchments, hysteresis effects
are frequently observed. Several researchers (e.g., Singh et al.
2005;Iidaetal.2012) suggest that clockwise hysteresis loops
dominate during the entire ablation period. On the other hand,
Mao and Carrillo (2017) report that, for the small Andean
catchment they studied, clockwise hysteresis loops mainly
occur during snowmelt events, while anticlockwise loops pre-
vail during glacier melting. Furthermore, Favaro and
Lamoureux (2015) indicate that hysteresis patterns can change
because of global warming. Their study in the High Arctic
(2004–2012) has shown that daily hysteresis patterns shifted
from dominantly clockwise to a more frequent occurrence of
anticlockwise loops. Nonetheless, it remains unclear to what
extent these findings can be generalized, given the currently
limited number of case studies available that systematically
analyze the temporal dynamics of sediment concentrations
across a sufficiently large number of events.
In addition, the study of hysteresis patterns in relation to
sediment sources is often difficult because it is not always
straightforward to characterize or quantify hysteresis loops.
This is especially so when the number of observations during
the event is limited (e.g., Vanmaercke et al. 2010). To mitigate
this challenge, several studies have prepared hysteresis index-
es (Langlois et al. 2005; Lawler et al. 2006; Lloyd et al. 2016).
These indexes typically come with limitations. For example,
they do not always clearly depend on the magnitude of the
event (Langlois et al. 2005; Lawler et al. 2006), which restricts
their capability to quantify the importance of different hyster-
esis types for sediment transport. Later studies have proposed
improvements (Gao and Josefson 2012). However, such im-
provements also come with the requirement of more measure-
ments (e.g., Lloyd et al. 2016; Hamshaw et al. 2018).
Nonetheless, the dominant hysteresis pattern can usually be
determined with a relatively limited number of samples (Aich
et al. 2014).
To help address these research gaps, this paper investi-
gates the suspended sediment dynamics of the Djankuat, a
glaciated mountainous catchment located in the North
Caucasus (Russia). Based on a large set of measurements,
we aim to quantify the total suspended sediment load as
well as the relative importance of different sediment
sources at both the seasonal and event scales. For the latter,
we propose a new method/index to quantify runoff
discharge–sediment concentration hysteresis loops that
can also be applied when only a small number of samples
are available.
J Soils Sediments

2 Study area
Our research was conducted at the Djankuat Glacialogical
Station (DGS), located at the Russian part of Northern
Caucasus, near the Russian-Georgian border (43° 12′31.71″
N, 42° 44′05.93″E, altitude 2635 m). The Djankuat catchment
was selected as a representative case study for the Northern
Caucasus during the International Hydrological Decade
(Dyurgerov 2003). Detailed observations on the glaciology,
meteorology, and hydrology of the catchment have been made
since 1965. It is therefore one of the best-studied glaciers in
Russia (Stokes et al. 2007;Lavrentievetal.2015;Vasil’chuk
et al. 2016;Retsetal.2017; Toropov et al. 2017;Rybakand
Rybak 2018). Nonetheless, our understanding of its sediment
dynamics has remained very limited. Continuous SSC measure-
ments have only been conducted since 2015 (Rets et al. 2019).
The Djankuat catchment is a typical high mountain catch-
ment with steep slopes, alpine meadows, and glacial-nival
terrains. Its catchment area is 9.1 km
2
.In2017,27%ofits
surface area was covered by glaciers (Rets et al. 2019;
Table 1). The largest glacier (2.42 km
2
), i.e., the Djankuat
glacier, is the source of the Djankuat river (cf. Fig. 1). In
addition, there are three smaller glaciers (< 0.5 km
2
)located
in the study basin: Koiavgan, Viatau, and Visyachii. The av-
erage precipitation depth (in rainfall equivalents) is around
4090 mm year
−1
. However, precipitation is characterized by
an important temporal variability with daily amounts up to
97.2 mm day
−1
. During the ablation period (May–
September) at the DGS air temperature ranges from −1.1 to
24.2 °C with an average of 10.2 °C. At the glacier (3000 m)
temperature during this period ranges from −9to17.6°Cwith
an average of 6.6 °C (Rets et al. 2019).
Typically, the catchment is almost completely covered with
snow until late May–earlyJune,whenthemeltingseasonbegins.
At the beginning of July, only a few patches of snow remain at
the higher elevations and north-facing slopes of the non-glaciated
area. Due to this specific hydrological regime, an estimated 98%
of the total sediment flux and water runoff occurs in the ablation
period (Durgerov et al. 1972;Retsetal.2017).
Mean water discharge at the outlet of the Djankuat catch-
ment is 1.38 m
3
s
−1
, with typical water discharges ranging
between 1 and 2 m
3
s
−1
(Rets et al. 2019). Peak discharges
can reach values up to 3–4m
3
s
−1
. Such high values are most-
ly observed in the second half of the ablation period and are
often linked to high intensity rainfall events (Rets et al. 2019).
Earlier work indicates that snow and ice-melting processes
are, with 44%, the main sources of runoff in the catchment;
37% comes from groundwater and only 19% of the discharge
is due to rainfall events (Rets et al. 2017).
3 Materials and methods
3.1 Layout of the stations and field data collection
All fieldwork was carried out during the ablation period of
2017. Runoff discharge (Q,m
3
s
−1
) and turbidity (T,NTU)
were monitored at the downstream gauging station of the
Djankuat (OUT, cf. Fig. 1) from June 6, 2017, to September
24, 2017. A digital pressure logger (Solinst Levellogger
Junior) recorded the pressure (water and atmospheric) every
10 min. In addition, water stages (H, cm) were measured man-
ually at 9:00, 11:00, 13:00, 15:00, 17:00, 19:00, 21:00, and
23:00 every day. These data allowed the conversion of the
automatic pressure readings to flow depths and to verify the
correct functioning of the pressure logger. Nearly 60 water
discharge (Q) measurements were made based on the dilution
method, using NaCl as a tracer (Dobriyal et al. 2017). Based
on these measurements, a robust stage-discharge (H–Q) rela-
tionship was developed, which allowed the conversion of ob-
served flow depths into runoff discharges (for details; see Rets
et al. 2019). To account for potential changes in bed morphol-
ogy, H–Q rating curves were established for every month of
the measuring period (Rets et al. 2019).
Turbidity measurements were made at least eight times per
day at the same times of the manual Hrecordings, using a
portable turbidimeter Hach 2100P.Toconvertopticalturbidity
values (T, NTU) to suspended sediment concentration (SSC,
gm
−3
), we established an empirical relationship between both
variables. During the fieldwork campaigns in the ablation sea-
sons of 2015–2017, we collected 39 water samples at the same
time and location as the Tmeasurements. These samples were
filtered with a Millipore vacuum pump, using 0.45 μm
Millipore membrane filter papers. These papers were then
dried for 2 h at 105 °C and weighted. Based on this weight
and the volume of the water sample, the SSC could be derived
and compared with the corresponding T. Based on this com-
parison, the following empirical relationships were
established (Fig. 2):
SSC ¼19 0:217Tþ0:00226T2;T<1000
556 þ1:21T;T1000
(ð1Þ
Table 1 Some key characteristics of the stream draining to the three
gauging stations (cf. Fig. 1)
OUT MID GL
Area (km
2
) 9.1 8.09 4.24
Minimum elevation (m) 2648 2682 2722
Maximum elevation (m) 3848 3848 3848
Elevation range (m) 1200 1166 1115
Glacierized area (%) 27 30.4 58.0
J Soils Sediments

