GOCE Project acronym: Dl_3. Natural. the. December December Month. Actual submission

Transcription

1 Project no: Project acronym: Instrument: Thematic Priority: GOCE BRAHMATWINN Specific Targeted Research Project Global Change and Ecosystems Project title: Twinning European and South Asian River Basins to enhance e capacity and implement adaptive management approaches Deliverable Report Dl_3 Assessment Environment of the Natural Due date of deliverable Actual submission date Start date of project: Duration: December 2007 December Month Organisation name of lead contractor for this deliverable: University of Salzburg (ZGIS) [CR 4] Project homepage: Dissemination Level: PU

2 Contents Executive Summary... 8 Introduction Deliverable 3.1 SRTM based terrain classifications according to IPCC recommendations and digital terrain models (DTM) for the UDRB and UBRB Lead partner and partners involved Rationale and main outcome Methodology Input data Methods applied Achievements Map showcase Data sets produced Project relevant achievements Lessons learned Deliverable 3.2 Analysis of historical glacier and snow distribution with change detection analyses Lead partner and partners involved Rationale and main outcome Methodology Input data Methods applied Achievements Map showcase Data sets produced Project relevant achievements Lessons learnt Deliverable 3.3 Terrain based permafrost distribution and vulnerability analyses of slope stability Lead partner and partners involved Rationale and main outcome Methodology Input data Methods applied / 91

3 3.4 Achievements Map showcase Data sets produced Project relevant achievements Lessons learnt Deliverable 3.4 IPCC based classification of land us and land cover (LULC) with change detection analysis Lead partner and partners involved Rationale and main outcome Methodology Input data Methods applied Achievements Map showcase Data sets produced Project relevant achievements Lessons learnt Deliverable 3.5 Assessments of wetlands, their functions and groundwater recharge Lead partner and partners involved Introduction Assessment of wetlands and wetland functions Methodology Achievements Assesment of groundwater recharge Methodology Achievements Discharge Modelling Rationale and main outcome Methodology Achievements Lessons learned Deliverable 3.6 Hydrological systems analysis and delineation of Hydrological Response Units (HRU) Lead partner and partners involved Rationale and main outcome / 91

4 6.3 Methodology Input data Methods applied Achievements Map showcase Data sets produced Project relevant achievements Lessons learnt Deliverable 3.7 Documented results for reporting and RBIS population Lead partner and partners involved Rationale and main outcome Achievements Population of RBIS List of Figures Figure 1. Histograms of glacier cover in the test catchments Figure 2. Permafrost elevation histograms in the UBRB Figure 3. Levels of details for BRAHMATWINN LULC classifications Figure 4. Illustration: fine scaled land use classification for a subset of the Assam Brahmaputra floodplain Figure 5. Spatial indicators to be derived by the 3.4 LULC products Figure 6. Area distribution of reference catchment Lhasa river Figure 7. Area distribution of reference catchment Wang Chu Figure 8. Area distribution of reference catchment Brahmaputra in Assam Figure 9. Flow diagram of the main components of the SPATSIM version of the Pitman model Figure 10. Model performance for the Koshi basin. Coefficient of determination: R2 = Figure 11. Calibration results showing observed vs. simulated monthly flow volumes (m3) for the Koshi basin Figure 12. Baseflow analysis for the Koshi upper sub catchment Figure 13. Simulated monthly flow volumes (m3) for the Outlet point of the Lhasa basin Figure 14. Schema of HRU delineation Figure 15. Slope class distribution in the Lhasa river catchment / 91

5 List of Maps Map 1.1. Digital surface model and terrain classification of Upper Brahmaputra River Basin and Upper Danube River Basin (resolution 1 x 1 km, scale: 1: 3,000,000) Map 1.2. Digital surface model and terrain classification of UBRB reference catchments of Lhasa River, Wang Chu, and Brahmputra in Assam Map 1.3. Digital surface model and terrain classification of Upper Danube River Basin (resolution 1 x 1 km, scale: 1: 1,100,00) Map 2.1. Snow cover range April in the UBRB (scale: 1:3,000,000) Map 2.2. Snow cover mean April in the UBRB (scale: 1:3,000,000) Map 2.3. Snow cover range April in the UDRB (scale: 1:1,250,000) Map 2.4. Snow cover mean April in the UDRB (scale: 1:1,250,000) Map 3.1. Modelled permafrost distribution in the UBRB and glacier cover Map 3.2. Detail of Map 3.1 over the Himalaya main ridge in Bhutan Map 3.3. Detailed map of modelled permafrost distribution and glacier cover in the Lhasa catchment Map 3.4. Detailed map of modelled permafrost distribution and glacier cover in the Wang Chu catchment Map 3.5. Modelled permafrost distribution in the Salzach basin, UDRB, together with glacier cover (blue: glaciers of 1998, red: glacier retreat ) Map 4.1. Land Use Land Cover Map for Assam (year ), North East India (scale 1:1,000,000).. 40 Map 4.2. Land Use Land Cover Map for Lech reference catchment, Austria /Germany (scale 1:250,000) Map 4.3. Land Use Land Cover Map for Lhasa River Basin, China (scale 1:550,000) Map 4.4. Land Use Land Cover Map for Salzach River Basin, Austria (scale 1:300,000) Map 4.5. Land Use Land Cover Map for Wang Chu River Basin, Bhutan (scale 1:500,000) Map 4.6. Change analysis in a subset of the Brahmaputra in Assam catchment between 1990 and Map 4.7. Change analysis for a subset of Lhasa River Basin, China, between 1990 and Map 5.1. Ecoregions in the UBRB. Ecoregions describe similar ecosystem complexes within its biomes (Olson et al. 2001); black boundaries mark the Catchments within the UBRB Map 5.2. Global Lake and Wetland Database Layer 3 (B. Lehner & P. Döll, 2004) (top); Wetland area (left) and polygons (right) in the UBRB Map 5.3. Lhasa subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) Map 5.4. Assam subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) Map 5.5. Wang Chu subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) 54 Map 5.6. Ecoregions in the UBRB. Ecoregions describe similar ecosystem complexes within its biomes (Olson et al. 2001); black boundaries mark the Catchments within the UBRB Map 5.7. Global Lake and Wetland Database Layer 3 (B. Lehner & P. Döll, 2004) combined with the Corine Lake Cover (provided by EEA) (top); Wetland area (left) and polygons (right) in the UDRB / 91

6 Map 5.8. Lech subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) Map 5.9. Salzach subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) Map Groundwater availability in Assam. Figures are given per district in km³/yr (source CGWB, Assam); please note that large symbols indicate low availability Map Assam subcatchemt: Aquifer vulnerability Map Assam subcatchment: Estimated groundwater availability Map The Koshi and Lhasa basins. The hatched area in the Koshi basin is the modelled upper sub catchment Map 6.1. Soil type map of the Salzach river reference catchment Map 6.2. Reclassified soil information in the Lhasa river catchment Map 6.3. HRU delineation in the Upper Danube River Basin Map 6.4. HRUs in the Salzach reference catchment Map 6.5. HRUs in the Lech reference catchment Map 6.6. HRUs in the Lhasa river reference catchment Map 6.7. HRUs in the Wang Chu reference catchment Map 6.8. HRUs in the Brahmaputra in Assam reference catchment List of Tables Table 1 1: Data sets used as input data Table 1 2: Data sets produced and their metadata (as being entered in RBIS) Table 2 1: Data sets used as input data for Table 2 2: Data sets produced Table 2 3: Glacier areas and area changes Table 3 1: Data sets used as input data for Table 3 2: Data sets produced Table 3 3: Permafrost area in the UBRB and UDRB Table 4 1: Input data sets used Table 4 2: BrahmaTWinn land use / land cover (LULC) classification scheme for Level 1 and Level 2. LC classification using spectral signal processing and complement knowledge, i.e. from a DEM Table 4 3: Data sets produced in Table 5 1: Data sets used as input data Table 5 2: Hydrological Classes based on the hydrological dynamics of the different wetlands (Keddy 2000) Table 5 3: Lhasa Subcatchment: Ecosystem Services, Biodiversity and Vulnerability Assessment Table 5 4: Assam subcatchment: Ecosystem Services, Biodiversity and Vulnerability Assessment Table 5 5: Wang Chu Subcatchment: Ecosystem Services, Biodiversity and Vulnerability Assessment Table 5 6: UDRB: Ecosystem Services, Biodiversity and Vulnerability Assessment / 91

7 Table 5 7: Data sets produced and their metadata (as being entered in RBIS) Table 5 8: Weight of the parameters Table 5 9: Input data sets used for Table 5 10: Pitman model parameter description Table 6 1: Input data sets used Table 6 2: Data sets produced in Table 6 3: Number of HRUs produced in the respective reference catchments Table 7 1: Population of RBIS / 91

