Content uploaded by Ekaterina Rets
Author content
All content in this area was uploaded by Ekaterina Rets on Mar 18, 2018
Content may be subject to copyright.
Content uploaded by Ekaterina Rets
Author content
All content in this area was uploaded by Ekaterina Rets on Jun 17, 2016
Content may be subject to copyright.
Global Change: Facing Risks and Threats to Water Resources (Proc. of the Sixth World
FRIEND Conference, Fez, Morocco, October 2010). IAHS Publ. 340, 2010.
Copyright 2010 IAHS Press
1
Hazardous hydrological processes in mountainous areas under
the impact of recent climate change: case study of Terek River
basin
EKATERINA RETS & MARIA KIREEVA
Moscow State University, Leninskie gory, GSP-1, Moscow, 119991 Russia
retska@mail.ru
Abstract The study focused on hazardous hydrological processes in mountainous areas. The general
objective was to analyse the spatiotemporal distribution of the characteristics of such processes in the Terek
River basin and to examine the main approaches to calculating and forecasting these processes. The study
mostly deals with maximum and minimum water flow and debris flow.
Key words hazardous hydrological processes; mountain hydrology; climate change; Russia; Caucasus;
physically-based mathematical modeling; snow and ice melting
INTRODUCTION
Hazardous hydrological processes arise from the combination of natural factors and human impact.
They are dangerous because they may cause damage to the population and economy. The Terek
River basin is one of Russia’s regions with a problematic water supply. In the foothills of the
basin, the overall performance of water management suffers from slope gully erosion and
mudflows. Floods, channel deformations, landslides, etc. are very likely during wet seasons, while
water resources may be deficient in dry seasons.
One of the main tasks of hydrology is to ensure the hydrological security for a territory. This
means to provide the relationships between water bodies, the population, and its social and
economic activity that can guarantee dependable water supply, and eliminate (or minimize) the
potential causes of hazardous hydrological processes and water deterioration.
First, this task implies identifying the causes of hazardous processes, estimating their
repeatability, scope, and the unidirectional or cyclic changes in their hydrological characteristics.
Moreover, the hydrology should provide a means of calculating and forecasting the main
characteristics of hazardous hydrological processes.
One of the most promising methods for such calculations and forecasts is physically-based
mathematical modelling of flow formation. The main source of river water in mountainous basins
is snow and ice melting. Therefore the first step in the development of a hydrological model for
mountain rivers is to establish an algorithm that describes snow and ice melting. Although the
mechanism involved in melting is fully understood and has been described, the majority of the
best-known mountain river basin models usually deal with empirical relationships between air
temperature and melting depth. Such models cannot provide sufficient accuracy for flow
calculation and forecasting. This is why it is necessary to develop a model of snow and ice melting
in a mountain river basin that will correspond to the up-to-date measurement facilities. Such a
model has been developed in the course of our study.
Although physically-based mathematical models provide one the best opportunities of getting
accurate results, it is however difficult to meet their data demands. In general, the calibration,
validation, and verification of a physically-based model requires a large body of detailed
observational hydrological and meteorological data on the study area. However, we usually face
the lack of hydro-meteorological data in the majority of mountainous catchments. For this reason,
physically-based mathematical models can hardly be applied. Therefore, for catchments with
limited hydro-meteorological data, less exact methods have become more useful and thus more
commonly used in practice.
Ekaterina Rets & Maria Kireeva
2
WATER REGIME AND HAZARDOUS HYDROLOGICAL PROCESSES IN TEREK
RIVER BASIN
The Terek River basin is located in southern European Russia, within the Northern Caucasus
mountain system (Fig. 1). The basin area is 43 200 km2, its elevation ranges from –28 to 5642 m.
The Terek River is 623 km long, and its average annual flow is 294 m3/s. The climate here is
moderate continental. The precipitation decreases both southeastwards and with a decrease in
elevation. The average annual precipitation depth is about 400–600 mm in the lower part of Terek
basin and about 800–1000 mm in the mountainous area. The glacier-covered area is 597 km2 (1.4%
of the total area). The main source of river water in the Terek basin is snow and ice melting (44%), a
comparable amount is supplied by groundwater (37%), while the role of liquid precipitation is the
least (19%). The contrasting relief and climate create the prerequisites for the development of
various hazardous natural processes, constantly endangering the life and activity of people.