Fig. 2 Suspended sediment concentration (SSC)–turbidity (T) relationships, based on 39 water samples collected between 2015 and 2017. A distinction
is made for samples corresponding with a T < 1000 NTU (a) and samples with a T above 1000 NTU (b)
Fig. 1 Overview of the Djankuat catchment and position of the gauging
stations and meteorological stations used in this study. OUT gauging
station at the catchment outlet, GL gauging station at the glacier mouth,
MID gauging station between OUT and GL (Source Satellite image:
Worldview 2, 2010-08-20)
J Soils Sediments

where SSC is the suspended sediment concentration (g m
−3
)
and Tis the optical turbidity (NTU).
At the MID and GL stations (cf. Fig. 1), measurements were
conducted from June 30, 2017, to August 30, 2017. At the MID
station, only the water turbidity (T
MID
, NTU) was measured. At
the GL station, both turbidity (T
GL
,NTU)andwaterstage(H
GL
,
cm) were measured. The T-observations at the GL and MID
station were converted to SSC values, using Eq. 1(Fig. 2).
Typically, these measurements were conducted only twice a
day (morning and evening). In general, they were conducted at
the same time (within a 5–10-min interval) of each other and the
measurements conducted at the OUT station.
Meteorological data for the measuring period are available in
the database compiled by Rets et al. (2019). In this research, we
used daily precipitation depths (in water equivalents), which
were manually measured for every precipitation event near the
outlet (Fig. 1). The rain gauge was located 0.5 m above ground.
Overall, suspended sediments at the outlet of the Djankuat
catchment (OUT) can originate from four potential sources: the
main glacier, the smaller glaciers (drained by the Koiavgan
stream; Fig. 1), riverbed erosion and hillslope erosion (during
snowmelt and rainfall events). Given its position, the only poten-
tial sediment source at the GL station is the main glacier. For the
MID station, the potential sediment sources are similar as for the
OUT station. However, given that measurements at MID (and
GL) only started in the middle of the ablation season (July), the
role of snowmelt as a potential runoff and sediment source was
already negligible (Vasil’chuk et al. 2016).
3.2 Runoff discharge estimations for the GL and MID
stations
While discharge was monitored continuously at the outlet of the
catchment (OUT), continuous runoff discharge series were lack-
ing for the GL and MID stations (cf. Fig. 1) and had to be
estimated. For the GL station, we based our estimate on the
assumption that, during days without rain, the dominant sediment
sources at the OUT station is typically the main glacier (i.e.,
Djankuat glacier). This is justified by the fact that snowmelt
was no longer significant when the measurements at the GL
and MID station started (see above) and the relatively much
smaller size of the small glaciers drained by the Koiavgan stream
(cf. Study area; Fig. 1). In addition, visual inspections (and irreg-
ular Tmeasurements) clearly indicated that water coming from
the Koiavgan stream typically contained no significant amounts
of sediments (≈10–20 NTU) during non-rainy days. We further
hypothesized that, during non-rainy days, there is no significant
deposition of suspended sediments between the GL and OUT
station. This is justified by the short distance (≈1 km) between
both stations, the steep channel slope (≈100 m km
−1
), the high
observed stream velocities (1–2ms
−1
), and the overall fine tex-
ture of the sediments (clay to silt). From these assumptions, it
follows that the sediment discharge at the OUT stations during
non-rainy days (Q
OUT
× SSC
OUT
) should be about equal to the
sediment discharge of the GL station (Q
GL
× SSC
GL
). As a con-
sequence, the ratio SSC
OUT
/SSC
GL
should be equal to Q
OUT
/Q
GL
.
As such, this allowed for the estimation Q
GL
, based on observed
Q
OUT
,SSC
OUT
,andSSC
GL
values during non-rainy days.
We therefore isolated all measurements conducted during
days without rainfall (considering also the average concentra-
tion time of runoff) and compared the concentrations mea-
sured at the outlet (SSC
OUT
) with those measured at the glacier
mouth (SSC
GL
). The resulting scatterplot (Fig. 3) reveals some
outliers with clearly higher SSC concentrations at the OUT
station. These are likely linked to mass wasting events in the
Koiavgan tributary and/or channel erosion in the river section
between the MID and OUT station. Nonetheless, as expected,
concentrations at the OUT stations are typically lower than at
the GL station. A simple linear relationship excluding these
outliers indicates that sediment concentrations at the OUT
station are on average 15% lower than those at the GL station.
Following the reasoning above, this would imply that, on non-
rainy days, 85% of the runoff discharge at the OUT station
originates from the glacier mouth (GL) (i.e., melt water).
Generic hydrograph separation of the Djankuat river at the
end of the ablation season in 2014 showed similar values
(70–80%; Rets et al. 2017). From this we calculated Q
GL
as
0.85 × Q
OUT
during non-rainy days. These estimated dis-
charges at the glacier outlet could be confronted with their
corresponding observed flow depth (H
GL
)toestablisha
depth-discharge rating curve for the GL station (Fig. 4b).
While the adjusted R
2
value of this rating curve is rather low
(R2
adj = 0.58), it is still acceptable and comparable to the R
2
values for the H–Q curve for the OUT station (cf. Fig. 4a). The
scatter on Q
GL
–H
GL
relationship might be partially attribut-
able to the role of other runoff sources (or sinks) between
the GL and OUT station. However, observation errors on the
H
GL
and Q
OUT
values will also play a role. Overall, the rela-
tionship of Fig. 4b mainly indicates that our assumptions
discussed above are generally justified. As such, this relation-
ship allowed us to estimate the runoff discharge at the glacier
outlet, based on the observed flow depth at the GL station.
Contrary to Fig. 3, this relationship should be also valid on
days with rain.
No glaciers contribute water to the Djankuat stream be-
tween the MID and OUT station (Fig. 1). Likewise, the differ-
ence in contributing area is small (Table 1). We therefore as-
sumed that the runoff discharge at the MID station (Q
MID
)was
equal to the observed discharge at the out station (Q
OUT
).
3.3 Sediment load calculations
We calculated the suspended sediment load per hydrological
event at the OUT, MID, and GL gauging stations as:
J Soils Sediments

SL ¼∑n
k¼1QkSSCk
n106Δt;ð2Þ
where SL is the suspended sediment load (t event
−1
); Q
k
is the
measured or estimated water discharge at moment k(m
3
s
−1
);
SSC
k
is the corresponding measured suspended sediment con-
centration at moment k(g m
−3
); nis the number of pairwise (Q
and SSC) measurements taken during that event; and Δtisthe
duration (s) of the hydrological event.
Furthermore, we have assessed the relative importance of
the main glacier in the total sediment load as:
SLGL;REL ¼100⋅SLGL
SLOUT
ð3Þ
where SL
GL,REL
(%) is the relative proportion of the event
sediment load at the GL station (SL
GL
)ascomparedtothe
corresponding daily sediment load at the OUT station
(SL
OUT
). Likewise, we calculated the relative contribution
Fig. 3 Relationship between
suspended sediment
concentrations at the OUT
(SSC
OUT
)andtheGL(SSC
GL
)
gauging stations for non-rainy
days (see Fig. 1). The 1:1 line is
indicated in yellow. Outliers are
indicated in red with their corre-
sponding timing. The linear re-
gression is based on all other ob-
servations (n=79)
Fig. 4 Runoff discharge (Q)–flow depth (H) relationships: afor the OUT station during different months of the ablations season 2017; bfor the GL
station (cf. Fig. 1) during non-rainy days. H
GL
,H
OUT
,andQ
OUT
were manually observed at the station, while Q
GL
was estimated based on Fig. 3(see text)
J Soils Sediments