8 Executive Summary The Deliverable of Workpackage 3, composed by Dl_3.1 through Dl_3.7, comprises the results of the assessment of the different components of the natural environment (NE). Work was carried out within the first 24 months of the project. Deliverables allocated to Dl_3 were required input to WP_2 and WP_4 till WP_10, for acceptance to WP_1 and for dissemination to WP_11 (DOW p 53). Workpackage 3 pursued the following S&T objectives: (1) Assess, classify and quantify the components of the natural environment such as topography, climate, hydrology, snow and glacier cover, permafrost and slope stability, land use and land cover, soils and geology, sediments and erosion, water quality, eco hydrology and biodiversity. They are required inputs to all subsequent workpackages as well as relevant to the assessment of present IWRM practices and the development and evaluation of adaptive IWRM options in WP_10. The assessment will provide the parameter quantification used for the hydrological modelling in WP_7. (2) Based on the spatial representation of the NE, a regionalisation of the basins heterogeneity by means of Hydrological Response Units (HRU) is carried out. HRUs are considered the baseline information elements for the delineation of Water Resources Response Units (WRRU) applied as model entities for the development and evaluation of adaptive IWRM options in WP_10. Adhering to these objectives, work carried out in WP_3 produced the following results: Seven DSM data sets, two Basin DSM and five Reference Catchment DSM. Basin DSMs have a spatial resolution of 1 by 1 kilometre, whereas reference catchment DSMs are sampled in 90 by 90 meters. All DSMs were derived from corrected SRTM data (with voids filled, 2 nd and 3 rd generation) and/or adapted from existing national digital elevation information (Dl_3.1). (1) Analysis of glacier area percentages and the magnitude of glacier area shrinkage and area loss in the Lhasa river and Wang Chu catchments in the UBRB and the Salzach catchment in the UDRB, and an estimate of the loss of water reserve bounds in glaciers for the entire UBRB since the 1970ies. (2) Trends of snow cover changes between 2000 and 2006 and their interseasonal variations between April and October (Dl_3.2). (1) Modelling of mountain permafrost distribution in the UBRB, with estimates of total permafrost area and its comparison to the UDRB permafrost area. (2) Identification of areas particularly affected by permafrost changes as the present day lower boundary of permafrost distribution and the elevation zone where permafrost might disappear in the coming decades. (Dl_3.3) (1) Consistent land use / land cover GIS data set for the year 2000 in all reference catchments. (2) Change analysis in the Lhasa reference catchment and the Brahmaputra in Assam reference catchment. (3) Detailed, fine scaled land use analysis in selected areas (Dl_3.4). Wetland assessment and assessment of groundwater recharge are presented for the UBRB and UDRB as well as groundwater vulnerability assessments. (Dl_3.5) HRU delineations for all reference catchments within UBRB and UDRB (Dl_3.6). (1) RBIS populated by the results of WP 3. (2) Dissemination of project results in scientific publications and conference presentations (Dl_3.7). 8 / 91

9 BRAHMATWINN > Assessment of the Natural Environment > Introduction Introduction Sustainable IWRM aims to distribute water both in sufficient quantity and quality based on detailed knowledge of the regeneration potential inherent in the hydrological regime of the river basin, and balance this potential against projected demands. BRAHMATWINN addressed this notion by a comprehensive assessment of the interactive dynamics between the natural environment (NE) and its human dimension (HD). Each basin, country or region has its specific natural environment and human dimension, a history of water use, cultures and customs, different situation of economic development, and various value judgement exists that is based on these. BRAHMATWINN by its integrated approach and IWRMS toolset has accounted for the specific situations given in the various regions covered by the project scope. It is common understanding that future challenges for a sustainable human development in river basins are closely related to the global dynamics of both NE and HD. To understand these interactions, to predict, quantify and evaluate such scale related impacts in favour of adaptive IWRM, an integrated holistic approach is required that utilises synergetic integration methods from socio economic as well as natural science and technologies from remote sensing and geoinformatics. The innovative IWRM challenge of BRAHMATWINN was to elaborate on in the macro scale twinning basins the spatial distributed modelling and evaluation of adaptive IWRM options for mitigating likely impacts from climate change on the hydrological regimes. This challenge required a thorough river basin systems analysis as a prerequisite to identify and classify systems similarities that indicates common dynamics in the twinning UDRB and UBRB. BRAHMATWINN has provided the basis for management and technological tools by assessing both NE and HD, e.g. the water cycle, water quality and availability, water allocation and sanitation, water related issues of gender, poverty and others by integrated indicators identified for the UDRB and UBRB to monitor the respective system status. Workpackage 3 dealt with a comprehensive assessment of the natural environment. Objective STO_3 of the BRAHMATWINN projects reads as follows: A comprehensive assessment and analysis of the natural environment (NE) comprising e.g. runoff, groundwater, glaciers, permafrost, terrain, land use, land cover, and eco hydrological research to derive the interactive dynamics of the system s components. Workpackage 3 pursued the following S&T objectives: (1) Assess, classify and quantify the components of the natural environment such as topography, climate, hydrology, snow and glacier cover, permafrost and slope stability, land use and land cover, soils and geology, sediments and erosion, water quality, eco hydrology and biodiversity. They are required inputs to all subsequent workpackages as well as relevant to the assessment of present IWRM practices and the development and evaluation of adaptive IWRM options in WP_10. The assessment will provide the parameter quantification used for the hydrological modelling in WP_7. (2) Based on the spatial representation of the NE, a regionalisation of the basins heterogeneity by means of Hydrological Response Units (HRU) is carried out. HRUs are considered the baseline information elements for the delineation of Water Resources Response Units (WRRU) applied as model entities for the development and evaluation of adaptive IWRM options in WP_10. 9 / 91

10 BRAHMATWINN > Assessment of the Natural Environment > Introduction The work to be carried out in this WP comprises field based research, data integration, remote sensing and GIS analyses and also makes use of the established stakeholder processes in the twinning basins. It combines innovative techniques with traditional field campaigns for ground validation, and has integrated available datasets from global and national depositories. The regionalization concept of Response Units (RU) was applied to the assessment results to delineate Hydrological Response Units (HRU) that classify the distributed heterogeneity of the natural environment in the UDRB and the UBRB. Results are collected and presented in a comprehensive digital GIS map assembly populated into the RBIS. 10 / 91

11 BRAHMATWINN > Assessment of the Natural Environment > SRTM based terrain classifications 1 Deliverable 3.1 SRTM based terrain classifications according to IPCC recommendations and digital terrain models (DTM) for the UDRB and UBRB 1.1 Lead partner and partners involved FSU (ZGIS) 1.2 Rationale and main outcome The rationale of 3.1 was to acquire, prepare, compile and visualise digital elevation datasets correct and calibrate the DSM classify geomorphometric and hydrological units As main outcomes of task 3.2 we delivered: Elevation models (DSM; in total seven DSM data sets for two scale domains Basin DSM and Reference Catchment DSM) with different resolutions; Geomorphometric parameter datasets generated and distributed to the project partners for integration in different model applications. 1.3 Methodology Input data The Shuttle Radar Topography Mission (SRTM) obtained elevation data on a near global scale to generate the most complete high resolution digital topographic database of Earth. SRTM consisted of a specially modified radar system that flew onboard the Space Shuttle Endeavour during an 11 day mission in February of Currently the SRTM dataset is most accurate elevation model which is globally available with the resolution of 90m. The dataset is continuously updated and enhanced. For the purpose of this project Version 3 data has been used and validated against high resolution elevation data. Data set Source / Origin Scale / Resolution Costs involved Remarks SRTM NASA/CGIAR 90 m Free Version 3 Table 1 1: Data sets used as input data Methods applied The morphology of the surface is the primary data layer for assessing water resources. Two different representations are generally used: Digital Surface Models (DSM) the height above sea level including all features on the surface (e.g. trees, houses, etc.), and the Digital Terrain Models (DTM), the height a.s.l. truly to the surface. Height measurements from space or airborne platforms naturally render DSMs which can be converted to DTMs by correcting for land cover. For the scales 11 / 91

12 BRAHMATWINN > Assessment of the Natural Environment > SRTM based terrain classifications and applications of BRAHMATWINN the difference between the two models was considered negligible. In task 3.1 we generated in total seven DSM data sets in two scale domains depending on the category, Basin DSM or Reference Catchment DSM. The basin DSMs of the twinning river basins Upper Danube and Upper Brahmaputra have a spatial resolution of 1km, whereas the reference catchment DSMs Lhasa, Wang Chu, Brahmaputra in Assam, Salzach, and Lech have a resolution of 90m. The SRTM data needed to be corrected for data voids and major errors occurring at specific land cover features (e.g. water land boundary, mountainous areas). For compatibility reasons, a calibration of height values at reference survey points was carried out. Voids were filled by data derived by stereoscopic methods applied to optical satellite imagery from the Advanced Space borne Thermal Emission and Reflection radiometer (ASTER), an experimental sensor on the Terra Platform. For the river basin level (Upper Brahmaputra River Basin, Upper Danube River Basin) DSMs were resampled from a 90 m to a 1 km raster, and clipped to the basins boundaries. The 90 m DSMs served to represent the reference catchments. For providing adequate data for modeling, sinks in the DSM were filled where required. The resulting DSM have been classified according to geomorphometric and hydrological criteria required to delineate the Hydrological Response Units (HRU) as well as to model parameters such as permafrost distribution and slope instability. Therefore the morphometric parameters aspect, slope, profile curvature, solar radiation index, flow accumulation and flow direction were derived from the DSM. 12 / 91

13 BRAHMATWINN > Assessment of the Natural Environment > SRTM based terrain classifications 1.4 Achievements Map showcase Map 1.1. Digital surface model and terrain classification of Upper Brahmaputra River Basin and Upper Danube River Basin (resolution 1 x 1 km, scale: 1: 3,000,000) 13 / 91

14 BRAHMATWINN > Assessment of the Natural Environment > SRTM based terrain classifications Map 1.2. Digital surface model and terrain classification of UBRB reference catchments of Lhasa River, Wang Chu, and Brahmputra in Assam 14 / 91