Fig. 1 Terek River basin.
The water regime and main characteristics of related hazardous hydrological processes depend
on (Figs 2–4, 6):
– the share of snow and ice melting in the alimentation of rivers and hence, on the altitude and
the glaciation coefficient of the drainage basin,
– the size of the river,
– the morphological characteristics of the river basin.
Fig. 2 Dependence of the date of the establishment of the maximum water levels D Hmax (a) and average
of distribution of the number of month when the minimum flow is observed ml.m. (b) on mean altitude
of the watershed.
(a)
(b)
Hazardous hydrological processes in mountainous areas under the impact of recent climate change
3
The highest value of mean annual modulus of flow Mmeanann = 20–50 L/(s·km2) is
characteristic of relatively small rivers with drainage area not more then 2500 km2, which arises in
the highest altitude, fed by snow and ice melting in the nival belt (Fig. 4). The water-abundant
period, which lasts from April–May to September–October, is prolonged and steady (Fig. 5(b)).
The fundamental wave of runoff hydrograph, formed by snow and ice melting, is overlain with
sharp peaks of rain floods. The maximum water levels are usually recorded in July (Fig. 2(a)),
while their drop starts in August. A stable winter low-flow period with the minimum flow in
February–March is characteristic of this kind of river. The values of minimum monthly modulus
of flow Mminmon are the highest (4–13 L/(s·km2)). The maximum recorded amplitude of the level
variation ∆Hmax (2~3 m), corresponding to the difference between the values of annual daily
maximum Hmax and minimum Hmin of water level, is relatively low in the high-altitude territory of
the Terek River basin.
Fig. 3 Dependency of minimum month discharge divided by mean annual discharge on the mean
altitude of the river basin.
Fig. 4 The correlation between Mmeanann and the mean elevation of the river basin drawn for different
parts of the Terek River basin.
Ekaterina Rets & Maria Kireeva
4
Fig. 5 Runoff hydrograph (1936): (a) Terek River – St. Kotliarevskaya; (b) Chegem River – s.Nizhniy
Chagem; (c) Sounzha River – pgt. Karabulak
Debris flow is among the most widespread and destructive hydrological processes in the high-
altitude territory of the Terek River basin. Mudflows are observed from May to September. The
marks of debris flow activity are discovered throughout the mountainous area of the basin (The
Map of Glacial and Debris Flow Hazard, 2007). Debris cover is situated on the elevation of more
then 2000 m and is universally not covered with turf; its thickness may reach tens of metres. This
provides a sufficient amount of material for debris flow formation.
The big and medium rivers of foothills and plains with the drainage area from 3000 to
38 000 km2, which arise on the high altitude and are fed by snow and ice melting in nival belt,
inherit the main features of water regime in their upstream parts. But due to a decrease in the
altitude of drainage basins to 1200–2000 m, it is transformed to some degree. The mean modulus
of flow decreases to 5–20 L/(s·km2) (Fig. 4). Although minimum monthly modulus of flow is also
lower (Mminmon = 3–10 L/(s·km2)), the ratio of minimum monthly discharge to the mean annual
discharge increases up to 0.4–0.5 (Fig. 3). This fact suggests the growth of baseflow share in the
alimentation of rivers. The rise of water level starts earlier, the date of maximum level
establishment shifts to earlier dates (Fig. 2(a)), as well as the average of the distribution of the
number of the month when the minimum flow is recorded (Fig. 2(b)). The shape of the hydrograph
becomes more uneven due to an increase in rainfall (Fig. 5(a)). The maximum recorded amplitude
of level variation ∆Hmax shows a substantial rise up to 3–5 m.