(SL
MID,REL
(%)) of the sediment load at the MID station
(SL
MID
)as:
SLMID;REL ¼100⋅SLMID−SLGL
SLOUT
ð4Þ
Overall, lower SL
GL,REL
values indicate that a relatively
larger proportion of sediments do not originate from the gla-
cier, but rather from the catchment hillslopes, the river bed or
river banks. Values above 100% indicate that a proportion of
the sediments originating from the glacier are deposited within
the catchment before reaching the catchment outlet. Similarly,
low SL
MID,REL
values indicate that a significant amount of
sediments originate from the river reach between the MID
and OUT stations (and/or its bordering hillslopes; Fig. 1).
Values above 100% indicate that sediment deposition occurs
along this reach.
3.4 A SHI
As explained in the introduction, we aimed to systematically
analyze observed event-hysteresis patterns and their potential
link with sediment sources. For this, we develop a simple
hysteresis index (SHI). This index was calculated for all sed-
iment transport events recorded at the OUT station for which
at least five SSC
OUT
observations were available, including
one observation at (or very close to) the peak discharge. For
this, we first demarcated all hydrological events. This was
done by smoothing the hourly water discharge values
(Q
OUT
) using a linear moving median function with a window
size of 3 h (as suggested by Rodda and Little (2015)). Next,
we identified the start and end point of each event, using the
local minimum method (Sloto and Crouse 1996).
The proposed SHI is illustrated in Fig. 5. First, we calcu-
lated a log-linear regression between Qand SSC for all obser-
vations made during that event (with a minimum of five). This
regression line represents the SSC that could be expected for a
given Qduring that specific event, should there be no hyster-
esis at all. Next, we calculated the residues of all samples as
the difference between the observed SSC and the estimated
SSC, based on this event rating curve. The SHI then consists
out of the difference between the mean residue of the samples
in the rising limb and the mean residue of the samples in the
falling limb, normalized for the maximum observed SSC:
SHI ¼
∑n
i¼1dRiþ…þdRn
n−∑m
j¼1dFjþ…þdFm
m

SSCmax
;ð6Þ
with dR
i
, the difference between observed and estimated SSC
during the rising limb; dF
j
, the difference between observed
and estimated SSC during the falling limb; n,m, the number of
samples in respectively the rising and falling limb; and
SSC
max
, the maximum observed SSC of the event.
The SHI (Eq. 6) yields values between 1 and −1. As a
result of its structure (cf. Fig. 5), values closer to 1 can be
expected to correspond with events having a strong clock-
wise pattern, while values close to −1 should correspond
to more anti-clockwise events. Events with no strong hys-
teresis or a complex hysteresis pattern (e.g., figure-eight
pattern) can be expected to have a SHI-value closer to zero.
To test the validity of this interpretation, we visually
inspected all events for which the SHI could be calculated
and classified them according to their main hysteresis pat-
tern, as proposed by Williams (1989): anticlockwise loops
(AW), clockwise loops (CW), and figure-eight loops (F8).
Events with no clear hysteresis pattern were classified as
“not applicable”(NA).
4 Results
4.1 Field data
Between June 6, 2017, and September 22, 2017, 37 rainfall
events were recorded with a mean rainfall depth of 13 mm
(standard deviation 18.4 mm) and a mean duration of 5.96 h.
The total precipitation depth of 482 mm was close to average
precipitation amount for this area during the ablation period
(Vasil’chuk et al. 2016). The mean rainfall intensity was
2.63 mm h
−1
(standard deviation 3.21 mm h
−1
) with a maxi-
mum of 17.7 mm h
−1
. The latter occurred during a 100-mm
rainfall event in the night of 31 August–1 September. This
event caused extreme erosion and triggered an outburst of
the Bashkara Lake, located in a neighboring valley
(Chernomorets et al. 2018). The event also damaged the gaug-
ing stations. For the OUT stations, the necessary reparations
were conducted by noon on September 1 and measurements
could continue. However, for the MID and GL station, mea-
surements were stopped after August 30.
At the OUT station (cf. Fig. 1),themeanrecordedwater
discharge during the observation period was 1.39 m
3
s
−1
(stan-
dard deviation 0.46 m
3
s
−1
). The maximum recorded dis-
charge was 3.21 m
3
s
−1
, observed on September 1 (Table 2;
Fig. 6). However, due to the abovementioned damages, the
real occurring maximum discharge might have been higher.
The mean recorded SSC at OUT was 725 g m
−3
(standard
deviation 2980 g m
−3
;cf.Table2; Fig. 6), with a maximum
of 53,800 g m
−3
(likewise measured in the night of September
1, 2017). Recorded SSCs at the MID and GL stations were
relatively lower (respectively 325 and 365 g m
−3
). However,
the median SSC did not differ significantly between the OUT,
MID, and GL station (Table 2).
The mean event suspended sediment load at the OUT station
was 68.3 t event
−1
(standard deviation 170 t event
−1
). The total
measured suspended sediment load (SL) was 9224 t over the 110-
day observation period (136 events). The total SL for the entire
J Soils Sediments

2017 ablation season is expected to be in the order of 9408 t,
corresponding to a sediment yield of 1033 t year
−1
km
−2
.The
mean SL at GL and MID was 37 and 37.4 t event
−1
, respectively
(Table 3). The maximum SL at OUT station was measured during
the hydrological event of 31 August (2017-08-31 22:00:00–2017-
09-01 04:00:00). However, during these events, no measurements
could be collected at the GL and MID station. As such, the aver-
age and maximum SL for these stations over the whole ablation
period of 2017 were likely higher.
4.2 Patterns and trends of hysteresis in sediment
concentrations
For 136 events recorded at the OUT station, enough samples
were collected to calculate the SHI. Of these events 84 were
visually classified as having a clockwise (CW) hysteresis loop
pattern, 16 as anticlockwise (AW), seven as figure eight (F8)
and 29 as having no clear hysteresis loop pattern (NA). As
such, clockwise hysteresis patterns were the most observed
loop type, occurring during 61.8% of the events.
Anticlockwise and figure-eight loops occurred for, respective-
ly, 11.8% and 5.15% of the events (Table 4).
Overall, the mean SSC during anticlockwise events was
higher (but not significantly, according to a Wilcoxon
ranksum test at significance level 0.05) than those of clock-
wise events (668 and 548 g m
−3
, respectively; Table 4).
Clockwise events also typically had a slightly (but significant-
ly, according to a Wilcoxon rank sum test at significance level
0.05) lower mean water discharge (Q) than anticlockwise
events (Table 4). No significant differences in mean Qwere
detected between the other types of events.
We also computed the time lag for each event, i.e. the differ-
ence in timing between peak water discharge and peak SSC.For
CW events, the SSC peaks on average 1.5 h before the Qpeak
(standard deviation 2.32 h). For AW events, the SSC peaks on
average 2.06 h after the Qpeak (standard deviation 1.95 h). We
detected no significant differences in the duration of the events
between the different hysteresis types (see Table 4).
Fig. 5 Example of an observed hysteresis loop, illustrating the
calculation of the proposed simple hysteresis index (SHI;cf.Eq.6). a
Hydrograph and observed sediment concentrations of the event. b
Corresponding hysteresis loop of the event. The dashed line shows the
fitted runoff discharge (Q)–suspended sediment concentration (SSC)
rating curve for this event. In this case, the samples taken during the rising
limb will have clearly positive residues, while those of the falling limb
will be clear negative. As a result, the SHI for this event (Eq. 6)willhavea
clearly positive value, corresponding with the observed a clockwise hys-
teresis pattern
Table 2 Summary statistics of the water discharge (Q
i
,m
3
s
−1
) and suspended sediment concentration (SSC
i
,gm
−3
) at different gauging stations at the
Djankuat river for the 2017th ablation period (with “i”the gauging station; cf. Fig. 1)
Variable Mean Standard deviation Median Maximum Time of maximum Minimum Time of minimum
Q
OUT
(m
3
s
−1
) 1.39 0.465 1.44 3.21 2017-09-01 14:00:00 0.345 2017-06-18 10:00:00
Q
GL
(m
3
s
−1
) 1.38 0.238 1.34 1.94 2017-08-09 19:35:00 0.655 2017-08-28 07:30:00
SSC
OUT
(g m
−3
) 725 2980 237 53,800 2017-09-01 00:00:00 213 2017-06-23 09:00:00
SSC
GL
(g m
−3
) 367 328 267 3010 2017-08-23 17:00:00 215 2017-07-12 13:30:00
SSC
MID
(g m
−3
) 325 292 245 2600 2017-08-23 17:00:00 214 2017-07-07 20:00:00
J Soils Sediments