15 BRAHMATWINN > Assessment of the Natural Environment > SRTM based terrain classifications Map 1.3. Digital surface model and terrain classification of Upper Danube River Basin (resolution 1 x 1 km, scale: 1: 1,100,00) Data sets produced Data set Source / Origin Scale / Resolution DSM UBRB SRTM 1 km x 1 km DSM UDRB SRTM 1 km x 1 km DSM Wang Chu Catchment SRTM 90 m x 90 m DSM Lhasa River Catchment SRTM 90 m x 90 m DSM Brahmaputra in Assam SRTM 90 m x 90 m DSM Salzach River Catchment SRTM 90 m x 90 m DSM Lech River Catchment SRTM 90 m x 90 m DSM derivatives (such as slope, aspect, profile curvature, solar radiation index, flow accumulation, flow direction) for UDRB and UBRB SRTM Table 1 2: Data sets produced and their metadata (as being entered in RBIS) 15 / 91

16 BRAHMATWINN > Assessment of the Natural Environment > SRTM based terrain classifications Project relevant achievements Elevation datasets are an important requisite for model inputs. The different elevation based datasets have been made available to the project and have specifically been used in this WP in the following tasks: Deliverable 3.2: Analysis of historical glacier and snow distribution with change detection analyses Deliverable 3.3: Terrain based permafrost distribution and vulnerability analyses of slope stability Deliverable 3.4 IPCC based classification of land us and land cover (LULC) with change detection analysis Deliverable 3.6 Hydrological systems analysis and delineation of Hydrological Response Units (HRU) Lessons learned A validation (in the high mountain area of SE Bavaria) of the version 3 datasets of the SRTM has been carried out as an important milestone in this deliverable. The conclusions can be summarised as follows: Previous vertical accuracy assessments of the SRTM data seem a bit too optimistic, particularly for high mountain areas Although it version 3 gives a better impression, improvement is limited for the validated area (it might be for areas where other DTM data was available) Trend of underestimating within interpolated areas and overestimating outside them Areas with interpolated surfaces (ex voids) should not be used for terrain analysis or as test areas without checking the vertical accuracy 16 / 91

17 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution 2 Deliverable 3.2 Analysis of historical glacier and snow distribution with change detection analyses 2.1 Lead partner and partners involved UniOs (FSU, GDS, ZGIS, IITR) 2.2 Rationale and main outcome The rationale of task 3.2 was to compile the past and present glacier distribution in the UBRB and the UDRB for suitable points in time in order to detect and analyse recent changes in glacier cover in the basins; analyse satellite data series for retrieving and analysing recent changes in snow cover in the two basins. As main outcomes of task 3.2 the group has found that the Lhasa river and Wang Chu catchments in the UBRB and the Salzach catchment in the UDRB had in 2000 a glacier area percentage in the order of 1.1 % to 1.6 %. this percentage glacier area shrank since the 1970s by around 0.2 % due to a glacier area loss of around 6 7 % per decade. that the glacier volume loss corresponding to the above glacier area loss, i.e. the loss of water reserve bound in glaciers, must have been in the order of 20 % since the 1970s for the entire UBRB. MODIS data are well suited to perform snow cover change detection. No trends could be observed for the time span between due to the short period. However it could be observed that significant annual variation take place, especially in lower to middle altitude lying areas. Snow cover occurs more stable in April with higher fluctuations in October. 2.3 Methodology Input data The 1970s glacier inventory compiled for the UBRB was mainly based on the Chinese Glacier Inventory. This inventory is based on airphotos that were, for the area investigated in detail, acquired between 1970 and For the Wang Chu catchment in Bhutan, the inventory was digitized from Corona satellite data from The year 2000 glacier inventories in the UBRB were mainly compiled from Landsat7 ETM+ multispectral satellite data, supported by ASTER satellite data (Advanced Spaceborne Thermal Emission and Reflection Radiometer). For the UDBR, the pre existing glacier inventories were digitized from airphotos of 1969 and For retrieving the snow cover and snow cover change since 2000 until 2006 a series of the MODIS (Moderate resolution Imaging Spectroradiometer) satellite derived snow cover product Snow Cover 8 Day L3 Global 0.05Deg CMG, Version 4 (MOD10C2) was used and parameters of mean snow cover and snow range (variability) for the months of April and October have been calculated. The three dimensional parameters of glacier and snow cover were intersected from the SRTM elevation model (see section 1). 17 / 91

18 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution Data set Source / Origin Scale / Resolution Costs involved Remarks Chinese Glacier GLIMS data base / From airphotos Free inventory Chinese Academy with scale number of Science of several Landsat5 TM and US Geological Spatial resolution Ca. 600$ per scene, Landsat7 ETM+ Survey / NASA 30x30m per pixel partially free satellite images ASTER satellite images NASA 15x15m 30x30m Ca. 80$ per scene, partially free CORONA satellite images SRTM elevation model USGS 6x6m 9x9m Ca. 30$ per image See section 1 MODIS NSIDC EOS 5 km x 5 km Free Snow cover 8 day composite CMG at 0.05 (MOD10C2) Table 2 1: Data sets used as input data for Methods applied The Chinese Glacier Inventory based on airphotos from the 1970s and 1980s was carefully checked glacier by glacier for the Lhasa river test catchment and an additional test area in the north western UBRB. This check was supported by overlay and comparison of the CGI glacier outlines with Landsat and Corona satellite data. Glaciers with obvious digitization or georeference errors were excluded from multitemporal analyses, e.g. from comparisons with the 2000 glacier inventories. In the Lhasa river and Wang Chu catchments, and the additional test region the north western UBRB, the glacier outlines for around the year 2000 were obtained by semi automatic segmentation. A automatic glacier mask was derived from a ratio between visible and short wave infrared channels of Landsat ETM satellite imagery. Initial glacier outlines were obtained by thresholding and rastervector conversion. Substantial manual editing of these initial outlines based on the original satellite images was used to correct segmentation errors (e.g. frozen lakes) and include missing areas (e.g. debris covered ice). For reference in the Lhasa river basin, the Chinese glacier inventory from the year 1970 was used. For the Wang Chu basin reference data for 1974 were manually digitized from CORONA satellite data. The results (glacier area and area changes) were up scaled to the entire UBRB using the Wang Chu, the Lahasa river and the north western test area. However, uncertainty on how representative the 3 test areas are for the glacierization of the entire UBRB adds uncertainty to the values for the entire UBRB compared to the directly measured values for the test areas. This upscaling to the entire UBRB was done through relating the area change of a certain glacier size class in the test areas to the sample size of this size class in the entire UBRB (from Chinese Glacier Inventory and ICIMOD data base) The glacier volumes in the test areas were estimated using two widely used empirical area volume relations. 18 / 91

19 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution For the Salzach catchment (UDRB) glacier outlines were existing for two points in time. The data were manually digitized from airphotos of around 1969 and Using MODIS snow cover 8 day composite CMG grids as input files, the applied method calculates on a per pixel basis for both the UDRB and the UBRB for each month April and October of the period (1) the snow cover change of two subsequent years, (2) the annual difference of snow cover to average snow cover and (3) the annual proportion of snow cover as percentage of average snow cover. Data was downloaded through the FTP pull option offered by the EOS Data Gateway. As one month is covered by four 8 day CMG datasets 56 scenes were collected for each test area to cover April and October for each year. For each month the composite with the least cloud cover was identified by visual interpretation in order to guarantee maximum cloud free land surface. The MODIS snow mapping algorithm already builds a cloud mask to which all cells with a CCA of more than 80% are assigned. This information is provided with the datasets. In this analysis we defined a CCA threshold of 20% to minimise the number of cells, where snow and cloud confusion persists. Therefore, the maximum CCA was calculated for each monthly time series (7 datasets per series) by pixel based extracting the highest CCA value of the relevant scenes. The identified cloudcovered cells were added to the existing MODIS cloud mask. This mask is further extended by areas that are free of snow (SCA = 0%) leading to an overall exclusion mask for each monthly time series. Thus, only the snow covered and at the same time cloud free (CCA < 20%) areas are considered for change detection analysis. Results of each monthly time series and test site were put into a sequence in order to animate snow cover change. Thus, interpretation of results was facilitated and change detection could be easier performed. Moreover, animations are an appealing way of presenting research outputs to a wider audience. 19 / 91

20 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution 2.4 Achievements Map showcase Map 2.1. Snow cover range April in the UBRB (scale: 1:3,000,000) 20 / 91

21 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution Map 2.2. Snow cover mean April in the UBRB (scale: 1:3,000,000) Map 2.3. Snow cover range April in the UDRB (scale: 1:1,250,000) 21 / 91

22 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution Map 2.4. Snow cover mean April in the UDRB (scale: 1:1,250,000) Data sets produced Data set Wang Chu catchment glacier inventory for 2000 Wang Chu catchment glacier inventory for 1974 Lhasa river catchment glacier inventory for 2000 North west UBRB glacier inventory for 2000 Revised version of 1970 Lhasa river Chinese Glacier Inventory (CGI) Revised version of 1970 north west UBRB Chinese Glacier Inventory for 1980 Glacier inventories for Salzach catchment for 1969 and 1998 Glacier change detection by intersection of repeat glacier inventories Snow cover change of subsequent years (2000/2001, 2001/2002, 2002/2003, 2004/2005, 2005/2006) Annual difference of snow cover to average snow cover (2000, 2001, 2002, 2003, 2004, 2005, 2006) Remarks Semi automatically derived from Landsat data Manually digitized from Corona data Semi automatically derived from Landsat data Semi automatically derived from Landsat data Manually corrected from original CGI Manually corrected from original CGI Compiled from existing data Available for Lhasa river, Wang Chu, NW UBRB and Salzach Available for UDRB and UBRB Available for UDRB and UBRB 22 / 91

