Role of Glaciers and Snow Cover in Rivers Discharge and LCLU Changes over Central Asia and South...
1
Role of Glaciers and Snow Cover in Rivers Discharge and LCLU Changes over Central Asia and South Siberia Vladimir Aizen, Elena Aizen, Arzhan Surazakov (University of Idaho, U.S.A., [email protected]) Roland Geerken (Yale University, U.S.A., [email protected]) Stanislav Nikitin (Tomsk State University, Russia, [email protected]) Genadiy Nosenko (Institute of Geography, RAS, Russia, [email protected] ) Alexander Finaev (IEH, TAS, Tajikistan, [email protected] ) Valeriy Kuzmichenok (IWRH, KAS, Kyrgyzstan, [email protected] ) Introduction: The central Asian High Mountain System (Siberian Altai, Tien Shan, Pamir) is a Water Tower for about 100 millions of people living in this region. The central Asia arid and semi- arid area covered of 12 million km 2 with mid- latitudinal extend over 5,000 km and comprise approximately of 3,133km 3 of fresh water. The central Asia arid and semi- arid area is one and half times larger than whole territory of the United States. Central Asia is a natural barrier for air masses penetrating and transformed over the mountain ridges and deserts and therefore has great influence on global and regional atmospheric processes and water cycle. Seasonal snow cover and glaciers of central Asia are supplying water and generate river flow vital not only for the millions of people living downstream, but also for forestry, agriculture, industry, and urban areas in adjacent lowlands. The human activity has been identified as contributing to land degradation. Since the middle of 19th century glaciers of central Asia lost 25% of their total area and rate of the glacier recession accelerated since 1970th. Another important source of water is the seasonal snow that contributes up to 80% of water in total river runoff in non-glacial basins, but from the 1960th to the present, the maximum snow depth and snow duration decreased. Apparently, changes in water regime may destroy stability of central Asia hydrological systems affecting on land degradation, regional agriculture, and further social and economical sustainability in central Asia and north-western China. THE GOAL: Estimate changes in the half- century of snow- and glacial-covered areas, and glaciers’ volume, simulate and predict the dynamics and feedbacks of a half- century of changes in seasonal snow/glaciers/lakes water resources and their effects on land degradation in arid and semi- arid river basins of central Asia. RS data: Aerial photos (1943-1990); DEMs resulted from SRTM, 2000; Corona KH-9 (1972-1973), Landsat TM (1982-present), ETM+ (1999-present), ASTER (1999- present); ALOS/PRISM NOAA AVHRR (1987-1999), MODIS 8-day snow product (2000-prsent), (MOD12Q1) to get a suite of Land Cover classifications, MODIS Collection 5, MODIS NDVI. 4.2% 11% 11.1% 16.3% 19.2% 17.1% 5.2% Pamir Tien Shan Altai Akshiirak glacial surface change between 1977 and 2000 evaluated by aerial photogrammetry and SRTM data Glacier area changes 1973- 2007 DEM generation: Alpine relief introduces strong geometric distortions. Some stipulation has been applied to accurate DEM development in the process of orthorectification. An unedited version of SRTM 3 arc-second data acquired in the C-band frequency. In areas of rugged terrain, the SRTM data often has void areas due to InSAR technological limitations in mountainous areas. The voids were filled with ASTER DEM generated in the “Leica” Photogrammetry Suite software package (LPS) using ASTER bands 3N and 3B. “Corona” image processing. Corona KH-9 images were acquired by Hexagon mapping camera in 1970s and declassified by US government in 2002 (although the internal camera parameters remained classified). The “frame geometry” camera was designed specifically for mapping and geodetic applications. KH-9 image “footprint” is about 120 km by 240 km and the nominal spatial resolution is 9 m. 21 KH-9 