In terms of sediment load, almost half of the total measured
SL at the OUT station was exported during events with a
clockwise hysteresis loop (4409 t or 47.8%). The total SL
occurring during anticlockwise events was 1325 t (14.4% of
the total SL during the observation period). More than 30% of
the SL was exported during events for which no clear hyster-
esis pattern could be identified.
Figure 7displays the calculated SHI versus the correspond-
ing SL
OUT
and maximum Q
OUT
of each event. For clockwise
events, higher peak runoff discharges and SL typically result
in a higher SHI. As such, the degree of hysteresis overall
increases with the magnitude of the event. For the other event
types, patterns are less clear. For AW events, larger runoff
peak discharges tend also to be associated with slightly lower
SHI values and, hence, a more pronounced negative hysteresis
pattern. However, this trend does not hold for SL (Fig. 7).
4.3 Hysteresis patterns and sediment budget analyses
For 58 out of the 136 events that were recorded at the OUT
station, simultaneous recordings were also available for the MID
and GL station (between June 30 and August 30). Comparison of
the SL from GL and OUT indicated that sediment deposition
occurred in 14 out of the 58 events (24%). The total difference
corresponded with an estimated 305 t of material deposited in the
section between GL and OUT (cf. Fig. 1). For 44 events (76%),
the SL at OUT was higher than at GL, indicating that additional
erosion in the catchment and the channel occurred. These differ-
ences amounted to a total of 642 t of eroded materials. Comparing
the amount of deposition/erosion with the event hysteresis type
revealed no clear pattern (Fig. 8).
To assess the overall importance of the Djankuat gla-
cier in the sediment budget, we calculated SL
GL,REL
(cf.
Fig. 6 aGraphs of 2 m air temperature (°C), precipitation (mm) at the DGS. bWater discharge (m
3
s
−1
) and suspended sediment concentration (g m
−3
)at
the Djankuat river outlet station (OUT) during the 2017 measuring period
Table 3 Suspended sediment load (SL) at the different gauging stations of the Djankuat catchment (cf. Fig. 1) for the 2017 ablation season
Gauging
station
Number of
events
SL during
observation period
1
(t)
Total estimated SL for
ablation period (t)
Mean
(t event
−1
)
Standard deviation
(t event
−1
)
Median
(t event
−1
)
Max
(t event
−1
)
Min
(t event
−1
)
GL 58 2148 –37 37.9 26.5 210 6
MID 58 2168 –37.4 32.6 30.5 220 7
OUT 135 9224 9408 68.3 170 30 1800 1.9
1
Observation period: for GL and MID: 30/06/2017–30/08/2017; for OUT: 06/06/2017–24/09/2017
J Soils Sediments

Eq. 3) for 44 events. From this, we estimated that the
glacier contributed on average 71.8% (median 77.6%)
of the suspended sediment load recorded at the OUT
station.
To obtain more insight in the sediment dynamics between
the GL, MID and OUT station, we also compared the SL
GL,REL
(Eq. 3)andSL
MID,REL
(Eq. 4) with the corresponding SHI of
the event (as recorded at the OUT station) and explored the
observed relationships with respect to time period, rainfall
conditions, magnitude of the event, etc. Figure 9displays
some key results. For non-rainfall events, the SHI is negative-
ly correlated with the relative contribution of the glacier to the
Table 4 Main characteristics of the runoff and sediment transport
events at the OUT station, for which at least five suspended sediment
concentration (SSC) measurements were conducted. A distinction is
made according to the type of observed hysteresis loop (AW,
anticlockwise loops; CW, clockwise loops; F8, figure-eight loops; NA,
not applicable). Qwater discharge, SHI simplified hysteresis index (cf.
Eq. 6); SSC suspended sediment concentration, SL suspended sediment
load
AW CW F8 NA Total
Number of events 16 84 7 29 136*
Proportion of events (%) 11.8 61.8 5.15 21.3 100*
Mean event duration (h) 19.3 20.3 22.3 12.2 18.5**
Mean Qlag time (h) −2.06 1.5 −2.29 –−0.95**
Standard deviation in Qlag time (h) 1.95 2.32 3.99 ––
Mean SHI −0.166 0.134 0.00663 −0.0291 −0.014**
Maximum SHI −0.000628 0.464 0.186 0.516 0.516***
Minimum SHI −0.597 0.000839 −0.117 −0.508 −0.597****
Mean Q(m
3
s
−1
) 1.65 1.32 1.66 1.47 1.52**
Maximum Q(m
3
s
−1
) 3.2 2.8 2.4 2.7 3.2***
Mean SSC (g m
−3
) 668 548 347 2650 1052**
Maximum SSC (g m
−3
) 11,000 19,000 1900 54,000 54000***
Mean SL (t event
−1
) 82.8 52.5 50.9 112 74.5**
Tot al SL [t] 1320 4410 356 3130 9224*
Contribution to total SL [%] 14.4 47.8 3.86 34 100*
*Total sum
**Total mean
***Total maximum
****Total min
Fig. 7 Simple hysteresis index (SHI) for events recorded at the OUT station, versus maximum water discharge (max Q
OUT
(m
3
s
−1
), plot a) and the
logarithm of the suspended sediment load of the event (SL
OUT
(t event
−1
), plot b)
J Soils Sediments

sediment load (SL
GL,REL
; cf. Fig. 9a). As such, relatively larger
contributions of the main glacier (Djakuat glacier) to the total
sediment load are typically associated with less positive
(clockwise) hysteresis loops.
Interestingly, during non-rainfall events, SHI shows no real
correlation with the contribution of the middle reach (Fig. 9b).
However, for significant rainfall events (> 8 mm), SL
MID,REL
shows a clear positive correlation with the SHI (Fig. 9d).
During such rainfall events, additional erosion from the
proglacial area can be expected, resulting in a higher
SL
MID,REL
. Hence, as indicates, larger contributions from the
proglacial areas are also reflected in more positive (clockwise)
hysteresis effects. By consequence, more positive hysteresis
loops during rainfall events are also associated with lower
contributions of the glacier (Fig. 9c).
5 Discussion
5.1 Sediment load and relative contribution
of the glacier
The total measured suspended sediment yield for the Djankuat
catchment for the ablation period of 2017 was
1033 t year
−1
km
−2
. Given that the Djankuat stream has no
significant water discharge outside the ablation period, this
yield will also closely approximate the annual sediment yield.
As always, this sediment yield should be interpreted with cau-
tion as it represents a value for only 1 year and sediment yields
are generally characterized by a large inter-annual variability
(Vanmaercke et al. 2012). Likewise, measuring errors can in-
duce significant uncertainties. A key factor influencing this
uncertainty is generally the SSC sampling frequency (e.g.,
Phillips et al. 1999; Moatar et al. 2006). Overall, the sediment
and runoff sampling frequency of this study can be considered
quiet high as compared to other SY studies (Vanmaercke et al.
2011,2014). Nevertheless, also in our case, measuring errors
likely induced uncertainties. This is especially so for larger
events, as the safety situation often limited the opportunities
to collect samples and the very high recorded turbidities
(above 16,000 NTU) might underestimate the actual SSC.
Despite these potential uncertainties, our measurements in-
dicate that the sediment yield for the Djankuat catchment is
rather high as compared to other estimates available for the
region. Gabrielyan (1971) reported a mean annual suspended
sediment yield for the Northern Caucasus of
900 t year
−1
km
−2
. Mozzherin and Sharifullin (2015)estimat-
ed that the mean contemporary annual denudation rate of the
Greater Caucasus ranges between 0.1 and 0.25 mm year
−1
.
According to their map, the mean annual suspended sediment
yield in the study area would be around 660 t year
−1
km
−2
.A
more recent interpolation of gauging station data in the region
concurs with this, resulting in an estimated mean annual
suspended sediment yield for the study area of approximately
700 t year
−1
km
−2
(Tsyplenkov et al. 2019). This relatively
higher SY for the Djankuat catchment is most likely attribut-
able to the presence of the glaciers and the large sediment
availability in the sparsely vegetated proglacial area. As
such, our estimated SY is relatively low as compared to
other glaciated catchments. For example, Gurnell et al.
Fig. 8 Difference in the sediment load at the GL and OUT station for
different events for which samples were simultaneously collected at both
stations. Negative values indicate additional erosion, while positive
values indicate sediment deposition. Hydrological events are colored
according to their hysteresis type. Event numbers are chronological and
refer to the events in Supplementary 1
J Soils Sediments