The highest value of ∆Hmax (>5, up to 7.9 m) is recorded in rivers with comparatively low
mean altitude of drainage basin (<1200 m). The contribution of nival belts to the alimentation of
these rivers is very small or absent. The result is a decrease in the mean annual modulus of flow to
about 10 L/(s·km2) (Fig. 4) and the formation of low-low period both in winter and summer–
autumn seasons. The water-abundant period begins in March–April, while water level drop starts
in June–July with the end of mass snow melting in the drainage area. The sharp rises of river
discharge, timed to rainfalls, alternate with the drops to the low flow level (Fig. 5(c)).
The extreme irregularity in the hydrological regime creates prerequisites for the rise of such
opposite hazardous hydrological processes as floods and extremely low runoff in the lower part of
the Terek River basin.
Floods in the rivers of the Terek basin occur during the high-water period. Outstanding and
catastrophic floods are always associated with the maximum height of the spring flood, accompanied
by significant or extreme rain-flood. Such floods occur in the Terek basin approximately every 5
years (Taratunin, 2000; Dobrovolsky & Istomina, 2006). In the Terek basin, over 200 000 hectares of
Hazardous hydrological processes in mountainous areas under the impact of recent climate change
5
agricultural fields, populated areas with a population of 140 000 are in the zone of possible
inundation. A dike system was constructed on the Terek River to control floods. Breaking of earth
dikes takes place every year, causing catastrophes every 11–14 years.
Fig. 6 Dependence of ∆Hmax on the catchment area for the lowland observation stations (1), medium
mountain (Fb from 700 to 3200 km2) (2) and small mountain (Fb<1100 km2) (3) rivers.
In the low-water period, a lack of water resources is possible. A sharp increase in the
population and the density of industrial facilities is connected here with an increase in water
consumption from surface and underground sources. The dramatic expansion of the irrigated land
area in the flat part of the basin suggests the greater dependence of water-economic complex on
the availability of necessary water resources during limiting seasons of the year. At the same time,
the volumes of water withdrawal quickly increase in the foothill zone, producing an adverse effect
on river water quality.
CLIMATE CHANGE AND THE TEMPORAL DISTRIBUTION OF MAIN
CHARACTERISTICS OF HAZARDOUS HYDROLOGICAL PROCESSES
The investigation of recent climate changes in the northeastern Caucasus was based on the analysis
of temporal variations in climate characteristics measured at 34 meteorological stations, located in
different elevation belts in the study area. Although the overall picture is quite heterogeneous, it is
possible to point out some common tendencies:
– an increase in total annual precipitation in the central and western part of the territory and a
decrease in its driest, eastern part;
– an increase in the snow-to-rain ratio (Psnowann/Prainann) in most sites;
– a decrease in the number of days with negative temperatures in the year, which is also quite
common;
– an increase in the mean summer temperature is characteristic of a number of sites;
– all the enumerated tendencies are more clear-cut in the foothills and plains;
– a decrease in mean winter temperature in the mountain territory.
The degradation of glaciation caused by climate change started at the beginning of the
regressive stage of the Little Ice Age (late 18th–early 19th century). During 1970–2000, the area of
glaciation dropped by 12.6%, the volume of ice by 14.9%, the number of glaciers increased by 2.4%,
and the length of glaciers dropped by 100 m on average (Voitkovskiy & Volodicheva, 2004).
The naturally renewed water resources of the region tended to increase during the last 50
years. The main cause of this can be an increase in precipitation and the active degradation of
Ekaterina Rets & Maria Kireeva
6
glaciers. The increase of debris flow activity in the northern Caucasus started in the beginning of
the regressive stage of the Little Ice Age. The scales of glacial debris flow disasters reached the
level of extreme maximum over the last millennium in the 20th and 21st centuries. Seynova (2008)
substantiates a hypothesis of the tendency toward an increase in the debris flow hazard due to the
extremely unstable condition of the periglacial zone in the first half of the 21st century. The
regressive changes in the snow line and glaciation increase the part of high-altitude zones where
huge volumes of sediments (the material for mudflows) “are prepared”.