23 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution Mean snow cover for April and August ( ) Snow cover range (variability) for April and August ( ) Available for UDRB and UBRB Available for UDRB and UBRB Table 2 2: Data sets produced Project relevant achievements Relevance Glacier inventories, i.e. glacier outlines and areas provide a map of glacier covered areas for a certain time step, defined by the base data used. A glacier inventory is an important data base for retrieving all kind of glaciological indicators, and for visualizing, analyzing and investigating land cover and its changes. Repeat glacier inventories reveal the degree of glacier area loss. Glacier area loss is an important indicator for climate change in the reference area, climate change impact on cryospheric systems in the reference area, climate change impact on glacier hydrology in the reference area and downstreams, and probability and distribution of some glacier related hazards, e.g. lake outbursts. The glacierized area percentage gives a general indication on the ice resources in a catchment and the degree of glacial character of it. The latter influences a number of factors such as seasonality of river runoff and dependency of dry season river discharge on glacier melt sensitivity / vulnerability of the catchment to climate change induced glacier changes Glacier hypsography gives the total glacier area per elevation bin in a catchment. This vertical glacier area distribution indicates how sensitive the glacier areas potentially are to climate changes and how the impacts of climate changes on glaciers might look. For instance, glacier cover with large area percentages at low elevations might be more sensitive to atmospheric warming than a glacier cover with a large area percentage at high elevations. Glacier hypsography is particularly important for monsoon type glaciers, where a change in air temperature not only affects the degree of ice melt (glacier ablation) but also the degree of accumulation (solid or liquid). Snow cover estimates provide important results on the variability and amount of snow covered areas. This is a crucial prerequisite for hydrological modelling as the snow cover is next to glacierisation an important annual reservoir of water. Knowledge on the snow cover dynamics become more important with the decrease of glaciers. Results and Conclusions Repeat glacier inventories of the Lhasa river and Wang Chu catchments have been produced and submitted to RBIS. The glacier area in the Lhasa river catchment is about 8 times larger than in the Wang Chu catchment. The glacier area change in both areas is similar in both catchments with around 7 % per decade, though slightly lower in the Wang Chu, presumably due to the significant debris cover of the glacier tongues in the Wang Chu. Such debris cover reduces glacier ablation and thus reduces glacier mass loss and retreat. The glacier area loss in the Salzach catchment during ~1970 to ~2000 was similar, slightly lower, to the area loss in the UBRB catchments. Total glacier area percentages for the catchments are low with slightly above 1 %. The Lhasa river catchment is slightly more glacierized than the Wang Chu one. The glacier area percentage of the Salzach catchment is between the one of Lhasa river and Wang Chu. All three catchments lost a few 23 / 91

24 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution 0.1 % in total glacier area percentage between the 1970s and ca The total glacier area percentage for both UBRB catchments is about half of that for the entire UBRB. Glacier changes in the UBRB will thus have potentially slightly more impact than in the two UBRB test catchments investigated in detail. The glacier hypsographies show that the maximum glacier areas are at around 5300 m asl for the entire UBRB, 5200 m for Wang Chu, 5800 m for Lhasa river and 2700 m for Salzach. The Lhasa river glacier cover at comparably high elevations might therefore be less vulnerable to a certain rise in air temperature than the other catchments. Compared to the entire UBRB, the Lhasa river glacier cover is restricted to a comparably small elevation band. The Wang Chu glacier cover shows a second peak at around 4500 m asl. representing the large debris mantled glacier tongues that survive at lower elevations due to their debris insulation. These low elevation glacier parts are, though, particularly exposed to air temperature rise and could be stagnantly down wasting. This process is known to potentially lead to the development of glacier lakes. Through empirical area volume scaling and upscaling to the entire basin it was found the glaciers in the UBRB lost about 20 % of their volume between around 1970 and This totals to a ice volume loss of 175 km 3, or 7 km 3 per year, or a glacier mass balance of about 0.3 m water equivalent per year, or about mm sea level equivalent per year. Lhasa river catchment Wang Chu catchment UBRB (*estimated from upscaling) Salzach Catchment area km km km km 2 Glacier area ~ km 2 50 km km 2 * 79 km 2 Glacier area ~ km 2 60 km km 2 95 km 2 Glacier area change per decade Glacier area percentage ~2000 Glacier area percentage ~ %/10yr 6.6 %/10yr 7.5 %/10yr * 6.3 %/10yr 1.3 % 1.1 % 2.8 % * 1.2 % 1.6 % 1.3 % 3.4 % * 1.4 % Table 2 3: Glacier areas and area changes 24 / 91

25 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution Figure 1. Histograms of glacier cover in the test catchments Results of snow cover change of two subsequent years show significant annual variation in the distribution of snow for both test areas due to high yearly differences of snow covered and snow free areas. Areas affected by high change rates are difficult to delineate, but are basically located in the western and northern UBRB, and in the central and southern UDRB. The yearly difference slightly depends on the elevation. Snow cover change is more likely to appear in low altitude areas, whereas there is less variation in the high mountains. The higher the area the more stable is the percentage of snow cover over time. This is especially true for the southern and eastern border regions of the UBRB snow covered area. However, there is no observed trend neither for the UBRB nor the UDRB of how snow cover change behaves and will behave in future. All results show significant annual variation of snow distribution with higher fluctuations for the October time series of both the UBRB and UDRB. It must be considered that the high change of annual snow cover in October either results from the differences up to three weeks in time span covered by each input dataset of one monthly time series or from differences in the occurrence of the first bigger snowfall event. The applied methods show valid results and therefore can be widely used to analyse the effects of climate change Lessons learnt Multispectral, semi automatic remote sensing based glacier classification based on nearinfrared/shortwave infrared ratios proofed to be a successful and robust way of mapping glaciers and glacier changes over large areas. Such initial glacier outlines require careful manual editing by experts in order to allow for compilation of glacier inventories, that consist of individual glaciers. Once glacier inventory data from different sources are used to retrieve glacier area changes (here: Chinese glacier inventory and own satellite image derived glacier inventory), a careful glacier byglacier check is necessary in order to exclude or correct false glacier outlines (from georeference errors, misclassifications, interpretation uncertainties, etc.) from the multitemporal analysis. The applied methods for snow cover estimations show valid results and therefore can be widely used to analyse the effects of climate change. However, in this study the investigation time period still is too short to observe potential climate trends in relation to snow cover distribution, but cannot be expanded as MODIS products are not available before the year A second constraint of MODIS data refers to missing information on snow depth, which would be essential for further studies on snow water equivalent, and further on climate change impacts and water resources. Nevertheless, 25 / 91

26 BRAHMATWINN > Assessment of the Natural Environment > Analysis of historical glacier and snow distribution MODIS is the only sensor that provides high quality snow cover data at moderate resolution on a global and near daily level. 26 / 91

27 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution 3 Deliverable 3.3 Terrain based permafrost distribution and vulnerability analyses of slope stability 3.1 Lead partner and partners involved ITP (ZGIS, UniOs) 3.2 Rationale and main outcome The rationale of task 3.3 was to model the mountain permafrost distribution in the UBRB and the UDRB, and to validate the model results in particular for the UBRB where such model was developed for the first time; identify areas which are particular prone to climate change induced permafrost warming and associated slope instability problems. As main outcomes within task 3.3 the mountain permafrost distribution in the UBRB was modelled for the first time, revealing a total permafrost area percentage of around % (compared to around 3 % glacier area percentage). In contrast, the permafrost area in the UDRB is in the order of 3 4 % in maximum. areas potentially particularly affected by permafrost changes were identified as the presentday potential lower boundary of permafrost distribution and the elevation zone where permafrost might disappear in the coming decades. 3.3 Methodology Input data The permafrost distribution models applied to the UDRB and UBRB (PermaKart) were mainly based on the SRTM elevation model (see section 1) and the climatic parameters elevation of the zero degree isoline and temperature gradient with elevation (lapse rate). The latter parameters were compiled from meteorological time series obtained from different partners and sources. Data set Source / Origin Scale / Resolution Costs involved Remarks SRTM elevation See section 1 model Meteorological Chinese, Bhutanese Very scattered Free through temperature series and Austrian distribution, in project partners meteorological or particular in the hydrological offices UBRB Table 3 1: Data sets used as input data for Methods applied The permafrost distribution in the UDRB was modelled as permafrost occurrence probability. The latter was modelled as a function of the 27 / 91

28 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution mean annual air temperature at a certain point as derived from the elevation of the 0 isoline and the mean temperature lapse rate in the area of concern, and the potential incoming short wave radiation Both factors, elevation and radiation, were derived using the SRTM elevation model. The 0 isoline elevation and temperature lapse rate were derived from a number of meteorological station data.. An empirical relation was then applied to estimate the local ground temperature as a function of local radiation and air temperature, and empirical thresholds used to then separate between probable (< 3 C), possible ( 2 C 3 C) and no permafrost occurrence (> 2 C). Since such model was for the first time applied to the UBRB, the model results were validated using a more physically based permafrost model in a small validation area, alongside with a rockglacier inventory that was compiled from high resolution satellite data. Rockglaciers are permafrost features and should thus not far extent under the modelled lower permafrost limit. These two validation measures both supported the model results for the entire UBRB. For the UDRB, a model similar to the above was applied, but using no radiation term and an aspectdependent probability threshold instead. The model resulted in two classes: permafrost possible or not possible. The permafrost model estimates permafrost occurrence irrespective of glacier cover. Usually, permafrost is not existent under glaciers due to the heat transport to the glacier bed by melt water. However, the high and cold glaciers in Tibet might well be underlain by permafrost, at least in parts. The glacier areas were therefore not subtracted from the permafrost map but rather two alternatives given (permafrost with and without glacier areas). 28 / 91