(1996) report a SY of 1800 t year
−1
km
−2
for the Glacier de
Tsidjiore Nouve in the Alps (Switzerland). Hodgkins et al.
(2003)reportaSY of 2250 t year
−1
km
−2
for an Artic Glacier
on Svalbard. Nevertheless, as compared to other reported de-
nudation rates for glaciated catchments worldwide, the SY of
the Djankuat falls within the expected order of magnitude for
its catchment size and glacial extent (Hallet et al. 1996).
Taking less than 0.02 % to the Terek river catchment area,
the Djankuat river in 2017 contributed to 0.12 % of the
Terek river sediment yield measured at the dowstream
Kartgalinsky gauging station in Dagestan.
Our measurements indicated that the Djankuat glacier is the
dominant suspended sediment source, accounting for about
71.8% of the sediment export. By consequence, 28.2% of
the sediment load likely originates from the proglacial zone
(corresponding to about 291 t km
−2
year
−1
). Also, these values
correspond well to those reported by other studies; e.g.,
O’Farell et al. (2009) report that in a small glaciated catchment
in Alaska (catchment area 16.7 km
2
), 10 ± 7% of the sediment
yield comes from the proglacial zone. Likewise, in the Alpine
catchment of the Arolla glacier (catchment area 7.56 km
2
),
approximately 77% of the sediment load originates from the
stream at the glacial snout and the moraine deposition zone
(Warburton 1990).
Nevertheless, these sediment contributions should be
interpreted with caution as they may vary over time. For ex-
ample, Mao and Carrillo (2017) showed for a Chilean Andean
glacier that slope and channel processes are the dominant
source at the beginning and end of the ablation season. Such
potential temporal dynamic could not be observed in the
Djankuat glacier, due to lack of data in June and September.
Nevertheless, we detected no significant differences between
the period of snowmelt (July) and the period where only the
glacier delivered meltwater (August). Our measurements did
indicate, however, that erosion due to rainfall becomes in-
creasingly important through August–September.
Our analyses of hysteresis patterns might shed some addition-
al light on the relative contribution of different sediment sources.
Overall, the fact that clockwise hysteresis loops are the most
frequently occurring type observed in the Djankuat catchment
(cf. Table 4) concurs with other studies in similar environments.
For example, a study in the catchment of the Gangotri Glacier
(western Himalayan) showed that clockwise diurnal loops are the
most commonly occurring during the entire ablation season
Fig. 9 The relationship between
SL
GL,REL
(%) (cf. Eq. 3),
SL
MID,REL
(%) (cf. Eq. 4)andSHI
(cf. Eq. 6) for non-rainfall events
(a,b), as well as between
SL
GL,REL
(%), SL
MID,REL
(%) and
SHI for rainfall events (c,d)
J Soils Sediments

(Singh et al. 2005).Likewise,inJapan,Iidaetal.(2012) report
that intra-event clockwise loops are more common than anti-
clockwise, which were only observed during the snow-melt pe-
riod. On the other hand, a similar study for the Vantaa River in
southern Finland (Kämäri et al. 2018) indicated that events in
snowmelt periods are typically characterized by clockwise hys-
teresis loops, while counterclockwise loops are more common in
other seasons.
A commonly accepted explanation for positive (clockwise)
hysteresis loops is the depletion of sediment sources in the
proglacial area (Klein 1984; Mao and Carrillo 2017). Also,
in non-glacial environments, the occurrence of clockwise
loops is often attributed to the depletion of sediments during
runoff events that cause overland erosion (Williams 1989;
López-Tarazón et al. 2009; Vanmaercke et al. 2010;Sun
et al. 2016). Anti-clockwise loops are usually linked to the
release of sediments after a certain threshold of water dis-
charge has been exceeded. This can due to streambank erosion
or mass transport (Williams 1989; Iida et al. 2012). Other
causes for anti-clockwise loops can be the rupture of an armor
layer on the river bed (Morehead et al. 2003; Vanmaercke
et al. 2010) or the entrainment of sediments that were previ-
ously deposited in floodplains (e.g., Oeurng et al. 2010).
Our statistical analyses, including the proposed SHI,allow
to further test and explore these different hypotheses. For ex-
ample, the idea of sediment depletion as a cause of clockwise
loops seems consistent with the fact that larger clockwise
events are characterized by a more positive SHI and thus a
relatively larger proportion of sediments transported during
the rising limb of the event (Fig. 7). As such, larger events
seem to be capable to flush out and deplete available sedi-
ments relatively faster than smaller events. Furthermore, for
most clockwise events, the sediment load at the outlet is larger
than at the glacier snout (Fig. 8), pointing to the entrainment of
additional sediments from the proglacial area. In addition,
during none-rainfall events (when the dominant source of
the river discharge is the glacier), relatively larger contribu-
tions of the glacier to the total sediment load are also associ-
ated with less positive SHI values (Fig. 9a). This suggests that
the glacier itself provides a relatively constant supply of sed-
iments (resulting in little to no hysteresis patterns). However,
when also the proglacial areas supply sediments (resulting in a
relatively lower SL
GL,REL
), events are characterized by more
positive hysteresis patterns. The later could be an indication of
sediment depletion. This tendency becomes especially clear
during rainfall events (when not only the proglacial channels
but also hillslope erosion can contribute to the sediment load).
Higher SHI values clearly appear more associated with rela-
tively larger contributions from the proglacial area (Fig. 9c, d).
Overall, this indicates that larger contributions from the
proglacial area result in more clockwise hysteresis loops.
Although clearly less frequent, also anti-clockwise loops were
observed in several cases (Table 4). Nonetheless, anti-clockwise
loops do seem to be associated with relatively large peak dis-
charges (Table 4;Fig.7). This indicates that they may be caused
by the re-entrainment of sediments stored along the riverbanks.
More specifically, the valley Sandur between the OUT and MID
station (Fig. 1) may function as a temporal sediment storage that
re-releases sediments during sufficiently large subsequent events.
They may also be attributable to changes in sediment supply
from the subglacial area or contributions of more distant sedi-
ment sources (e.g., Klein 1984; Mao and Carrillo 2017).
However, the later may be less likely, given the limited size of
the catchment. Another explanation can be the supply of sedi-
ments due to streambank erosion or rock and boulder falls from
lateral moraines during later stages of the event (Williams 1989;
Vanmaercke et al. 2010;Iidaetal.2012).
Nevertheless, these sediment dynamics remain difficult to
capture based on SSC observations and hysteresis patterns alone.
For example, several events with a similar peak discharge result-
ed in a clockwise rather than an anti-clockwise loop (Fig. 7). For
three of these events, our data indicates that sediment deposition
occurred along the above-mentioned Sandur (Fig. 8). As such,
not all clockwise loops are necessarily attributable to sediment
depletion. Also, the overtopping of riverbanks and the conse-
quent settling of suspended sediments in the flooded zone may
cause a clockwise hysteresis loop. However, our analyses re-
vealed no systematic pattern (e.g. in relation to timing, weather
conditions, event magnitude or hysteresis loop characteristics) as
to why some events seem to result in a net deposition or net
entrainment of sediments (Fig. 8). Fully disentangling the sedi-
ment dynamics of the Djankuat and other similar catchments will
therefore require the use of additional tools, such as sediment
fingerprinting techniques.
Furthermore, the results and hysteresis patterns presented
in this paper should be interpreted with caution as they repre-
sent only one ablation season. The expected glacier retreat
may make new sediment sources available over time while
older ones may become exhausted (Leggat et al. 2015).
Furthermore, sediment contributions can vary strongly in re-
lation to the timing and magnitude of precipitation or melt
events. Continued high-resolution monitoring over consecu-
tive years is therefore required to better understand the
suspended sediment dynamics of proglacial catchments.
6 Conclusions
In this paper, we presented the results on runoff and sediment
transport in glaciated Djankuat catchment (Northern
Caucasus) for the ablation period of 2017, based on an exten-
sive dataset of runoff discharge and sediment concentration
observations collected at three gauging stations. The total
suspended sediment export at the catchment outlet during
the observation period (June 6–September 24) was 9224 t.
An estimated 71.8% of this sediment load originated from
J Soils Sediments