images (1970 th ) were scanned at 14 micron resolution (1800 dpi) at USGS EROS data center. Each photograph has a rectangular grid of 1081 calibration marks exposed every 1 cm. Using the calibration grid we estimated film distortion to achieve 6.5 – 18.5 micron (1 st. dev.). Aerial triangulation and orthorectification of the images were accomplished in LPS using Ground Control Points (GCP) collected from 1:25,000 and 1:50,000 maps. The triangulation RMS residuals were 6 – 12 m. The 1973 glacier boundaries were manually digitized. ASTER images processing: The ASTER images were orthorectified in LPS using the GCPs from maps and the orthorectified Corona images. The triangulation RMS residuals were 8 – 18 m. Due to high reflectance difference of snow and ice at visible and short-wave infrared spectrum, glaciers are spectrally well separable from nonglacial terrain. We used segmentation of ratio images (ASTER band 3N/ ASTER band 4) and additional NDVI masking to derive glacier boundaries. Modis data processing. To assess the impacts of cloud coverage and to estimate the date of maximum snow depth, we first performed an analysis using fractional snow data from MODIS (Collection 5) and MODSCAG product. The MODIS 8- day product was covered area above the 1000m contour encompassing the entire Tien Shan. This analysis was performed for the winter/late spring, capturing the peak and decay of snow cover. For the large area contained within the 1000 m contour the fraction of pixels that are cloud-obscured for every day in an 8-day period is small, usually less than 5% but occasionally, less than one 8-day period every other year, the value may be 20% or greater. In winter clouds are occur over snow, to avoid misidentification of snow as cloud, we added the cloud fraction to the snow fraction to give a probable snow maximum extent. When MODIS Collection 5 completed, we made comparison with both Terra and Aqua. The result is generally confirms that Aqua sees more cloud and less snow than Terra. AVHRR data processing. To establish the longest possible time series of snow-cover area for the Tien Shan we relied on the NOAA/NASA polar-orbiting satellites optical sensors. The AVHRR LAC data used with a fractional snow cover algorithm (AVTREE) developed by Southwest Regional Earth Science Applications Center (SW-RESAC). The SAF NWC cloud cover masking algorithms was developed and validated with standard meteorological and snow cover records to compliment the AVTRRE algorithm, which produced snow covered area maps from images that include cloud cover. Days of Snow in 2006-07 Days of Snow in 2000-01 0 20 40 60 80 Tien Shan snow covered area (%), by AVHRR and MODIS data using AVTREE and SAF NWC algorithm 1987 2000 2001 2002 2003 2004 2005 2006 2007 1999 1998 1997 1996 1995 1994 1993 1992 1991 1990 1989 1988 Surface data: Air temperature, precipitation, snow thickness, dates of snow appearance and disappearance, number of days with snow cover, date of max snow thickness, snow water equivalent, and river runoff discharge. High resolution topographic maps, data of geodetic, photogrammetry, GPS bounded glacier radio-echo sounding surveys. -3.0 -4.0 -5.0 -6.0 -7.0 1940 1960 1980 2000 300 400 500 600 700 800 1940 1960 1980 2000 2000-3000 m a.s.l Liner trend = 0.01 o C year -1 Ann air temperature, o C Ann precipitation, mm Liner trend = 3.8 mm year -1 2000-3000 m a.s.l Altai Altai Results: Changes in air temperature and moisture income may influence on the glacier and seasonal covered areas differently in different mountain regions at macro- and meso- scale and even at a scale of small catchments. In central Asia increase of air temperature accompanied by increase of precipitation, notably in summer and autumn, and at high elevations over 3 000 m. In Altai, at the northern periphery of the central Asia mountain system the glacier area shrank by 7.2% in average for the period from 1952 to 2006. Altai mountains