CLIMATE CHANGE AND WATER POTENTIAL OF THE PAMIR MOUNTAINS

The Pamir region supplies water for most countries of the Central Asia. Discussions and arguments with regard to reduction of water resources related to climate change are popular today among various governmental and international institutions being a great concern for modern society. Probable decrease of the Pamirs runoff will affect downstream countries that can face water deficiency. However, there is no scientific rationale behind such disputes. The Pamir region is a remote, high-mountainous and hard-to-access area with scarce observation network and no reliable data. It is not sufficiently investigated in order to perform any assessment of climate change. This article represents results of study of climate parameters change (such as temperature, precipitation and river discharge) in the Pamirs. The study area covers all countries included in this mountain region (Tajikistan, China, Afghanistan and Kyrgyzstan). Observation records, remote sensing data and GIS modeling were used in the present work. Chronological data series were divided into two equal time intervals and were treated as climatic periods. Further analysis of climate change helped to estimate its influence on change of water potential in the Pamirs. The paper considers issues of liquid and solid precipitation change in the study area.


INTRODUCTION
The Pamir Mountains are bordering the north-west of the Tibetan Plateau and stretch through the territories of Tajikistan, Kyrgyzstan, Afghanistan and China.High elevated mountain ranges make a great barrier to air streams coming from the west.This results in intense precipitation in the western part of the whole area and windward slopes of high ranges.The most extensive contemporary glaciers are found in the Central Pamir of Tajikistan and in the Kashgar mountains of China.The Pamirs, being a source of major rivers in Central Asia and China, is divided into two large watersheds -the Amu Darya river in the west and upper reaches of the Tarim river in the east.Global warming of climate undoubtedly impacts the snow cover, glaciers' dynamics and water potential of this area.Change of water resources has direct influences on irrigation sector in Central Asian countries and northwestern China.A comprehensive description of the Pamir glaciers was done by Schetinnikov [1998]; theoretical basis of glacial study for this area was developed by Glazyrin [1991].Estimation and forecast of glacial changes for that part of the region in case of various scenarios have been done by several scientists [Glazyrin and Finaev, 2003]; the Book of articles "Glacial areas of the Pamir-Alay" [1993]; Finaev [1999].Such estimations were generally done on the basis of records obtained from some glaciers of Tajikistan before 1980.The recent development of remote sensing and geo-information technologies (GIS) allows estimation of glacial changes on a large scale, for instance the Chinese part of the Eastern Pamirs [Liu and others, 2008] or the Western Pamirs in Tajikistan [Finaev, 2013].A thorough review made by some researches [Unger-Shayesteh and others, 2013] proved that Central Asia has experienced accelerated warming since the 1970s which resulted in shrinkage of glaciers in the Pamirs and Tieng Shan.
In the paper published by Tandong Yao and other [2012], the authors have shown that glacial areas of the Himalayas were shrinking due to change of atmospheric circulation and precipitation decrease, while in the Pamirs precipitation increase was observed from 2006 to 2010.It should be noted that Yao et al, [2012] studied only Chinese part of the Pamirs (the Kashgar ridge) which included the highest peak of Muztagh Ata.However, the greatest glacial areas here stretch to the west and can be found in the Central Pamirs of Tajikistan.
Modeling of climate change and glacial areas under the ADB project TA-7599 [Climate Resiliency for Natural Resources Investments, 2011] has shown that the Pamir glaciers were tending to reduce.There are many earlier works proving degradation of some glaciers [Schetinnikov, 1998]

STUDY REGION
Usually "the Pamirs" describes the mountains located between the Pyanj and the Vakhsh rivers (the conditional boundaries are identified along these two rivers).However, this approach is not correct from the hydrological point of view, because river basins include mountain ranges which are not part of the Pamirs by the above definition.Thus, the study area covers the entire basins of the Vakhsh, the Pyanj and the Tarim rivers (including ranges adjacent to the Pamirs).Therefore, the present study considers the area which is larger than the total area of the Pamirs typically mentioned in the scientific literature.