29 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution 3.4 Achievements Map showcase Map 3.1. Modelled permafrost distribution in the UBRB and glacier cover 29 / 91

30 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution Map 3.2. Detail of Map 3.1 over the Himalaya main ridge in Bhutan Map 3.3. Detailed map of modelled permafrost distribution and glacier cover in the Lhasa catchment 30 / 91

31 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution Map 3.4. Detailed map of modelled permafrost distribution and glacier cover in the Wang Chu catchment Map 3.5. Modelled permafrost distribution in the Salzach basin, UDRB, together with glacier cover (blue: glaciers of 1998, red: glacier retreat ) 31 / 91

32 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution Data sets produced Data set UBRB modelled permafrost distribution UDRB modelled permafrost distribution Modelled increase of lower permafrost limit in test areas Remarks Modelled for DEM resolutions 100m, 200m, 400m and 1km Modelled for SRTM resolution of 90m Based on IPCC scenarios provided by WP2 Table 3 2: Data sets produced Project relevant achievements Relevance The existence of mountain permafrost has significant effects on hydrology, slope stability and other natural hazards, sensitivity to climatic changes, landscape development, erosion processes, permafrost glacier interactions, etc. The permafrost area percentage gives a first general indication of the degree of the periglacial character of a region. The latter influences a number of processes such as mass turn over and erosion processes sensitivity / vulnerability of the catchment to climate change induced changes in the ground thermal regime probability of permafrost related hazards, e.g. slope instabilities. The permafrost area hyspography gives a general indication on the vertical distribution of perennially frozen ground in a catchment. This indicates in particular the sensitivity / vulnerability of the ground to climate change induced ground changes, i.e. periglacial slope destabilsation. Results and Conclusions Permafrost distribution in the UBRB and the UDRB was modelled and can be intersected with glaciers, glacier lakes, steep terrain etc. in order to identify potential interactions and climate change impacts. Compared to the Lhasa river catchment and the entire UBRB, the larger part of permafrost occurrence in the Wang Chu catchment is possible, not probable. This indicates a larger area of permafrost close to the melting point and thus a higher sensitivity/vulnerability of the permafrost in the Wang Chu catchment to changes in boundary conditions most important air temperature and snow cover. The permafrost area percentage in the UBRB is comparably high with %, underlining the strong periglacial character of the basin. For comparison, the glacier area percentage is significantly lower with around 3 %. Nearly half of the Lhasa river catchment is presumably underlain by permafrost, in contrast to 3 4 % for the Wang Chu. Periglacial processes dominate in the Lhasa river catchment, whereas they play only a minor role in the Wang Chu catchment. As a consequence, changes in the ground thermal regime due to climatic changes will therefore have significantly more impact in the Lhasa river than the Wang Chu catchment. The largest permafrost areas, in the UBRB and Lhasa river catchment are at around 5200 m asl., and around 4900 m asl. for the Wang Chu catchment. The permafrost area histograms reflect due to the 32 / 91

33 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution strong elevation dependency of permafrost distribution mainly the topographic elevation distribution. The significantly lower elevation of permafrost areas in the Wang Chu catchment indicates a comparably large sensitivity to changes in air temperature and snow cover. Though, these potential impacts affect only small areas due to the small permafrost area percentage in the Wang Chu. In the Salzach basin, UDRB, the area percentage underlain by permafrost is in the order of 3 4 %, perhaps even less, because the model applied to the UDRB assumes slightly warmer conditions for the existence of permafrost compared to the UBRB permafrost model. Similar percentage numbers are found for the entire UDRB than for the Salzach. Lhasa river Wang Chu UBRB Salzach UDRB catchment catchment Catchment area km km km km 2 Permafrost area probable possible total total without glaciers (1970s) km km km km 2 70 km km km km km km km km km km km 2 Permafrost area percentage total total without glaciers 49 % 47 % 4 % 3 % 25 % 21 % 4% 3% Table 3 3: Permafrost area in the UBRB and UDRB 33 / 91

34 BRAHMATWINN > Assessment of the Natural Environment > Terrain based permafrost distribution Figure 2. Permafrost elevation histograms in the UBRB Lessons learnt First order permafrost distribution modelling over large remote areas produces reasonable results when based on comparably simple and robust models. However, the model results depend strongly on available temperature station time series that might in remote areas such as the UBRB be very scarce in horizontal and vertical distribution. The resulting lack of representativeness of the temperature data for the entire model area adds therefore a significant degree of uncertainty that is, however, difficult to quantify. The above uncertainty from lack of representativeness of climatic station data together with the fact that only the most important spatial determinants of mountain permafrost, elevation and shortwave radiation, are modelled and more local influences such as surface cover are omitted leads to a regional scale permafrost distribution model that should not be interpreted on local scales even when plotted at such scales. 34 / 91

35 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover 4 Deliverable 3.4 IPCC based classification of land us and land cover (LULC) with change detection analysis 4.1 Lead partner and partners involved Z_GIS (GDS, ZGIS, ICIMOD, IITR) 4.2 Rationale and main outcome Land use land cover information is a crucial basis for natural resource management, environmental variables, global and regional change affecting ecological processes. Despite the significance of land cover as an ecological variable, our understanding of land cover dynamic is poor. In task 3.4 for the common land cover and land use classification, the LULC group of BRAHMATWINN has developed an adaptive and harmonized classification scheme. During an expert meeting this scheme has been tested by jointly applying the scheme to the different reference areas selected in the twinning basins. Land use and Land Cover (LULC) should adhere to the scheme proposed by the IPCC, but adapted to the specific requirements and settings imposed by the BRAHMATWINN twinning approach. The main outcomes are: A consistent land use / land cover GIS data set for the year 2000 in all reference catchments Change analysis in the Lhasa reference catchment and the Brahmaputra in Assam reference catchment. Detailed, fine scaled land use analysis in selected areas 4.3 Methodology Input data Geo corrected Landsat ETM+ images in WGS84 UTM projection used in this study were obtained from Global Land Cover Facility (GLCF) web portal. Digital elevation model (DEM) with 90m spatial resolution downloaded from Shuttle Radar Topography Mission (SRTM) was also used. ETM+ data original contains 8 bands; in this study we have used only 7 bands. Besides this we have also used administrative maps of Lhasa basin, and for Bhutan Wang Chu basin we have used Bhutan administrative boundary, roads, river, and available topographic land use map together with field data collected by the means of GPS. Data set Source / Origin Resolution Remarks Landsat ETM+ Global Land Cover Facility 30x30m TERRA: ASTER / MODIS GLCF 15x15m / 250m Populated areas global data set Digital Surface Models (from 3.1) Quickbird Digital Globe ~0,6m Quickbird data acquired for specific areas GPS measurements for verification and validation Table 4 1: Input data sets used ICIMOD / FSU n/a 35 / 91

36 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Methods applied With the objective to enhance the classification results different approaches have been applied to the LULC classification process, such as the application of an ecologically driven spectral decision tree approach in combination with Cognition Network language (CNL), which includes image segmentation and object based feature extraction. Administrative and topographic map layers have been integrated together with field data collected by the means of GPS. Depending on the scale domain of the application, Landsat ETM+ data (~ the year 2000) have been used on basin scale and recent Quickbird data for fine scale studies. Landsat TM datasets with complete coverage available from the year 2000 (+/ 1 year) together with complement satellite images and already published data available from the Web were used to produce LULC classification for all test catchments. For compatibility reasons the classification was carried out according to the IPCC (2003) guideline which uses six land use and cover classes. Suitable satellite data were acquired, selected and compiled. A LULC expert workshop in Salzburg (07.5. to ) was organized to ascertain compatible classifications in Asia and EU. To ascertain comparative studies between river basins, a harmonized classification scheme was developed which links to the IPCC guideline (2003) based on six LULC classes. The classification scheme used is adaptive and hierarchical, where different levels are defined by the underlying data type, the availability of external knowledge and ground truthing. Level 1 and level 2 comprise eight main classes and 22 subclasses, respectively. These classes can be derived from satellite imagery and additional auxiliary data (like DSM derivatives). Level 3 and level 4 require additional external knowledge, either form experts or from field surveys. Level 3 contains structurally defined subclasses (e.g. dense vs. sparse forest), whereas level 4 reflects specific land use types (e.g. irrigated vs. inundated). Level 5, finally, the species level requires very high spatial resolution data and field verifications, and thus can only be applied in selected fine scaled test beds. As an input to subsequent modeling tasks, Level 1 and Level 2 have been applied to the reference catchments of the twinning basins. Figure 3. Levels of details for BRAHMATWINN LULC classifications. 36 / 91