the glacier. However, during rainfall events, also hillslope ero-
sion in the proglacial area can become a very significant sed-
iment source.
We further explored intra-event sediment dynamics of >
130 events by statistically analyzing occurring hysteresis pat-
terns. For this, we proposed a new simple hysteresis index
(SHI;Eq.6). This index weighs the relative importance of
sediment concentrations during the rising and falling limb. It
can be calculated based on a fairly limited number of samples
(five or more) and can be used to robustly quantify hysteresis
patterns in a continuous way. This not only avoids interpreta-
tion difficulties when manually classifying hysteresis loop
patterns, but also allows more refined quantitative analyses.
Like earlier studies, we found that the majority (> 60%) of
hysteresis loops are clockwise (Table 4). Analyses based on
the SHI further showed that larger events (both in terms of
sediment load and runoff peak discharge) are typically asso-
ciated with more pronounced clockwise patterns (Figs. 7and
8). This may point towards the importance of sediment deple-
tion during events. Nonetheless, also sediment deposition dur-
ing riverbank overtopping may result in a clockwise loop. As
such, disentangling the exact causes of a certain hysteresis
loop type remains very difficult. Nevertheless, our analyses
(e.g., Fig. 9) clearly indicate that clockwise loop patterns are
mainly caused by sediment contributions from the proglacial
area rather than from the glacier itself. Additional tools (such
as sediment fingerprinting) are required to further confirm and
investigate this.
Overall, our results indicate that hysteresis loop types
can indeed provide (to some extent) meaningful infor-
mation on sediment sources within the catchment. They
can help in better understanding the sediment dynamics
of glaciated catchment. This is especially relevant in the
context of climate change. Therefore, our methodology
(and in particular the proposed SHI) may also be useful
in other contexts.
Acknowledgments This study was conducted within the project of the
Russian Scientific Foundation No. 19-17-00181 and a stay with an
Erasmus+ International Credit Mobility scholarship. This study contrib-
utes to the State Task no. 0148- 2019-0005, Institute of Geography
RAS. The help of Ekaterina Rets, Andrey Smirnov, Pavel Toropov,
Varvara Bazilova, Anna Avilova, Ekaterina Kornilova, and all other stu-
dents who helped with the collection and processing of the field data is
gratefully acknowledged.
Author contributions Conceptualization: Matthias Vanmaercke, Anatoly
Tsyplenkov, Valentin Golosov; Methodology: Matthias Vanmaercke,
Anatoly Tsyplenkov; Data collection and analyses: Anatoly Tsyplenkov
with input from Matthias Vanmaercke; Writing: Anatoly Tsyplenkov,
Matthias Vanmaercke, Valentin Golosov; Funding acquisition: Valentin
Golosov, Sergey Chalov; Resources: Sergey Chalov; Supervision:
Valentin Golosov, Matthias Vanmaercke.
Funding This study was funded by project of the Russian Scientific
Foundation No. 19-17-00181.
Data availability Reproducible R code is available at the GitHub repos-
itory (https://github.com/atsyplenkov/intra-event-djankuat). Contact
Anatoly Tsyplenkov (atsyplenkov@gmail.com) for more information.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
Ethical approval This chapter does not contain any studies with human
participants or animals performed by any of the authors.
Informed consent Informed consent was obtained from all individual
participants included in the study.
References
Aich V, Zimmermann A, Elsenbeer H (2014) Quantification and inter-
pretation of suspended-sediment discharge hysteresis patterns: how
much data do we need? Catena 122:120–129. https://doi.org/10.
1016/j.catena.2014.06.020
Barry RG (2006) The status of research on glaciers and global glacier
recession: a review. Prog Phys Geogr 30:285–306. https://doi.org/
10.1191/0309133306pp478ra
Beniston M (2003) Climatic change in mountain regions: a review of
possible impacts. Clim Chang 59:5–31. https://doi.org/10.1023/A:
1024458411589
Bogen J (1989) Glacial sediment production and development of hydro-
electric power in glacierized areas. Ann Glaciol 13:6–11
Carrivick JL, Geilhausen M, Warburton J, Dickson NE, Carver SJ, Evans
AJ, Brown LE (2013) Contemporary geomorphological activity
throughout the proglacial area of an alpine catchment.
Geomorphology 188:83–95. https://doi.org/10.1016/j.geomorph.
2012.03.029
Chernomorets SS, Petrakov DA, Aleynikov AA et al (2018) The outburst
of Bashkara glacier lake (Central Caucasus, Russia) on September 1,
2017. Kriosf Zemli XXII:70–80. https://doi.org/10.21782/KZ1560-
7496-2018-2(70-80) (In Russian)
Dedkov AP, Moszherin VI (1984) Erosion and sediment runoff on the
Earth. Izd-vo Kazansk. un-ta, Kazan (In Russian)
Dobriyal P, Badola R, Tuboi C, Hussain SA (2017) A review of methods
for monitoring streamflow for sustainable water resource manage-
ment.ApplWaterSci7:2617–2628. https://doi.org/10.1007/
s13201-016-0488-y
Durgerov MB, Freindlin VS, Chernova LP (1972) Suspended sediment
load in the Djankuat glacier basin during the summer of 1970. Mater
Glaciol Issled 19:253–254 (In Russian)
Dyurgerov M (2003) Mountain and subpolar glaciers show an increase in
sensitivity to climate warming and intensification of the water cycle.
J Hydrol 282:164–176. https://doi.org/10.1016/S0022-1694(03)
00254-3
Farnsworth KL, Milliman JD (2003) Effects of climatic and anthropo-
genic change on small mountainous rivers: the Salinas River exam-
ple. Glob Planet Chang 39:53–64. https://doi.org/10.1016/S0921-
8181(03)00017-1
Favaro EA, Lamoureux SF (2015) Downstream patterns of suspended
sediment transport in a high Arctic river influenced by permafrost
disturbance and recent climate change. Geomorphology 246:359–
369. https://doi.org/10.1016/j.geomorph.2015.06.038
Gabrielyan GK (1971) The intensity of denudation in the caucasus.
Geomorphol RAS 1:22–27 (In Russian)
J Soils Sediments