The topography and river systems
The study area is located at altitudes from 1100 m up to above 7700 m asl.Particular features of the Pamir highland allow dividing the whole territory into two parts -the Eastern and the Western Pamirs.
The land forms of the Western Pamirs vary and can be characterized by separate high ridges (up to 3.0-4.0km) alongside with flat river valleys affected by erosion.The main ridges of the Western Pamirs stretch from the southwest to the northeast, where the highest peak is Somoni (Communism) with an elevation of 7495 m asl.Catchment areas of the Western Pamirs belong to the Pyanj and Vakhsh river basins.
The Eastern Pamirs is a high mountain plateau located at the elevation of 3.8-4.5 km.Small mountain ridges rise over flat valleys at the height of 0.5-1.5 km, and some ranges reach 6.2-6.9 km.This high elevated region is covered with permafrost due to low temperatures.The Kashgar ridge with the highest peaks of Muztag-Ata (7546 m asl.) and Kongurt (7719 m asl.) is located eastward.In the central part of the Eastern Pamirs there is an orographic depression with a closed (drainless) basin of the Karakul lake.This territory is protected by mountains from penetration of damp air masses, thus receiving very small amount of precipitation.High mountain plateaus are characterized by wide and flat river valleys.The catchment areas of the Eastern Pamirs belong to the Pyanj and Tarim river basins (Fig. 1).Altitudes of main river basins are presented in Table 1.Despite the fact that all meteorological stations are located in mountain valleys, they are at different altitudes.Since the whole area is not big, change of temperature in case of altitude increase is more significant than its variations related to the stations' locations.Thus, the altitude gradient was used to calculate spatially distributed monthly air temperatures for the Pamir area, which is presented in the formula below: where a and b -coefficients of regression equation for each month in two climatic periods; H -elevation of the station extracted from the DEM; T -average monthly air temperature.DEM resolution was 0.833 km.
Distribution of precipitation in mountains does not depend on elevation change, but relates more to land forms and air masses movement.Thus, the method used for calculation of temperature is not applicable for precipitation modeling.Information from the World Climate Database (WCD) [Hijmans and other, 2005; www.worldclim.org] was used for calculation of precipitation fields throughout the Pamirs.Time intervals used in the present study (1927-1967 and 1970-2009) and the period represented in the WCD  did not match.This fact causes differences between precipitation grid from the WCD and the average monthly records from the meteorological stations.However, the advantage of the WCD is that it can characterize relative distribution of precipitation over a particular territory taking into account surface elevation and various climate conditions.In order to have reliable precipitation distribution throughout the Pamirs, calibration between data from the WCD and the stations records have been done for each month in both CP.Error (Er st ) between precipitation data from WCD (P wcd ) and records at each station was calculated for this purpose.
Er st = P wcd -P st , where Er st -the WCD error at the station; P wcd -precipitation at the station according to WCD; P st -actual precipitation at the station.
The field of errors (F Er ) for the entire territory has been calculated according to the obtained error results Er st .Then the WCD field has been calibrated using the error field (F Er ).
where WCD crrect -the WCD corrected field; F WCD -the WCD field; F Er -the error filed.
As a result, the corrected precipitation grids for each month in both climatic periods were calculated.Assessment of grids and volume of solid precipitation (snow) has been done for conditions when temperature was below or equal to zero (T < = 0).

Hydrological and glaciological data
Available observation records obtained from hydrological stations of Tajikistan have been used for water discharge analysis (Fig. 2).Using the abovementioned methodology, all observation records have been equalized to

RESULTS AND DISCUSSION
Analysis of changes in climate data is based on the calculated temperature and precipitation fields.

Temperature
The analysis of temperature records from meteorological stations showed that average air temperature during the second CP has increased by 0.42°С (Table 2).It proves positive temperature trend with the rate of 1.01°С/100 years.
The analysis of simulated air temperature fields in the Pamirs showed that average temperature has increased by 0.44°С (or 1.06°С/100 years) in the second CP.That value is 0.02°С higher than the values calculated using the observation records from the stations.Maximum warming occurred in autumn and winter seasons (up to 1°С), and minimum warming was observed during summertime (0.3°С).Decrease of temperature by 0.1°С was observed in March (Table 3).Induced by temperature growth, the elevation (asl) of the zero degree isotherm increased.The altitude of zero degree isotherm increased by 66 m on average in the second CP.The average elevation of the zero degree isotherm was 3220 m asl in the first CP, and 3286 m asl in the second CP.In July, the zero degree isotherm was 5090 m asl during the first CP, and 5168 m asl during the second CP.Thus, the temperature below zero is higher than this level within the whole year.