also define the southern periphery of the Asian Arctic Basin, and the Ob and Yenisei rivers are only Siberian rivers that fed from Altai glaciers. In the last 54 years the annual precipitation in Altai alpine areas increased by 3.2 mm yr -1 , notably in spring months, while at adjusted low lands the precipitation trend is insignificant. Similar tendency observed in Pamir where annual precipitation increased by 18% (8.1 mm yr -1 ) on average in the last 17 years, notably in summer and autumn, and at high elevations of north-western and central Pamir over 3 000 m. However, the main cause of the continue glacier recession in Altai and Pamir is the increase in summer air temperatures by 0.03 o C yr -1 , which intensified snow/glacier ice melt and increased 7% discharge to the Ob and Yenisei rivers and 13.5% in Pamir river heads . Snow dominates the montane hydrology, accounting for 50-70% of the total precipitation and providing 60% of the total river runoff . The annual runoff of the major Tien Shan rivers is on average 67 km 3 yr -1 , which includes glacial melt of about 14 km 3 yr -1 (20%). During droughts, the proportion of glacial runoff increases to 30% of the total as a result of decrease in precipitation and increase in glacier melt. In some Tien Shan river basins, the proportion of glacial runoff can be as high as 40% of total runoff . Unlikely Pamir and Altai, precipitation in Tien Shan increased mainly at the western and northern part of the mountain system while air temperature increased at elevations, below 3 000 m (0.02 o C yr -1 ) and particularly since the middle of 1970 th . Evaluation of the Tien Shan glacier changes during last thirty years has revealed the glacier area reduction on 1,617 km 2 (-10.1%) . The rate of glacier recession between 1973 and 2003 increased by three times comparing to rate estimated for 1943-1973 period. It was found that during last twenty years, the duration of snow melts from the date of maximum snow cover to date of its disappearance reduced by 30 days and in 2007 was equal to 138 days in Tien Shan. The seasonal snow covered area in Tien Shan decreased by 15% (approximately 120 000 km 2 ), and there is a tendency for later dates of the maximum snow cover. The decrease of seasonal snow cover is not a linear process. The further decrease may be accelerated due to increase of rainfall instead of snowfall in early spring months at high elevations, and consequently a lesser heat expenditure for the snowmelt, which initiate early floods and peak of the rivers discharge. Mathematical simulation of the current glacier state and forecast the potential impact of global and regional climate change on the glaciers and glacier river runoff in the Tien Shan estimated that an increase in air temperature of 1 o C at Equilibrium Line Altitude (ELA) must be compensated by a 100 mm increase in precipitation at the same altitude to maintain glaciers in their current state. An increase in mean air temperature of 4 o C and precipitation by 1.1 times of the current level that is predicted for the 21st century may uplift ELA by 570 m during the 21century. The number of glaciers, glacier covered area, glacier volume, and glacier runoff are predicted to be 94%, 69%, 75%, and 75% of current values. The maximum glacier runoff may reach as much as 1.25 -150 -50 50 150 1940 1960 1980 2000 -0.8 0 0.8 1.6 1940 1960 1980 2000 Ann air temperature, o C Ann precipitation, mm 1960 1980 2000 -1.0 -0.5 0 1.0 1.5 0.5 1940 0 500 1000 1940 1960 1980 2000 Ann air temperature, o C Ann precipitation, mm Liner trend = 0.02 o C year -1 Tien Shan Tien Shan 2500-3500 m a.s.l 2500-3500 m a.s.l Pamir Pamir 3000-4000 m a.s.l 3000-4000 m a.s.l Liner trend = 0.01 o C year -1 Liner trend = 8.1 mm year -1 Tien Shan
-
Upload
shavonne-phelps -
Category
Documents
-
view
215 -
download
0
Transcript of Role of Glaciers and Snow Cover in Rivers Discharge and LCLU Changes over Central Asia and South...