According to Lourje (2002) more often repetition of extreme hydrometeorological phenomena
is a negative consequence of climate fluctuations. The formation of catastrophic flooding is
connected with the formation of significant rain high waters. Khristoforov et al. (2007) concluded
that the likely climate changes have no effect on the height and shape of individual flood peaks,
but influence their average number and distribution during the year. The average number of peaks
increases during the cold season and decreases during the warm season with increasing air
temperature. The increase in annual precipitation for all examined rivers will lead to a general
increase in the number of floods in all seasons. Calculations show that the danger of occurrence of
catastrophic floods in the rivers of the northern Caucasus will increase in nearest decades with
growing air temperature.
The analysis of temporal variability of the maximum water levels showed that in the 20th
century the increase in the maximum water level took place almost at all observation stations of the
Terek River basin (Fig. 7). It is associated both with changes in climatic characteristics and the
peculiarities of the sediment load and accumulation processes, resulting in a gradual rise of the level
(at Q = const) as was the case with Kotlyarevskaya gauging station on the Terek River (Fig. 7).
Fig. 7 The maximum water levels in 1966–2002 for (a) the Terek River, station Kotlyarevskaya and
(b) the Baksan River (village of Zayukovo).
The minimum monthly discharge series are mostly homogeneous and do not show any trend.
On some gauging stations in the Terek River basin, a high value of autocorrelation was recorded.
For one-year time lag, the coefficient of autocorrelation r1 can amount to 0.85. Due to the high
value of r1, the negative bias of standard estimation of coefficient of minimum monthly discharge
variation Cvminmon and r1 can amount to 20% or even higher, depending on the value of the
corresponding statistical significance.
MAIN APPROACHES TO CALCULATING AND FORECASTING THE
CHARACTERISTICS OF HAZARDOUS HYDROLOGICAL PROCESSES IN
MOUNTAIN AREAS
Methods unexacting to input data
Hydrological analogy and hydro-geographical generalization can be named as examples of
unexacting methods for calculating the characteristics of the hazardous hydrological processes in
Hazardous hydrological processes in mountainous areas under the impact of recent climate change
7
mountain areas. The result of the implementation of these methods is a map or a dependency of the
sought characteristic upon some predictor, whose value can be gained for the analysed river basin
(Figs 2–4, 6).
One of the least fastidious in the input data method for flood forecasting in mountainous areas
is the method of corresponding levels. It makes it possible to describe the movement and
transformation of flood waves in a channel network on the basis of hydrometric observational data.
Although it is rather simplistic, it allows one to give satisfactory forecasts for different locations in
the Terek River basin with a lead time of 1–3 days (Khristoforov, 2007).
The most commonly used methods of forecasting the river discharge during the low-flow
period are based on the method of tendency. Usually the dependence of the nth month’s flow on
the n–kth month’s flow is drawn. It was estimated that under the conditions of the Terek River
basin, this method gives satisfactory forecasts of minimum month discharge, with the lead time of
up to 3 months.
The major approaches to the debris flow forecasting are:
– Geomorphologic forecasts based on the investigation of the hotbeds of erosion. On the
definite state of erosion of mountain slopes, the debris flow hazard arises (Gavardashvili,
2008).
– Background forecasting of debris flow hazard, based on the analysis of antecedent
meteorological condition (Adzhiev et al., 2008).
Physically-based mathematical modelling of snow and ice melting in the nival zone
A physically-based mathematical model of snow and ice melting in the nival zone, corresponding
to the up-to-date measurement facilities, has been developed in the course of our investigation
(Fig. 8). Ice/snow melting calculation is based on the surface heat balance equation:
Lm
Q
lTz
E
a
EA
df
S
b
S
L
hm
+ω±ω±−+−+
=
ω
=
)1)((
Fig. 8 Schematic representation of the developed model. Black arrows show characteristics, calculated
by means of given parameters. Abbreviations are the same as those used in the text, except that Ai is the
surface area of the ith zone, km2, H i is the mean elevation of the ith zone, m, εм is the debris emissivity.