37 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Level 1 Level Agriculture 1201 arable land 1202 pasture and/or meadow 1203 plantation 2000 Bare Ground 2201 soil 2202 rock and debris 3000 Forest 3201 coniferous 3202 deciduous 3203 evergreen 3204 mixed 3205 plantation 4000 Non Forest Vegetation 4201 bushland 4202 alpine grassland 4203 grassland 4204 shrubland 5000 Ice and Snow 5201 glacier 5202 snow 6000 Bulit Up Areas 6201 urban 6202 rural 7000 Open Water 7201 water courses 7202 water bodies 8000 Unclassified 8201 clouds 8202 shadow Table 4 2: BrahmaTWinn land use / land cover (LULC) classification scheme for Level 1 and Level 2. LC classification using spectral signal processing and complement knowledge, i.e. from a DEM In the recent years a new methodology and concept known as object based image analysis (OBIA) was introduced into remote sensing sciences as implemented in a commercial software environment. Satellite images are partitioned into image segments which can be modeled to meaningful objects. Their attributes can be used for advanced, knowledge based classifications. The generated image objects have spatial properties such as size and form, neighborhood relations, hierarchical relationships between different levels of aggregation, etc. In addition to the spectral properties of the objects, all these additional information can be used in the classification process. 37 / 91

38 BRAHM MATWINN > Asssessment of th he Natural Envirronment > IPCC C based classificcation of land u us and land cove er Automated LULC C classificatio on is often hampered h in n areas of hiigh relief due to shadow wing effects,, h results in different values v for one o and the e same land d cover classs (Fuerederr, in press).. which Topographic norm malization as a a pre proccessing step p aims at compensating for the top pographicallyy W the caatchments off either high altitudes orr inducced illuminattion variation by means of a DEM. Within very rugged terraain, Landsat TM data as well as a SRTTM (resampled to 30 m spatial resolution) weree used as input data to calculaate differentt methods of o topograph hic normalizaation (cosine e correction,, Minn naert correcttion, C correcction and staatistic empirrical correction). The results were evvaluated and d comp pared visually and statisttically concerning qualityy and usability in order tto improve th he followingg LULC classificatio on. Minnaertt correction, C correctio on and statiistic empiricaal correction n proved to o succeessfully redu uce topograp phically indu uced illuminaation variations. Overco orrection, ho owever, also o occurrs in areas off low illumin nation due to o the inadequate estimattion of the d diffuse irradiance as welll as the insufficien nt resolution of the DEM M, which norrmally should be at least as fine as the satellitee imagee (Civco 1989). The classification c ns in the UB BRB catchment (especiaally those areas heavilyy influenced by shadow w effects, mainly occcurring in th he Wang Chu u catchmentt and the Lhaasa catchment) have bee en improved d by ground truthin ng missions carried c out in Oct/Nov Verificaation has beeen based on a collection n of GPS meaasured grou und referencce points in n Lhasa and d 112 refereence points in Bhutan,, respeectively. Change analysis has been caarried out in the Guwahati floodplaiin test area based on a comparison n ween the Landsat ETM mo osaic of 2000 and the Laandsat TM mosaic m of With a fo ocus on bankk betw erosion, the change of the river bed, along a with in ncrease or decrease d of agricultural fields weree invesstigated (see map showcaase below). In su upport to th he evaluatio on of eco hyydrological relevant lan nd cover types (see WTT 3.5), high resolution satellite imagery (Quickbird) was used fo or a detailed d, fine scaleed LULC classsification. A subseet of this classification caarried out in n the wetland d areas of th he Brahmapu utra floodplaain in Assam m is illustrated in th he figure belo ow. Figuree 4. Illustratio on: fine scaled d land use cla assification for a subset of the t Assam Brrahmaputra flloodplain Other specific rem mote sensingg image proccessing meth hods used within 3.4 com mprise: ed Differencce Vegetatio on Index). NDVI N is a reemote sensing index offten used in n NDVII (Normalize monitoring vegettation. It is defined d as th he difference e of value in near infrareed and red band b divided d by su um of these two t bands. 38 / 91

39 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover NDSI (Normalized Difference Snow Index). NDSI is also a ratio index, which can be used to detect snow cover using spectral difference of snow in visible band and short wave infrared. Unsupervised classification. Unsupervised classification is used to extract shadows caused by the big mountains which can t be removed by topographic normalization due to the low son elevation in winter. Vector layer integration. Manual interpretation is adopted to digitize the object which can t be automatically extracted. For instance, built up area and bare ground have similar spectral value and most of small towns are too small to distinguish in ETM+ images with 30 meter resolution. Farmland and wetland below 4200m are also digitized manually. After digitizing, the vector maps can be integrated into expert classification system. Digital Elevation Model (DEM), slope map and aspect map. DEM and slope are two important environment factors of land resource, especially in mountainous area. Basically, they determine land use type. DEM and slope are always introduced and integrated with TM data to improve computer based automatic classification. DEM, slope and aspect can modify part of misclassified object because of spectral similarity. Landsat ETM+ B7 (Short wave infrared). ETM+ B7 lies in water absorption zone. It is sensitive to the water content of ground features. (B4 B2) / (B4+B2). This index is found very well in extraction of alpine frigid meadow, especially in very high mountain zone with large area of Kobresia pygmaea meadow. Grassland zones (Z1, Z2 and Z3). Dividing study area into several sub areas based on similar physiognomy and ecological function before classification can minimize spectral value difference in homogeneous sub area. Water body, ice, snow and glacier. First of all, water body, ice, snow and glacier were extracted from land by NDSI at a threshold of 0.33 and easily separated from shadow. However there was mixture of these three objects, for instance, there s snow on the high altitude water body and some water bodies were freezing. We considered water body and ice are one type and snow as another type. DEM and snow mask were used in this study for dividing these two types. 39 / 91

40 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover 4.4 Achievements Map showcase Map 4.1. Land Use Land Cover Map for Assam (year ), North East India (scale 1:1,000,000) 40 / 91

41 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Map 4.2. Land Use Land Cover Map for Lech reference catchment, Austria /Germany (scale 1:250,000) Map 4.3. Land Use Land Cover Map for Lhasa River Basin, China (scale 1:550,000) 41 / 91

42 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Map 4.4. Land Use Land Cover Map for Salzach River Basin, Austria (scale 1:300,000) Map 4.5. Land Use Land Cover Map for Wang Chu River Basin, Bhutan (scale 1:500,000) 42 / 91

43 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Map 4.6. Change analysis in a subset of the Brahmaputra in Assam catchment between 1990 and 2000 Map 4.7. Change analysis for a subset of Lhasa River Basin, China, between 1990 and / 91

44 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Data sets produced Data set LULC classification in the Wang Chu reference catchment 2000 LULC classification in the Lhasa reference catchment 2000 Remark Classification enhannced by ground measurements Atmospheric correction carried out due to pronounced topographic effects LULC classification in the Brahmaputra in Assam reference catchment 2000 LULC classification in the Salzach reference catchment 2000 LULC classification in the Lech reference catchment 2000 Change analysis in the Lhasa catchment between 1990 and 2000 Change analysis in the Brahmaputra in Assam catchment (subset) between 1990 and 2000 Fine scale land use classification based on recent QuickBird imagery (2006) Change analysis has been carried out with slightly modified class scheme due to different atmospheric conditions Change analysis carried out in a subset of actual reference site Data sets available for selected sites in the Brahmaputra in Assam reference site and the Wang Chu catchment Table 4 3: Data sets produced in Project relevant achievements For all reference catchments consistent LULC classifications are available derived from satellite imagery provided by the same sensor (Landsat ETM+) for the time span around the year Due to the uniform, hierarchical classification scheme the product geospatial layers are intercomparable. Next to documentary character of the produced maps representing the actual distribution of land cover and land use types (see also the area distribution tables for the UBRB reference catchments below), various spatial indicators can be derived from the 3.4 LULC products. These indicators, potentially used for the developed IWRM approach, are shown in the following figure. Indicator Quantification 1 Quantification 2 Quantification 3 Quantification 4 Agriculture Arable land Agriculture Pasture and/or meadow Agriculture Plantation Bare Ground Soil Bare Ground Rock and debris Forest Coniferous Forest Deciduous Forest Evergreen Forest Mixed Forest Plantation Non Forest Vegetation Bushland Non Forest Vegetation Alpine grassland Absolute values [km²] 1990/2000 Relative value [% of total coverage] 1990/2000 Change of total area [km²] between 1990 and 2000 Relative change between 1990 and 2000 [%] 44 / 91

45 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Non Forest Vegetation Grassland Non Forest Vegetation Shrubland Ice and Snow Glacier Ice and Snow Snow Built up areas Urban Built up areas Rural Open Water Water courses Open Water Water bodies Slope [Digital Surface Model] Aspect [Digital Surface Model] In Percent or degrees In Degrees Snow Cover ( ) Difference Year to Year Difference to Mean Percentage of Mean Glacier Inventory (1968/1998)* Area in km² (68/98) Change of total area Relative change in % Permafrost areas Area in km² * only Europe Figure 5. Spatial indicators to be derived by the 3.4 LULC products The following sequence of figures shows the area distribution of LULC types in the three reference catchments in the UBRB. Figure 6. Area distribution of reference catchment Lhasa river 45 / 91

46 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Figure 7. Area distribution of reference catchment Wang Chu Figure 8. Area distribution of reference catchment Brahmaputra in Assam The results from the land use land cover classification of Lhasa River basin and Wangchu River basin shows that the land use map of 2000 has 80.12% (Lhasa Basin) and % (Wangchu Basin) accuracy, which is relatively a high accuracy for the map produced from imagery of Landsat ETM+ with 30m resolution. 46 / 91