Role of Glaciers and Snow Cover in Rivers Discharge and LCLU Changes over Central Asia and South Siberia
Vladimir Aizen, Elena Aizen, Arzhan Surazakov (University of Idaho, U.S.A., [email protected])
Roland Geerken (Yale University, U.S.A., [email protected])
Stanislav Nikitin (Tomsk State University, Russia, [email protected])
Genadiy Nosenko (Institute of Geography, RAS, Russia, [email protected] )
Alexander Finaev (IEH, TAS, Tajikistan, [email protected] )
Valeriy Kuzmichenok (IWRH, KAS, Kyrgyzstan, [email protected] )
Introduction: The central Asian High Mountain System (Siberian Altai, Tien Shan, Pamir) is a Water Tower for about 100 millions of people living in this region. The central Asia arid and semi- arid area covered of 12 million km2 with mid- latitudinal extend over 5,000 km and comprise approximately of 3,133km3 of fresh water. The central Asia arid and semi- arid area is one and half times larger than whole territory of the United States. Central Asia is a natural barrier for air masses penetrating and transformed over the mountain ridges and deserts and therefore has great influence on global and regional atmospheric processes and water cycle. Seasonal snow cover and glaciers of central Asia are supplying water and generate river flow vital not only for the millions of people living downstream, but also for forestry, agriculture, industry, and urban areas in adjacent lowlands. The human activity has been identified as contributing to land degradation. Since the middle of 19th century glaciers of central Asia lost 25% of their total area and rate of the glacier recession accelerated since 1970th. Another important source of water is the seasonal snow that contributes up to 80% of water in total river runoff in non-glacial basins, but from the 1960th to the present, the maximum snow depth and snow duration decreased. Apparently, changes in water regime may destroy stability of central Asia hydrological systems affecting on land degradation, regional agriculture, and further social and economical sustainability in central Asia and north-western China.
THE GOAL: Estimate changes in the half-century of snow- and glacial-covered areas, and glaciers’ volume, simulate and predict the dynamics and feedbacks of a half-century of changes in seasonal snow/glaciers/lakes water resources and their effects on land degradation in arid and semi- arid river basins of central Asia.
RS data: Aerial photos (1943-1990); DEMs resulted from SRTM, 2000; Corona KH-9 (1972-1973), Landsat TM (1982-present), ETM+ (1999-present), ASTER (1999-present); ALOS/PRISM NOAA AVHRR (1987-1999), MODIS 8-day snow product (2000-prsent), (MOD12Q1) to get a suite of Land Cover classifications, MODIS Collection 5, MODIS NDVI.
4.2%
11% 11.1%
16.3%
19.2%
17.1%
5.2%
Pamir
Tien ShanAltai
Akshiirak glacial surface change between 1977 and 2000 evaluated by aerial photogrammetry and SRTM data
Akshiirak glacial surface change between 1977 and 2000 evaluated by aerial photogrammetry and SRTM data
Glacier area changes 1973-2007
DEM generation: Alpine relief introduces strong geometric distortions. Some stipulation has been applied to accurate DEM development in the process of orthorectification. An unedited version of SRTM 3 arc-second data acquired in the C-band frequency. In areas of rugged terrain, the SRTM data often has void areas due to InSAR technological limitations in mountainous areas. The voids were filled with ASTER DEM generated in the “Leica” Photogrammetry Suite software package (LPS) using ASTER bands 3N and 3B. “Corona” image processing. Corona KH-9 images were acquired by Hexagon mapping camera in 1970s and declassified by US government in 2002 (although the internal camera parameters remained classified). The “frame geometry” camera was designed specifically for mapping and geodetic applications. KH-9 image “footprint” is about 120 km by 240 km and the nominal spatial resolution is 9 m. 21 KH-9 images (1970th) were scanned at 14 micron resolution (1800 dpi) at USGS EROS data center. Each photograph has a rectangular grid of 1081 calibration marks exposed every 1 cm. Using the calibration grid we estimated film distortion to achieve 6.5 – 18.5 micron (1 st. dev.). Aerial triangulation and orthorectification of the images were accomplished in LPS using Ground Control Points (GCP) collected from 1:25,000 and 1:50,000 maps. The triangulation RMS residuals were 6 – 12 m. The 1973 glacier boundaries were manually digitized.