Ekaterina Rets & Maria Kireeva
8
where hm is the melting, mm; L is the latent heat of fusion, J/kg; ω is the net energy flux, W/m2; Sb
is the direct beam radiation flux, W/m2; Sdf is diffuse solar radiation flux, W/m2; A is surface
albedo; Ez is outgoing long-wave radiation, W/m2; Eа is the counter radiation of the atmosphere,
W/m2; ωT is the sensible heat flux density, W/m2; ωl is the latent heat flux density, W/m2; Qm is
the heat flux through the debris, W/m2.
The input data is represented by the results of complex meteorological observations, including
global short-wave radiation, Sg, incoming long-wave radiation, Ea, wind speed, U, arranged at a
reference site within the study area, and air temperature vs altitude profile, T(H). The area under
study is subdivided into zones differing by heat balance structure.
To divide Sg into Sb and Sdf components, a dependence Sdf / Sg = f (Sdf / Sg) =f (k) is used, where
S0 is short-wave extraterrestrial radiation, k is atmosphere transparency index (Boland & Ridley,
2008). In the distribution of the direct component Sb, it is corrected according to:
– its angle of incidence in each zone, which is calculated from the values of the angle of the sun
stand above the horizon, y, the azimuth angle of the sun Az, the mean slope of the ith zone
surface Ii , the mean aspect of the ith zone Exi;
– the horizon obstruction in each zone (αij is the mean angle of horizon screening by
surrounding landforms, in each j 2° sector of horizon, for the ith zone, degrees).
Diffuse solar radiation Sdf in each zone is composed from the diffuse short-wave radiation and
irradiance due to reflection by surrounding terrain. The albedo A of ith zone surface is derived
from snow, ice, firn, and debris. The albedo values, As, Ai, Af, Ad, which were assumed to be
constant, according to Asti – surface area fraction of the ith zone, devoid of the seasonal snow and
dli, dmi, dfi is the surface area fraction of the ith zone, free from the seasonal snow and represented
by ice, debris cover, and firn, respectively. The outgoing long-wave radiation, Ez, depends on the
surface temperature. As the surface temperature of snow and ice cover is mostly around zero, Ez
equals 316 W/m2 most of the time and can only experience small variations. An empirical
dependence on air temperature is used to distribute its value. Counter radiation of the atmosphere
Ea is accepted to be equal for the study area, because it is supposed to be of necessary size to
assume that atmosphere features are even over it. Sensible heat flux density ωT is calculated using
simple bulk aerodynamic formula derived by Kuzmin (Vazhnov, 1966). The latent heat flux
density ωl is neglected in the developed model, because its share in the net energy flux is rarely
more than 1–2% (Vazhnov, 1966). The ice, covered with debris, is melting under the influence of
the heat flux conducted through the debris by means of molecular thermal conductivity Qm:
Qm = λm
dz
dT
≈
λm
m
icemh
Т
T−
where λm is the thermal conductivity of the debris cover in the ith zone, W/(m·K); hm is the debris
cover thickness in the ith zone, m; Тice is the temperature of ice, underlying debris, which is
273.16 K (0°C); Tm is debris surface temperature, K, calculated according to its heat balance.
A computation program was developed for calculation. The model has been applied to the
Dzhankuat glacier system, which is situated in the northern Caucasus (Fig. 1). The comparison of
the calculated and measured melting depth showed that the developed model gives correct results
for all parts of the glacier (Fig. 9).
The created model also permits the simulation of changes in the glacier-derived liquid runoff
responding to anticipated climate change and progressive deglaciation. If the counter radiation of
the atmosphere increases by 1.5 W/m2 (which corresponds to the changes that have occurred over
the last 100 years), the mean air temperature increases by 4º C(according to the MGEIK forecast),
the debris cover increases by 180% (equal to the changes, occurred from 1968 to 1999), the
atmosphere transparency index decreases by 5% and the Dzhankuat glacier melting rate will
increase by 12–40%, depending on the elevation-slope zone of the glacier. Thus, as the area of
glacier is expected to grow, the increase in glacier melting volume will not be dramatic. If the area
decreases by 30%, which corresponds to the changes that occurred from 1910 to 1999, the increase
in the volume of Dzhankuat glacier melting will be 8%.