47 BRAHMATWINN > Assessment of the Natural Environment > IPCC based classification of land us and land cover Lessons learnt The validity of land use/land cover mapping in the study area using expert classification system could be demonstrated within 3.4. Combination of Landsat imagery and ancillary topographical and environmental data is proved an effective technique. This expert system can also be used in image classification in area with same or similar condition. Although expert system can improve classification accuracy, it is still a far way towards automatic classification. Because of cloud cover or other atmospheric effects (e.g. oblique sun angle due to winter season), as well as the constraints imposed by strong topographic effects, the LULC classification for 1990 was hampered. In cases where these data sets were required, analysis was done on MODIS data sets, with a significant lower spatial resolution (250m) as compared to Landsat (30m), but with sufficient coverage. 47 / 91

48 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands 5 Deliverable 3.5 Assessments of wetlands, their functions and groundwater recharge 5.1 Lead partner and partners involved UniVie, ICIMOD, UniBu, ITP, CARR, GDS, Vodni, IITR 5.2 Introduction In this chapter wetland assessment and assessment of groundwater recharge are presented for the UBRB and UDRB. Additionally, a case study of modelling the discharge for the Lhasa river is included. Wetlands provide for the water flow between the ground water system and the surface water. The numerous functions of wetlands, and their ecosystem services, depend on various factors. One major role is played by the gradient of the groundwater table. Other factors are: geomorphology, geologic conditions, climate, land use, land cover, soil erosion. The modelling of discharge for the Lhasa River in the UBRB relates to one essential variable regulating many other of the complex interactions in river systems. 5.3 Assessment of wetlands and wetland functions Rationale and main outcome The values of the classified wetlands in WP 3.4 were evaluated following the Millennium Ecosystem Assessment (2005). UniVie adapted the framework for the wetland classes at subcatchment scale. The assessment was conducted using the information from the following sources: (I) local data provided by the partner in WP3, (II) data which was collected during the field trips, and (III) through the evaluation of the wetland databases (Ramsar Site Information Service, Global Wetland Inventory Database, Wetlands of India, European Environmental Agency). The Biodiversity of the wetlands was assessed using specific wetland functions and their services as basis for the evaluation in combination with available local data. The character and extent of wetland vulnerability was based on wetland type and information from literature and expert judgement. It was assessed particularly with regard to (I) the threat exerted by the density of settlements located near wetlands and (II) the possible effects of climate change on wetland distribution and sustainability. Density of settlements was partly derived from information provided by partner GeoData Methodology Input data Data set Source / Origin Scale / Resolution Costs involved Remarks Global Lake Lehner & Döll 2004 global free and Wetland ml Database (GLWD) 48 / 91

49 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Ecoregions Olson, D. M, E. Dinerstein, E.D. Wikramanayake, N.D. global free of the Burgess, G.V.N. Powell, E.C. Underwood, J.A. D'amico, I. world Itoua, H.E. Strand, J.C. Morrison, C.J. Loucks, T.F. Allnutt, T.H. Ricketts, Y. Kura, J.F. Lamoreux, W.W.Wettengel, P. Hedao, & K.R. Kassem Terrestrial Ecoregions of the World: A New Map of Life on Earth. BioScience 51: Corine Land Cover (CLC) EEA landcover regional free only UDRB Ramsar sites Wetlands of India local free local free Only UBRB Table 5 1: Data sets used as input data Methods applied The wetland distribution was conducted by a mulit scale approach: low resolution geo data, terrain features (provided by FSU), and the Global Lake and Wetland Database (B. Lehner & P. Döll, 2004) build the framework to identify potential wetland areas. High resolution images, local data collected by the Asian partners (UniBu, IITR, ICIMOD and ITP) and during the field trips, literature reviews, and different wetland databases (Ramsar Site Information System, Global Wetland Inventory Database, Wetlands of India) were used to gain information about the wetlands. A rule based expert system was used to combine the data at subcatchment scale, which uses the LULC classification provided by Z_GIS and the NDVI as basis to identify the different wetland classes Achievements Upper Brahmaputra River Basin (UBRB): For the UBRB 11% (approx. 56,000km²) of the surface could be delineated as wetlands: 84% low altitude wetlands (203 patches), 6% high altitude wetlands (260 patches), 4% lakes (340 patches) and 6% rivers and streams (639 patches). The Lhasa River subcatchment cover three percent (~1100km²) wetlands of the total surface: 40 % alpine Meadows (1466 patches), 31% alluvials (4539 patches), 23% alpine swamps (5265 patches), 3% floodplains (1026 patches) and 3% lakes (112 patches). The Brahmaputra Valley subcatchment in Assam hold 11% wetlands with a total value of 6800km²: 66% alluvials (10,770 patches), 22% beels (2620 patches), 11% natural seminatural floodplains (14 patches) and 1% artificial lakes (171 patches) The Wang Chu River subcatchment shows the smallest wetland area with only 171km² (2% of the surface): 76% alluvials (178 patches), 12% beels (836 patches), 5% swamps (161 patches), 5% lakes (155 patches) and 2% natural seminatural floodplains (43 patches). Upper Danube River Basin (UDRB): 49 / 91

50 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands The UDRB covers approx. 9800km² wetlands (0.01% of the total UDRB surface): 53% lakes (330 patches), 14% rivers (30 patches), 21% high altitude wetlands (137 patches) and 12% low altitude wetlands (137 patches). In the Lech subcatchment 0.04% (~80km²) wetlands could be identified: 53% Lakes (112 patches, including reservoirs); 16% Alluvials (361 patches, underrepresented); 16% swamps (910 patches, including bog, fens, mires and swamps); 15% floodplains (910 patches). 33% Alluvials (14,718 patches); 30% lakes (109 patches, including reservoirs); 22% swamps (225 patches, including bog, fens, mires and swamps); 15% floodplains (514 patches Map showcase Map 5.1. Ecoregions in the UBRB. Ecoregions describe similar ecosystem complexes within its biomes (Olson et al. 2001); black boundaries mark the Catchments within the UBRB. 50 / 91

51 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.2. Global Lake and Wetland Database Layer 3 (B. Lehner & P. Döll, 2004) (top); Wetland area (left) and polygons (right) in the UBRB 51 / 91

52 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.3. Lhasa subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) 52 / 91

53 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.4. Assam subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) 53 / 91

54 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.5. Wang Chu subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) 54 / 91

55 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.6. Ecoregions in the UBRB. Ecoregions describe similar ecosystem complexes within its biomes (Olson et al. 2001); black boundaries mark the Catchments within the UBRB. 55 / 91

56 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.7. Global Lake and Wetland Database Layer 3 (B. Lehner & P. Döll, 2004) combined with the Corine Lake Cover (provided by EEA) (top); Wetland area (left) and polygons (right) in the UDRB 56 / 91

57 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.8. Lech subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) 57 / 91

58 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map 5.9. Salzach subcatchment: Wetland classes (top); Wetland area (left) and polygons (right) Ecosystem Services, Biodiversity and Vulnerability Assessment Local and River Basin scale were combined by a rule based expert system at subcatchment scale, whichh uses the LULC classification (Z_GIS) and the NDVI as basis to identify alluvials, lakes, alpine swamps and meadows, floodplains and beels. These different wetland classes were further classified by their hydrological dynamics into four hydrological classes (Tab.1 ). Hydrological class Flooding Groundwater Description wetlands depending mainly on flooding and inundation with sediment rich runoff from the river wetlands depending on groundwater dynamics with seasonal rise and falling of groundwater Wetland class alluvials, floodplains, beels lakes, swamps Hybrid wetlands of a hybrid nature comprising flooding and beels, lakes groundwater Slope wetlands at footslopes fed by interflow from the adjacent swamps, slopess meadows Table 5 2: Hydrological Classes based on the hydrological dynamics of the different wetlands (Keddy 2000) The Synthesis Report Wetland and Water (Millennium Ecosystem Assessment, 2005) assesses the relative magnitude [low ( ) medium ( ) high ( )] of the ecosystem services based on expert 58 / 91

59 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands opinions for a global average pattern for the wetlands. UniVie adapted the framework (Fig. 1) for the wetland classes at subcatchment scale. The assessment was conducted using the information from the following sources: (I) local data provided by the partner in WP3, (II) (III) data which was collected during the field trips, and through the evaluation of the wetland databases (Ramsar Site Information Service, Global Wetland Inventory Database, Wetlands of India, European Environmental Agency). The median of the values of the following ecosystem services was used to assess the biodiversity of the wetlands: Refugia Biological control Pollination Genetic, medicinal resources The vulnerability of the wetlands was assessed particularly with regard to (I) the pressure of the human population on wetlands and to (II) the possible effects of the climate change on the wetland distribution. The threat through the density of settlements was partly evaluated using population data provided by the socioeconomy partners (GeoDa). 59 / 91

60 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Ecosystem Services, Biodiversity and Vulnerability UBRB Ecosystem Services Alluvials Lakes Floodplains Swamps Meadows Alluvials Lakes Floodplains Swamps Meadows Provisioning Biodiversity Food, Raw materials Genetic, medicinal resources overall median Fresh water var (+-1) Regulating Vulnerability Climate regulation Human Dimension Water regulation Climate Change Water supply Waste treatment Erosion control and sediment retention Disturbance regulation Supporting Refugia Biological control Pollination Soil formation nutrient cycling relative magnitude [low ( ) medium ( ) high ( )] Table 5 3: Lhasa Subcatchment: Ecosystem Services, Biodiversity and Vulnerability Assessment 60 / 91