Gao P, Josefson M (2012) Event-based suspended sediment dynamics in
a central New York watershed. Geomorphology 139–140:425–437.
https://doi.org/10.1016/j.geomorph.2011.11.007
Geilhausen M, Morche D, Otto J-C, Schrott L (2013) Sediment discharge
from the proglacial zone of a retreatingAlpine glacier. Zeitschrift für
Geomorphol Suppl Issues 57:29–53. https://doi.org/10.1127/0372-
8854/2012/S-00122
Glazirin GE, Semakova ER (2019) Estimation of Dispersed Glaciation
Shrinkage Under Climate Change. Geography, Environment,
Sustainability 12(1):5–12
Gurnell A, Hannah D, Lawler D (1996) Suspended sediment yield from
glacier basins. In: Walling DE, Webb BW (eds) Erosion and sedi-
ment yield: global and regional perspectives. IAHS Publ 236. IAHS
Press, Wallingford, pp 97–104
Hallet B, Hunter L, Bogen J (1996) Rates of erosion and sediment evac-
uation by glaciers: a review of field data and their implications. Glob
Planet Chang 12:213–235. https://doi.org/10.1016/0921-8181(95)
00021-6
Hamshaw SD, Dewoolkar MM, Schroth AW, Wemple BC, Rizzo DM
(2018) A new machine-learning approach for classifying hysteresis
in suspended-sediment discharge relationships using high-frequency
monitoring data. Water Resour Res 54:4040–4058. https://doi.org/
10.1029/2017WR022238
Hinderer M (2012) From gullies to mountain belts: a review of sediment
budgets at various scales. Sediment Geol 280:21–59. https://doi.org/
10.1016/j.sedgeo.2012.03.009
Hodgkins R, Cooper R, Wadham J, Tranter M (2003) Suspended sedi-
ment fluxes in a high-Arctic glacierised catchment: implications for
fluvial sediment storage. Sediment Geol 162:105–117. https://doi.
org/10.1016/S0037-0738(03)00218-5
Iida T, Kajihara A, Okubo H, Okajima K (2012) Effect of seasonal snow
cover on suspended sediment runoff in a mountainous catchment. J
Hydrol 428–429:116–128. https://doi.org/10.1016/j.jhydrol.2012.
01.029
Jaoshvili S (2002) The rivers of the Black Sea. EEA Tech Rep 71:58
Jones JAA (1999) Climate change and sustainable water resources: plac-
ing the threat of global warming in perspective. Hydrol Sci J 44:
541–557. https://doi.org/10.1080/02626669909492251
Kämäri M, Tattari S, Lotsari E, Koskiaho J, Lloyd CEM (2018) High-
frequency monitoring reveals seasonal and event-scale water quality
variation in a temporally frozen river. J Hydrol 564:619–639. https://
doi.org/10.1016/j.jhydrol.2018.07.037
Khmeleva NV, Vinogradova NN, Samoilova AA, Shevchenko BF (2000)
Mountain river basin and exogenous processes within it. LMSU
Faculty of Geography, Moscow (In Russian)
Khmeleva NV, Shevchenko BF (2006) The role of screes in the formation
of the mountain rivers float and correlation of scree dynamics with
solar cycles (according to stationary time series data). Geomorphol
RAS 78:78-86. https://doi.org/10.15356/0435-4281-2006-4-78-86
(In Russian)
Klein M (1984) Anti clockwise hysteresis in suspended sediment concen-
tration during individual storms: Holbeck catchment; Yorkshire,
England. Catena 11:251–257. https://doi.org/10.1016/0341-
8162(84)90014-6
Koppes MN, Montgomery DR (2009) The relative efficacy of fluvial and
glacial erosion over modern to orogenic timescales. Nat Geosci 2:
644–647. https://doi.org/10.1038/ngeo616
Kutuzov S, Lavrentiev I, Smirnov A, Nosenko G, Petrakov D (2019)
Volume changes of Elbrus glaciers from 1997 to 2017. Front Earth
Sci 7:153. https://doi.org/10.3389/feart.2019.00153
Kutuzov S, Shahgedanova M (2009) Glacier retreat and climatic variabil-
ity in the eastern Terskey–Alatoo, inner Tien Shan between the
middle of the 19th century and beginning of the 21st century. Glob
Planet Chang 69:59–70. https://doi.org/10.1016/j.gloplacha.2009.
07.001
Langlois JL, Johnson DW, Mehuys GR (2005) Suspended sediment dy-
namics associated with snowmelt runoff in a small mountain stream
of Lake Tahoe (Nevada). Hydrol Process 19:3569–3580. https://doi.
org/10.1002/hyp.5844
Lavrentiev II, Kutuzov SS, Petrakov DA, Popov GA, Popovnin VV
(2015) Ice thickness, volume and subglacial relief of Djankuat gla-
cier (Central Caucasus). Ice Snow 128:7. https://doi.org/10.15356/
2076-6734-2014-4-7-19
Lawler DM, Petts GE, Foster IDL, Harper S (2006) Turbidity dynamics
during spring storm events in an urban headwater river system: The
Upper Tame, West Midlands, UK. Sci Total Environ 360:109–126.
https://doi.org/10.1016/j.scitotenv.2005.08.032
Leggat MS, Owens PN, Stott TA, Forrester BJ, Déry SJ, Menounos B
(2015) Hydro-meteorological drivers and sources of suspended sed-
iment flux in the pro-glacial zone of the retreating Castle Creek
Glacier, Cariboo Mountains, British Columbia, Canada. Earth Surf
Process Landf 40:1542–1559. https://doi.org/10.1002/esp.3755
Lloyd CEM, Freer JE, Johnes PJ, Collins AL (2016) Using hysteresis
analysis of high-resolution water quality monitoring data, including
uncertainty, to infer controls on nutrient and sediment transfer in
catchments. Sci Total Environ 543:388–404. https://doi.org/10.
1016/j.scitotenv.2015.11.028
López-Tarazón JA, Batalla RJ, Vericat D, Francke T (2009) Suspended
sediment transport in a highly erodible catchment: The River
Isábena (Southern Pyrenees). Geomorphology 109:210–221.
https://doi.org/10.1016/j.geomorph.2009.03.003
Mano V, Nemery J, Belleudy P, Poirel A (2009) Assessment of suspended
sediment transport in four alpine watersheds (France): influence of
the climatic regime. Hydrol Process 23:777–792. https://doi.org/10.
1002/hyp.7178
Mao L, Carrillo R (2017) Temporal dynamics of suspended sediment
transport in a glacierized Andean basin. Geomorphology 287:116–
125. https://doi.org/10.1016/j.geomorph.2016.02.003
Moatar F, Person G, Meybeck M, Coynel A, Etcheber H, Crouzet P
(2006) The influence of contrasting suspended particulate matter
transport regimes on the bias and precision of flux estimates. Sci
Total Environ 370:515–531. https://doi.org/10.1016/j.scitotenv.
2006.07.029
Moore RD, Fleming SW, Menounos B, Wheate R, Fountain A, Stahl K,
Holm K, Jakob M (2009) Glacier change in western NorthAmerica:
influences on hydrology, geomorphic hazards and water quality.
Hydrol Process 23:42–61. https://doi.org/10.1002/hyp.7162
Morche D, Baewert H, Schuchardt A, Faust M, Weber M, Khan T (2019)
Fluvial sediment transport in the proglacial Fagge river, Kaunertal,
Austria. In: Heckmann T, Morche D (eds) Geomorphology of
proglacial systems. Springer, Berlin, Germany, pp 219–229
Morehead MD, Syvitski JP, Hutton EWH, Peckham SD (2003) Modeling
the temporal variability in the flux of sediment from ungauged river
basins. Glob Planet Chang 39:95–110. https://doi.org/10.1016/
S0921-8181(03)00019-5
Mozzherin VV, Sharifullin AG (2015) Estimation of current denudation
rate of the mountains based on the suspended sediment runoff of the
rivers (the Tien Shan, the Pamir-Alai, the Caucasus, and the Alps as
an example). Geomorphol RAS 1:15–23. https://doi.org/10.15356/
0435-4281-2014-1-15-23 (In Russian)
O’Farrell CR, Heimsath AM, Lawson DE et al (2009) Quantifying
periglacial erosion: insights on a glacial sediment budget,
Matanuska Glacier, Alaska. Earth Surf Process Landf 34:2008–
2022. https://doi.org/10.1002/esp.1885
Oeurng C, Sauvage S, Sánchez-Pérez J-M (2010) Dynamics of
suspended sediment transport and yield in a large agricultural catch-
ment, southwest France. Earth Surf Process Landf 35:1289–1301.
https://doi.org/10.1002/esp.1971
Otto J-C, Schrott L, Jaboyedoff M, Dikau R (2009) Quantifying sediment
storage in a high alpine valley (Turtmanntal, Switzerland). Earth
Surf Process Landf 34:1726–1742. https://doi.org/10.1002/esp.1856
J Soils Sediments