Precipitation
Atmospheric precipitation at all stations has increased by 3.2 % on average during the second CP (Table 4).Analysis of simulated precipitation fields showed that the average amount of precipitation over the territory was 372 mm/year during the first CP, and Maximum of precipitation occurs in the highest part of the Pamirs (up to 1371 mm/year), and this is the reason for big concentration of glaciers here.In the Eastern Pamirs, precipitation is less than 100 mm/year.Eastward moving air streams are obstructed by the Kashgar ridge located in China, and that induces up to 400-500 mm of precipitation per year over mountains (Fig. 3).
Balance between liquid and solid precipitation (rainfall and snow) varies within a year.In winter season, snow is observed over the entire Pamirs, while during summer it occurs only in high mountains.In the second CP such balance changed due to temperature growth, thus increasing rainfall and decreasing snowfall (Fig. 4).
In the first CP, precipitation varied from 5.6 mm/year to 1337 mm/year.In the second CP, the range of precipitation extended from 0.3 mm/year to 1371 mm/year (see Table 3).Change of precipitation throughout the study area was not similar.In the second CP, precipitation increased by 40 mm/year in some high mountain areas, and decreased by 20 mm/year over other territories (Fig. 5).
Precipitation in the glacial area in the first CP was 6.72 km 3 /year, and in the second CP, it was 6.91 km 3 /year showing an increase of 2.91 %.Precipitation over the territory without glaciers was 51.4 km 3 /year during the first CP, and 52.1 km 3 /year during the second CP (see Table 4).Maximum of snow accumulation occurred in April.In the first CP, it was 30.6 km 3 / year, and in the second CP, it decreased by 2.85 % to 29.8 km 3 /year (Fig. 6).The maximum area of snow cover occurs in January when the whole Pamir region is covered with snow. to area reduction (Fig. 6).The minimum area of snow accumulation in the same month decreased by 26.05 % during the second CP.The maximum thickness of snow cover increased by 3.85 %.However, the average annual snow accumulation volume decreased by 3.28 % in the second CP (see Table 4).For a better understanding of climate change, calculation of the seasonal precipitation changes for the three decades from 1981 to 2010 has been made.
Analysis of precipitation change over decades showed the following picture.During the period of 1981-1990, amount of precipitation was 55.9 km 3 /year.In the next period of 1991-2000, precipitation increased up to 59.8 km 3 / year.In the period of 2001-2010, precipitation decreased again to 57.7 km 3 /year.In winter season continuous increase of precipitation was observed.During spring and summer seasons precipitation increase occurred in the second decade (1991-2000) (Fig. 7).

Glacial zone
The area of open ice calculated based on Landsat images was 10382 km 2 or 7 % of the total area of the entire Pamirs (Table 5).
The greatest glacial areas can be found in the Pyanj river basin (3702.82km 2 ); however, the share of glaciers in the total basin is only 5.2 %.The biggest share of glacial areas can be found in the Vakhsh river basin, which is 12.2 % of the total catchment.Mean annual snow cover area across the Pamirs decreased by 2.5 % during the second CP.Maximum reduction of mean annual snow cover area by 2.7 % occurred in the Vakhsh river basin.

Distribution of precipitation throughout river basins
Assessment of precipitation distribution throughout river basins showed increase from 1.3 % (the Vakhsh river) to 2.4 % (the Pyanj river) in western part of the Pamirs during the second CP .In the Karakul lake basin precipitation increased by 4.1 %, and in Chinese part of the Pamir it decreased by 0.9 % (Table 6).Precipitation increased by 1.7 % or 0.97 km 3 /year across the entire Pamirs in the second CP.
The Tarim river basin.During the first CP precipitation varied from 30 mm/year to 609 mm/year in the Tarim river basin near the Kashgar ridge, and the average value was 164.9 mm/year.During the second CP precipitation varied from 30 mm/year to 613 mm/year, and the average value was 163.5 mm/year (Table 7).Thus, a small reduction of precipitation volume (-0.9 %) with increased range was observed.
Despite decrease of snow cover area in the Tarim river basin by 4.9 % during the second CP, snow volume in glacial zone of the Kashgar ridge increased by 1.8 % (see Table 6).The share of average annual precipitation in the Tarim river basin relative to the entire Pamirs was 12.5 % during the first CP, and 12.2 % during the second one.The share of snow in the glacial areas did not change and was 1.06 % (Table 8).
The Karakul lake basin.Annual precipitation in the Karakul lake basin increased by 4.1 % during the second CP, i.e. from 199,8 mm/year to 208 mm/year, or from 0.91 km 3 /year to 0.95 km 3 /year (see Table 6, Table 7).Quantity of snow throughout the entire basin increased by 2.3 %; amount of snow in glacial areas increased by 4 %.The share of annual precipitation in the Karakul lake basin relative to the entire Pamirs remained constant at 1.6 %.The share of snow in the total precipitation was 1.3 %, and the share of snow in the glacial area was 0.35 %.
The Pyanj river basin.Average precipitation in the Pyanj river basin was 430.5 mm/year during the first CP and 440.9 mm/year during the second one (see Table 7), i.e., 30.85 km 3 /year and 31.59km 3 /year, accordingly (see Table 6).Amount of snow increased by 0.3 % reaching 24.09 km 3 /year during the second CP.In the glacial area, snow volume increased by 3.9 % reaching 2.87 km 3 /year during the second CP.The share of precipitation in the Pyanj river basin in the total area was 53.1 % in the first CP, and 53.5 % in the second one (Table 8).The share of snow in the total precipitation decreased from 41.3 % to 40.8 %, and in the glacial areas it increased from 4.75 % to 4.85 % during the second CP.
The Vakhsh river basin.The most intensive precipitation was observed in the Vakhsh river basin -from 657.57mm/year in the first CP to 666.35 mm/year in the second one (see Table 7).Volume of precipitation increased from 19.05 km 3 /year during the first CP to 19.30 km 3 / year during the second one (see Table 6).The average annual amount of snow throughout the basin decreased by 1.8 % in the second CP.
During the first CP, the amount of snow was 12.9 km 3 /year, and during the second one it was 12.7 km 3 /year.In the glacial area, snow was 3.14 km 3 /year during the first CP and 3.21 km 3 /year during the second one, which is 2.2 % greater.The share of precipitation relative to the total amount for the entire Pamirs was 32.8 % in the first CP and 32.7 % in the second one; the amount of snow was 22.2 % and 21.5 % respectively.The share of snow in the glacial area of the Vakhsh river basin relative to the total precipitation for the entire Pamirs was 5.41 % in the first CP and 5.44 % in the second one.
The share of snow in the first CP throughout the entire Pamirs was 71 % (Table 8); in the second one it was 1.7 % smaller.However, in the glacial area, the amount of snow was 0.14 % greater during the second CP compared to the first one.