ASTER images processing: The ASTER images were orthorectified in LPS using the GCPs from maps and the orthorectified Corona images. The triangulation RMS residuals were 8 – 18 m. Due to high reflectance difference of snow and ice at visible and short-wave infrared spectrum, glaciers are spectrally well separable from nonglacial terrain. We used segmentation of ratio images (ASTER band 3N/ ASTER band 4) and additional NDVI masking to derive glacier boundaries.
Modis data processing. To assess the impacts of cloud coverage and to estimate the date of maximum snow depth, we first performed an analysis using fractional snow data from MODIS (Collection 5) and MODSCAG product. The MODIS 8-day product was covered area above the 1000m contour encompassing the entire Tien Shan. This analysis was performed for the winter/late spring, capturing the peak and decay of snow cover. For the large area contained within the 1000 m contour the fraction of pixels that are cloud-obscured for every day in an 8-day period is small, usually less than 5% but occasionally, less than one 8-day period every other year, the value may be 20% or greater. In winter clouds are occur over snow, to avoid misidentification of snow as cloud, we added the cloud fraction to the snow fraction to give a probable snow maximum extent. When MODIS Collection 5 completed, we made comparison with both Terra and Aqua. The result is generally confirms that Aqua sees more cloud and less snow than Terra.
AVHRR data processing. To establish the longest possible time series of snow-cover area for the Tien Shan we relied on the NOAA/NASA polar-orbiting satellites optical sensors. The AVHRR LAC data used with a fractional snow cover algorithm (AVTREE) developed by Southwest Regional Earth Science Applications Center (SW-RESAC). The SAF NWC cloud cover masking algorithms was developed and validated with standard meteorological and snow cover records to compliment the AVTRRE algorithm, which produced snow covered area maps from images that include cloud cover.
Days of Snow in 2006-07
Days of Snow in 2000-01
0
20
40
60
80
Tien Shan snow covered area (%), by AVHRR and MODIS data using AVTREE and SAF NWC algorithms
1987 2000 2001 2002 2003 2004 2005 2006 2007199919981997199619951994199319921991199019891988
Surface data: Air temperature, precipitation, snow thickness, dates of snow appearance and disappearance, number of days with snow cover, date of max snow thickness, snow water equivalent, and river runoff discharge.
High resolution topographic maps, data of geodetic, photogrammetry, GPS bounded glacier radio-echo sounding surveys.
-3.0
-4.0
-5.0
-6.0
-7.01940 1960 1980 2000
300
400
500
600
700
800
1940 1960 1980 2000
2000-3000 m a.s.l
Liner trend = 0.01oC year-1
Ann
air
tem
pera
ture
, oC
Ann
pre
cipi
tatio
n, m
m
Liner trend = 3.8 mm year-1
2000-3000 m a.s.l
Altai
Altai
Results: Changes in air temperature and moisture income may influence on the glacier and seasonal covered areas differently in different mountain regions at macro- and meso- scale and even at a scale of small catchments. In central Asia increase of air temperature accompanied by increase of precipitation, notably in summer and autumn, and at high elevations over 3 000 m. In Altai, at the northern periphery of the central Asia mountain system the glacier area shrank by 7.2% in average for the period from 1952 to 2006. Altai mountains also define the southern periphery of the Asian Arctic Basin, and the Ob and Yenisei rivers are only Siberian rivers that fed from Altai glaciers. In the last 54 years the annual precipitation in Altai alpine areas increased by 3.2 mm yr -1, notably in spring months, while at adjusted low lands the precipitation trend is insignificant. Similar tendency observed in Pamir where annual precipitation increased by 18% (8.1 mm yr -1) on average in the last 17 years, notably in summer and autumn, and at high elevations of north-western and central Pamir over 3 000 m. However, the main cause of the continue glacier recession in Altai and Pamir is the increase in summer air temperatures