Phillips JM, Webb BW, Walling DE, Leeks GJL (1999) Estimating the
suspended sediment loads of rivers in the LOIS study area using
infrequent samples. Hydrol Process 13:1035–1050. https://doi.org/
10.1002/(SICI)1099-1085(199905)13:7<1035::AID-HYP788>3.0.
CO;2-K
Polade SD, Pierce DW, Cayan DR, Gershunov A, Dettinger MD (2015)
The key role of dry days in changing regional climate and precipi-
tation regimes. Sci Rep 4:4364. https://doi.org/10.1038/srep04364
Rets E, Chizhova JN, Loshakova N, Tokarev I, Kireeva MB, Budantseva
NA, Vasil’chuk YK, Frolova N, Popovnin V, Toropov P, Terskaya E,
Smirnov AM, Belozerov E, Karashova M (2017) Using isotope
methods to study alpine headwater regions in the Northern
Caucasus and Tien Shan. Front Earth Sci 11:531–543. https://doi.
org/10.1007/s11707-017-0668-6
Rets EP, Popovnin VV, Toropov PA, Smirnov AM, Tokarev IV, Chizhova
JN, Budantseva NA, Vasil&amp;apos;chuk YK, Kireeva MB,
Ekaykin AA, Veres AN, Aleynikov AA, Frolova NL, Tsyplenkov
AS, Poliukhov AA, Chalov SR, Aleshina MA, Kornilova ED
(2019) Djankuat glacier station in the North Caucasus, Russia: a
database of glaciological, hydrological, and meteorological obser-
vations and stable isotope sampling results during 2007–2017. Earth
Syst Sci Data 11:1463–1481. https://doi.org/10.5194/essd-11-1463-
2019
Rodda HJE, Little MA (2015) Understanding mathematical and statistical
techniques in hydrology. Wiley, Chichester
Rybak OO, Rybak EA (2018) Model-based calculations of surface mass
balance of mountain glaciers for the purpose of water consumption
planning: focus on Djankuat Glacier (Central Caucasus). IOP Conf
Ser Earth Environ Sci 107:012041. https://doi.org/10.1088/1755-
1315/107/1/012041
Singh P, Haritashya UK, Ramasastri KS, Kumar N (2005) Diurnal vari-
ations in discharge and suspended sediment concentration, including
runoff-delaying characteristics, of the Gangotri Glacier in the
Garhwal Himalayas. Hydrol Process 19:1445–1457. https://doi.
org/10.1002/hyp.5583
Slaymaker O (2018) A global perspective on denudation data, primarily
specific sediment yield in mountainous regions. Landf Anal 36:19–
31. https://doi.org/10.12657/landfana.036.003
Sloto RA, Crouse MY (1996) Hysep: a computer program for streamflow
hydrograph separation and analysis. U. S. Geological Survey Water
Resources Investigations Rep 96-4040, USGS, Washington, DC,
USA
Solomina ON (2000) Retreat of mountain glaciers of northern Eurasia
since the Little ice age maximum. Ann Glaciol 31:26–30. https://doi.
org/10.3189/172756400781820499
Stokes CR, Popovnin V, Aleynikov A, Gurney SD, Shahgedanova M
(2007) Recent glacier retreat in the Caucasus Mountains, Russia,
and associated increase in supraglacial debris cover and supra
−/proglacial lake development. Ann Glaciol 46:195–203. https://
doi.org/10.3189/172756407782871468
Stott TA, Mount NJ (2007) Alpine proglacial suspended sediment dy-
namics in warm and cool ablation seasons: implications for global
warming. J Hydrol 332:259–270. https://doi.org/10.1016/j.jhydrol.
2006.07.001
Sun L, Yan M, Cai Q, Fang H (2016) Suspended sediment dynamics at
different time scales in the Loushui River, south-Central China.
Catena 136:152–161. https://doi.org/10.1016/j.catena.2015.02.014
Syvitski JPM, Milliman JD (2007) Geology, geography, and humans
battle for dominance over the delivery of fluvial sediment to the
coastal ocean. J Geol 115:1–19
Tielidze LG, Lomidze N, Asanidze L (2015) Glaciers retreat and climate
change effect during the last one century in the Mestiachala river
basin, Caucasus mountains, Georgia. Earth Sci 4:72-79. https://doi.
org/10.11648/j.earth.20150402.12
Tielidze LG, Wheate RD (2018) The greater Caucasus glacier inventory
(Russia, Georgia and Azerbaijan). Cryosph 12:81–94. https://doi.
org/10.5194/tc-12-81-2018
Toropov PA, Shestakova AA, Smirnov AM (2017) Methodological as-
pects of heat balance components estimation on mountain glaciers.
Russ J Earth Sci 17:1–9. https://doi.org/10.2205/2017ES000605
Tsyplenkov A, Vanmaercke M, Golosov V (2019) Contemporary
suspended sediment yield of Caucasus mountains. In: Chalov S,
Golosov V, Collins A, Stone M (eds) Lane use and climate change
impacts on erosion and sediment transport. IAHS Pub 381. IAHS
Press, Wallingford, pp 87–93. https://doi.org/10.5194/piahs-381-87-
2019
Vanmaercke M, Poesen J, Broeckx J, Nyssen J (2014) Sediment yield in
Africa. Earth-Sci Rev 136:350–368. https://doi.org/10.1016/j.
earscirev.2014.06.004
Vanmaercke M, Poesen J, Radoane M, Govers G, Ocakoglu F,
Arabkhedri M (2012) How long should we measure? An exploration
of factors controlling the inter-annual variation of catchment sedi-
ment yield. J Soils Sediments 12:603–619. https://doi.org/10.1007/
s11368-012-0475-3
Vanmaercke M, Poesen J, Verstraeten G, de Vente J, Ocakoglu F (2011)
Sediment yield in Europe: spatial patterns and scale dependency.
Geomorphology 130:142–161. https://doi.org/10.1016/j.geomorph.
2011.03.010
Vanmaercke M, Zenebe A, Poesen J, Nyssen J, Verstraeten G, Deckers J
(2010) Sediment dynamics and the role of flash floods in sediment
export from medium-sized catchments: a case study from the semi-
arid tropical highlands in northern Ethiopia. J Soils Sediments 10:
611–627. https://doi.org/10.1007/s11368-010-0203-9
Va s i l’chuk YK, Rets EP, Chizhova JN et al (2016) Hydrograph separation
of the Dzhankuat River, North Caucasus, with the use of isotope
methods. Water Resour 43:847–861. https://doi.org/10.1134/
S0097807816060087
Vinogradova NN, Krylenko IV, Sourkov VV (2007) Some regularities of
channel formation in the upper course of mountain rivers (Baksan
river as an example). Geomorphol RAS 2:49-56. https://doi.org/10.
15356/0435-4281-2007-2-49-56 (In Russian)
Walling DE (1995) Suspended sediment yields in a changing environ-
ment. In: Gurnell AM, Petts GE (eds) Changing River Channels.
Wiley, Chiche ster, pp 149–176
Warburton J (1990) An alpine proglacial fluvial sediment budget. Geogr
Ann Ser A Phys Geogr 72:261–272. https://doi.org/10.1080/
04353676.1990.11880322
Williams GP (1989) Sediment concentration versus water discharge dur-
ing single hydrologic events in rivers. J Hydrol 111:89–106. https://
doi.org/10.1016/0022-1694(89)90254-0
Publisher’snoteSpringer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.
J Soils Sediments