River discharge
As it was mentioned above, the Vakhsh river basin discharge is characterized by records obtained from the Darbant station.The area of the Vakhsh river basin upstream from this station is 29005 km 2 .The area of the Pyanj river basin, upstream of the Shidz station, is 60044 km 2 .Historical data show a slight tendency for water discharge decrease during the entire observation period.However, increase of water discharge in the Vakhsh river can be noticed since 1974 (Fig. 8).
In the second CP precipitation increased by 2,8 % in the Vakhsh river basin, and by 1,7 % in the Pyanj river basin.However, river discharge in the second CP decreased by 5,1 % in the Vakhsh river basin, and by 1,8 % in the Pyanj river basin (Fig. 9).Reduction of water discharges while precipitation grows can be explained by two factors.Firstly, rise of temperature increases evaporation.Secondly, in high mountains precipitation grows without thawing, which results in snow accumulation increase.

Circulation Index and climate fluctuation in the Pamirs
The Pamir Mountains lie between 36 and 39 degree of the northern latitude.In the present study, the authors have tested the mentioned Classification of circulation patterns and periods in the Northern hemisphere (http://atmospheric-circulation.ru) for climate change assessment in the Pamir region.Thus, the total circulation index for decadal moving average has been used together with mean annual precipitation in the Pamirs for the same decades (Fig. 10).
Approximation by a polynomial of the sixth order showed nearly synchronically change of precipitation and circulation.Such correlation illustrates impact of global circulation processes on climate change in the Pamirs.