by 0.03 oC yr-1 , which intensified snow/glacier ice melt and increased 7% discharge to the Ob and Yenisei rivers and 13.5% in Pamir river heads . Snow dominates the montane hydrology, accounting for 50-70% of the total precipitation and providing 60% of the total river runoff . The annual runoff of the major Tien Shan rivers is on average 67 km 3 yr-1, which includes glacial melt of about 14 km3 yr-1 (20%). During droughts, the proportion of glacial runoff increases to 30% of the total as a result of decrease in precipitation and increase in glacier melt. In some Tien Shan river basins, the proportion of glacial runoff can be as high as 40% of total runoff . Unlikely Pamir and Altai, precipitation in Tien Shan increased mainly at the western and northern part of the mountain system while air temperature increased at elevations, below 3 000 m (0.02 oC yr-1) and particularly since the middle of 1970th. Evaluation of the Tien Shan glacier changes during last thirty years has revealed the glacier area reduction on 1,617 km 2 (-10.1%) . The rate of glacier recession between 1973 and 2003 increased by three times comparing to rate estimated for 1943-1973 period. It was found that during last twenty years, the duration of snow melts from the date of maximum snow cover to date of its disappearance reduced by 30 days and in 2007 was equal to 138 days in Tien Shan. The seasonal snow covered area in Tien Shan decreased by 15% (approximately 120 000 km 2), and there is a tendency for later dates of the maximum snow cover. The decrease of seasonal snow cover is not a linear process. The further decrease may be accelerated due to increase of rainfall instead of snowfall in early spring months at high elevations, and consequently a lesser heat expenditure for the snowmelt, which initiate early floods and peak of the rivers discharge. Mathematical simulation of the current glacier state and forecast the potential impact of global and regional climate change on the glaciers and glacier river runoff in the Tien Shan estimated that an increase in air temperature of 1 oC at Equilibrium Line Altitude (ELA) must be compensated by a 100 mm increase in precipitation at the same altitude to maintain glaciers in their current state. An increase in mean air temperature of 4 oC and precipitation by 1.1 times of the current level that is predicted for the 21st century may uplift ELA by 570 m during the 21century. The number of glaciers, glacier covered area, glacier volume, and glacier runoff are predicted to be 94%, 69%, 75%, and 75% of current values. The maximum glacier runoff may reach as much as 1.25 times current levels while the minimum will likely equal zero . While the Tien Shan glacier area decreases continuously, at the last decade the annual river discharge is growing mainly due to precipitation increase. One of the main predictors of the current year’s river discharge in Tien Shan is the volume of river runoff during the previous year, which could be replenished by ground water. The possible sharp change in river runoff indicated the non-linear system response caused mainly by non-linearly response of evapotranspiration to air temperature and precipitation changes. Thus, a surplus of precipitation accelerates growth of evapotranspiration when the air temperature increases, while a precipitation deficit slows this process even air temperature growth, which also increase albedo of glacier surface in summer and reduce potential snow and ice melt . Current glaciers recession, while initially considered as a positive factor that increased the river flow, at the end causes the runoff to decrease.
-150
-50
50
150
1940 1960 1980 2000-0.8
0
0.8
1.6
1940 1960 1980 2000Ann
air
tem
pera
ture
, oC
Ann
pre
cipi
tatio
n, m
m
1960 1980 2000-1.0
-0.50
1.01.5
0.5
1940
0
500
1000
1940 1960 1980 2000Ann
air
tem
pera
ture
, oC
Ann
pre
cipi
tatio
n, m
m
Liner trend = 0.02oC year-1
Tien Shan Tien Shan2500-3500 m a.s.l 2500-3500 m a.s.l
Pamir Pamir3000-4000 m a.s.l
3000-4000 m a.s.l
Liner trend = 0.01oC year-1 Liner trend = 8.1 mm year-1
Tien Shan