Hazardous hydrological processes in mountainous areas under the impact of recent climate change
9
Fig. 9 Comparison of the calculated and measured melting rate: (а) Zone 1 (2740–2870 m a.s.l.),
(b) Zone 4 (2850–3010 m a.s.l.) of Dzhankuat glacier.
In our further research, the presented model of snow and ice melting is supposed to be the
base for the model of flow formation processes in the mountain areas. The author plan it to be
maximally non-demanding for the input data, so that approximate results could be received, even
if the observation upon a number of hydrometeorological characteristics is lacking.
CONCLUSION
Contrasting relief and climate create the prerequisites for the development of various hazardous
natural processes in the Terek River basin, constantly endangering the safety of people. The
complex spatial distribution of the main characteristics of hazardous hydrological processes is
determined by the diversity of the river’s alimentation structure and morphometrical characteristics
of their basins. The increase in the frequency of extreme hydrometeorological phenomena is a
negative consequence of the climate changes. To provide hydrological security for a territory, a
means of calculating and forecasting the main characteristics of the hazardous hydrological
processes should be developed. Both need to be unexacting to input data methods and methods,
based on mathematical modelling. In the course of investigation, the main approaches of
calculation and forecast of the main characteristics of dangerous hydrological processes in
mountainous areas using the less fastidious in the input data methods were examined. Physically-
based mathematical model of ice/snow melting in the alpine zone, aimed at the calculation of the
run-off in mountain rivers, is introduced. It conforms to the up-to-day measurement facilities of
modern weather stations. Ice/snow melting calculation is based on surface heat balance equation.
The input data are represented by the results of complex meteorological observations, arranged at a
reference site within the study area, and air temperature vs altitude profile. Application of the
proposed model for the highly glaciated Djankuat River catchment area reveals a good
reproduction of directly measured melting values for all parts of the basin.
REFERENCES
Adzhiev, A. H., Kondratieva, N. V., Kumukova, O. A., Seynova, I. B. & Bogachenko, E.M. (2008) Background forecasting for
debris flow hazard in central Caucasus: method and application results. In: Debris Flows: Disasters, Risk, Forecast,
Protection (ed. by S. S. Chernomorets), 263–266. Pyatigorsk, Russia.
Boland, J. & Ridley, B. (2008) Models of diffuse solar fraction. In: Modeling Solar Radiation at the Earth’s Surface. (ed. by
V. Badescu), Springer-Verlag Berlin Heidelberg, 193–219.
Dobrovolskiy S. & Istomina M., (2006) Floodings of the World. GEOS, Moscow, Russia.
Gavardashvili, G. V. (2008) Forecast of erosion processes in Duruji River Basin. In: Debris Flows: Disasters, Risk, Forecast,
Protection (ed. by S. S. Chernomorets), 270-273. Pyatigorsk, Russia.
Khristoforov, A. V., Yumina, N. M., Kirillov, A. V. & Rets, E. P. (2007), Runoff forecasting for Terek’s Basin, Water Industry
of Russia 4, 25–37.
(a)
(b)
Ekaterina Rets & Maria Kireeva
10
Lurje, P. M. (2002), Water Resources and Water Balance of the Caucasus. Gidrometeoizdat, Saint-Petersburg, Russia.
Map of Glacial and Debris Flow Hazard in the Central Part of North Caucasus (2007) Composed by A. N. Kapitanov (ed. by S.
S. Chernomorets & O. V. Tutubalina,), Moscow, Russia (manuscript).
Seynova, I. B. (2008) Climatic and glaciological conditions of debris flow formation in the Central Caucasus at a stage of
regress of the little ice age. In: Debris Flows: Disasters, Risk, Forecast, Protection (ed. b y S. S. Chernomorets), 121–124.
Pyatigorsk, Russia.
Taratunin, A. A. (2000) Flooding in Russian Federation. Rosniivh, Ekaterinburg, Russia.
Vazhnov, A. N. (1966) Analysis and Forecast of the Caucasus River Flow. Gidrometeoizdat, Moscow, Russia.
Voitkovskiy, K. F., Volodicheva, N. A., (2004) Evolution of elbrus glacial system. Geography, Society and Environment 1,
377–395.