CONCLUSION
The present work allowed estimating climate change and its impact on water resources of the Pamir region during the observation period .The findings revealed average increase of air temperature and precipitation in the second half of the whole chronological interval.In the western and northern areas of the Pamirs precipitation increased, while in the eastern part a small decrease was observed.Snow accumulation increased by 2.9 % in the upper reaches of glaciers.Despite this, winter snow stocks decreased due to reduction of the entire snow area by 2.5 %.It can be assumed that in the next period glaciers located below 3200 m asl will continue to shrink until they gain mass balance.Thawing of glaciers will slightly increase at the elevation between 3200 m and 5100 m asl.At the same time, snow will be accumulated in the areas higher than 5100 m asl.When getting a critical mass, it is likely for glaciers to start shifting downwards.Thus, the glacial area at high altitudes can be extended.Hydrological characteristics in rivers of the Pamirs varied within several percent.Therefore, no trends of discharge change were revealed in the upper parts of the Pyanj and Vakhsh river basins.
Climate change in the Pamirs can, probably, be related to global circulation processes in the atmosphere.This point of view is supported by good correlation between the atmospheric circulation index and long-term variations in precipitation.
. The catalogue of the Pamir glaciers located in the USSR was published based on aerial photography performed in 1968 [Catalogue of glaciers of the USSR, 1968].It described parameters of glaciers for 1955-1960.Information provided in the second catalogue of glaciers included the period through 1980 [Catalogue of Pamir and Hissar-Alay glaciation for 1980, 2011].These data proved reduction of glaciers compared to the information from the first catalogue.No glacial inventory and assessment were done afterwards in the Pamirs.All mentioned studies include only the Tajik side of the Pamirs, although glacial areas are distributed over the territories of China and Afghanistan as well.The present paper represents an updated review of climate change in the Pamirs based on meteorological and hydrological records, space images and modeling results with the help of GIS.Analyses of different parameters' change such as air temperature, precipitation, snow cover and snow accumulation thickness (in water equivalent) for glacial areas across the whole Pamirs and particular basins of large rivers have been performed in the present study.
one time interval from 1932 to 2009.Records from hydrological stations of other countries were not available.The Darbant hydrological station is located right on the border of the investigated area, thus it observes water discharge of the Vakhsh river basin which is located in the Pamirs.The Shidz hydrological station represents only part of the entire Pyanj river basin.It was impossible to obtain discharge records for the entire Pyanj river basin.Thus, assessment of runoff change in the Pamirs was done using data from two hydrological stations -Darbant and Shidz.Hydrological records from the drainless Karakul lake basin and the Tarim river basin are absent, because there are no observations in these areas.Change of open ice area depends on climate change and can characterize glacial fluctuations.It is necessary to have long-term space images of the entire Pamirs to perform assessment.Unfortunately, such data are not available.The present study uses Landsat images covering the whole Pamir region for different time intervals.It helps to do the mosaic of debris-free ice areas.In the Pamirs, the proportion of ice covered with debris and stone fragments (moraines) is 10 % on average [Schetinnikov, 1998].Areas without debris are clearly identified and outlined on space images.

Fig. 5 .Fig. 6 .
Fig. 5. Change of annual precipitation in the Pamirs during the second CP (mm/year)

Fig. 10 .
Fig. 10.Total circulation index and average annual precipitation in the PamirsLegend: P10 Pamir -total annual precipitation in the Pamirs per 10 years; Circulation Index = Z + Vz + Mn + Ms (Z -zonal circulation index; Vz -zonal circulation disturbance index; Mn -northern meridional circulation index; Ms -southern meridional circulation index).

elevated Fig. 1. The main river basins in the study area Finaev A.F. , Shiyin L. et al. CLIMATE CHANGE AND WATER POTENTIAL OF THE PAMIR MOUNTAINS
In the Vakhsh river basin there are 7 stations ranging from 1258 up to 4169 m asl.In the Pyanj river basin there are 11 stations; the highest one is Shaimak (3840 m asl.) and the lowest one is Darvaz (1284 m asl).
term station records[Narovlyanskiy, 1968;  Pedhazur, 1982; Drozdov and others, 1989; The international meteorological dictionary, WMO, 1992].In the present study this procedure was done by tools of multiple linear regression available in Microsoft Excel 2010.In order to improve calculation results, 16 stations outside the Pamir territory from Tajikistan, China, Afghanistan, Kyrgyzstan and Pakistan have been included in the study.Totally, records from 36 stations for temperature and records from 35 stations for precipitation have been included.The term "climate" implies average weather conditions (meteorological parameters) for the period of 30 years and longer.In the present study the total investigated interval is 83 years.In order to estimate probable climate change, the whole time interval has been divided into two climatic periods (CP): from 1927 to 1969, and from 1970The average altitude of stations is 2740 m asl, which is 1020 m lower than the average elevation of the Pamirs according to the Shuttle Radar Terrain Mission (SRTM) digital elevation model (DEM).In conditions of complicated

Table 2 . Air temperature and precipitation from meteorological stations in the Pamirs
Thickness of snow accumulation for the same two ice core sites, based on our model, was 1230 mm/year and 1260 mm/year, accordingly.Thus, the difference between real measurements and calculations is 10.9 % and 39.7 %, accordingly.There are three reasons for such variance.Firstly, different time intervals were used when comparing available data.

Table 8 . Share of basin total precipitation, solid precipitation (snow) and solid precipitation in glaciated areas in total precipitation over the Pamirs
Later the same scientists developed Classification of atmospheric circulation features in the Northern hemisphere, which became useful for identifying of 41 Elementary Circulation Patterns (ECP).The researchers proved that during a long-time interval (of several years) there are certain types of circulation which influence cyclones or anticyclones shifting latitudinally or longitudinally.Thus, variable air stream routes can significantly impact on climate change in particular regions.