61 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Ecosystem Services Alluvials Lakes Floodplains Beels Alluvials Lakes Floodplains Beels Provisioning Biodiversity Food, Raw materials Genetic, medicinal resources overall median Fresh water var (+-1) Regulating Vulnerability Climate regulation Human Dimension Water regulation Climate Change Water supply Waste treatment Erosion control and sediment retention Disturbance regulation Supporting Refugia Biological control Pollination Soil formation nutrient cycling relative magnitude [low ( ) medium ( ) high ( )] Table 5 4: Assam subcatchment: Ecosystem Services, Biodiversity and Vulnerability Assessment 61 / 91

62 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Ecosystem Services Alluvials Lakes Floodplains Swamps Beels Alluvials Lakes Floodplains Swamps Beels Provisioning Biodiversity Food, Raw materials Genetic, medicinal resources overall median Fresh water var (+-1) Regulating Vulnerability Climate regulation Human Dimension Water regulation Climate Change Water supply Waste treatment Erosion control and sediment retention Disturbance regulation Supporting Refugia Biological control Pollination Soil formation nutrient cycling relative magnitude [low ( ) medium ( ) high ( )] Table 5 5: Wang Chu Subcatchment: Ecosystem Services, Biodiversity and Vulnerability Assessment 62 / 91

63 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Ecosystem Services, Biodiversity and Vulnerability in the UDRB: Ecosystem Services Alluvials Lakes Floodplains Swamps Alluvials Lakes Floodplains Swamps Provisioning Biodiversity Food, Raw materials Genetic, medicinal resources overall median Fresh water var (+-1) Regulating Vulnerability Climate regulation Human Dimension Water regulation Climate Change Water supply Waste treatment Erosion control and sediment retention Disturbance regulation Supporting Refugia Biological control Pollination Soil formation nutrient cycling Table 5 6: UDRB: Ecosystem Services, Biodiversity and Vulnerability Assessment Data sets produced Related data for the indicators concerning ecosystems and their services (maps, datasheets from WP 3.5 and WP 4) can be found on the project FTP server. Region Information type of data UDRM terrestrial ecosystems map UDRM wetland types map Lech subcatchment wetland classes map Lech subcatchment assessment of ecosystem services and ecosystem functions table Salzach subcatchment wetland classes map Salzach subcatchment assessment of ecosystem services and ecosystem functions table UBRM terrestrial ecosystems map UBRM wetland types map Lhasa subcatchment wetland classes map Lhasa subcatchment assessment of ecosystem services and ecosystem functions table Assam subcatchment wetland classes map 63 / 91

64 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Assam subcatchment assessment of ecosystem services and ecosystem functions table Wang Chu subcatchment wetland classes map Wang Chu subcatchment assessment of ecosystem services and ecosystem functions table Table 5 7: Data sets produced and their metadata (as being entered in RBIS) In this respect the DSMs of the five reference catchments are not merely subsets of the river basin DSMs Project relevant achievements Upper Brahmaputra River Basin (UBRB): Lhasa subcatchment: The overall relative value of the wetlands in the Lhasa subcatchment were assessed from low to high magnitude. In addition to rivers and their plains (alluvials), floodplains provide important regulation and supporting services. As azonal ecosystems, they are characterized by a rich biodiversity, but are highly endangered by human activities and climatic changes. Climatic changes will have the highest effect on alpine swamps and meadows, even their species composition is not completely known until now (according to Bernhard Dickoré participating in the Flora Tibetica) Assam subcatchment: The large wetland areas in the Brahmaputra Valley provide important provisioning, regulation and supporting services and a rich biodiversity the overall relative magnitude was evaluated from high to medium value. Most of the wetlands in the Valley are highly influenced by human activities since centuries. Rivers and their plains (alluvials) and the countless small oxbows, swamps and other waterlogged areas (beels) supply important resources for the local people. Although, like the seminatural/natural floodplains, they are highly vulnerable to increasing human activities and climatic changes. Wang Chu subcatchment: The mountainous regions in Bhutan cover only small wetland patches, which were evaluated high and low relative magnitude. Rivers and their plains (alluvials) and the small floodplain remnants provide important provisioning, regulating and supporting services. The diversity of the mountainous regions is huge, but mainly unexplored (especially some glacial lakes in the northern part of the subcatchment). The numerous beels in the lower part are highly endangered like in Assam. Upper Danube River Basin (UDRB): The wetlands in the Salzach and Lech subcatchments in the UDRB were evaluated from low to high relative magnitude. The azonal ecosystems are biodiversity hotspots. The floodplains provide, like in Asia, important regulating and supporting services. But most of the wetlands are highly altered by human activities and therefore reduced in their value and function. Despite the fact strategies like the Natura 2000 network or the EU Water Framework Directive will have positive effects on the endangered situation of the wetlands. 64 / 91

65 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands 5.4 Assesment of groundwater recharge Rationale and main outcome Groundwater resources have been assessed in regard to availablity, recharge and use. In order to study groundwater availability, one has to consider the different sources of recharge. For Assam, the river Brahmaputra and its tributaries are the main source of ground water recharge, as the flood plains in the vicinity of the rivers act as repositories for groundwater. To meet the additional demand during non monsoon periods, groundwater can be withdrawn from the flood plains. The exploited aquifer zones would then be recharged during the subsequent monsoon period through flood water. Aquifers near the surface are subject to annual recharge from precipitation, but the rate of recharge is affected by human interference. Deep aquifers, on the other hand, occur below layers of sedimentary rocks or in fault zones of the crystaline basement away from surface polution sources but are only recharged over long periods. Map Groundwater availability in Assam. Figures are given per district in km³/yr (source CGWB, Assam); please note that large symbols indicate low availability Methodology Input data sets According to government sources in Assam, the annual groundwater draft in the State corresponds to 5.44 billion cubic metres, from which 89% are diverted to irrigation, and the remainder is used for domestic and industrial consumption Method applied The aquifer sensitivity rating was based on the use of the DRASTIC model. DRASTIC stands for: 65 / 91

66 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands D Depth to water (difference between the well head elevation and that of the water level in the aquifer) R Net Recharge (amount of water that reaches the aquifer) A Aquifer Media (primary type of aquifer material) S Soil Media (primary type and size of soil particles) T Topography (the slope of the land surface) I Impact of the Vadose Zone (primary type and size of vadose zone material) C Hydraulic Conductivity (the ease at which water is able to move through the aquifer material) Parameter Weight Depth to groundwater 5 Net precipitation 4 Aquifer media 3 Soil media 2 Terrain slope 1 Impact of vadose zone 4 Hydraulic conductivity 2 Table 5 8: Weight of the parameters Main outcomes In general the region of Assam can be considered as a one with significantly high aquifer vulnerability. The main reason is a presence of very shallow aquifers all over the area, where the typical depth to groundwater is approximately up to 5 meters. In some regions the depth reaches 10 meters, but very rarely significantly deeper aquifers are exploited. Another major factor are the very high rates of groundwater recharge, counting up to 1350 mm/year. Although the prevailing part of soil is not characterized by relatively low permeability, the shallow aquifers of only a small part of the region are protected from downward leakage by fine clay soils. On the other hand, in the case of fine soils and shallow groundwater level the influence of the vadose zone is significantly growing. 66 / 91

67 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map Assam subcatchemt: Aquifer vulnerability 67 / 91

68 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Map Assam subcatchment: Estimated groundwater availability Achievements Accordingly to DRASTIC model analysis, the highest aquifer vulnerability has been determined in the districts of Nagaon, Karbi Anglong and in lower elevated parts of Sonitpur and Lakhimpur. The highest safety (lowest aquifer vulnerability) was detected in the mountain districts of Kokrajhar, Bongaigaon, Nalbari, Barpeta and higher parts of Karbi Anglong. 68 / 91

69 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands 5.5 Discharge Modelling Rationale and main outcome The figure below shows the location of the Lhasa and Koshi basins. In order to generate a discharge model for the Lhasa catchment, the Pitman model was first calibrated using rainfall runoff data from the Koshi Basin. The calibrated model parameters from the Koshi basin as well as climate data from the Lhasa basin were then used to generate flows in the Lhasa basin. The simulated monthly discharge curves can be used to understand rainfall runoff responses in the study catchments. Furthermore, the monthly flow volumes can be used for water accounting and management by comparing total water availability (simulated from the model) to water use/demand from the different water use sectors in the basin. Map The Koshi and Lhasa basins. The hatched area in the Koshi basin is the modelled upper subcatchment 69 / 91

70 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Methodology Input data Data set Source / Origin Resolution Remarks Landsat ETM+ Global Land Cover Facility 30x30m ASTER Advanced Spaceborne Thermal Emission and Reflection Radiometer) Populated areas global data set x15m Digital Surface Models (from 3.1) Quickbird Digital Globe ~0,6m Quickbird data has been acquired for test purposes in specific areas GPS measurements for verification and validation ICIMOD / FSU n/a Table 5 9: Input data sets used for Methods applied The SPATSIM (Spatial and Time Series Information Modeling) software was used in this study ( SPATSIM was developed at the Institute for Water Research (IWR) at Rhodes University, South Africa. It is an integrated data management and modelling software package developed in Delphi using the spatial data handling functions of Map Objects. SPATSIM includes several external model and data analysis tools, one of which is the Pitman Model. 70 / 91

71 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands Figure 9. Flow diagram of the main components of the SPATSIM version of the Pitman model The IWR (Rhodes University) semi distributed version of the Pitman model was applied in this study. This model is a conceptual model which runs on monthly values. Similar to many other conceptual models, the Pitman (1973) model consists of storages linked by functions designed to represent the main hydrological processes prevailing at the basin scale. Compulsory data requirements for the rainfall runoff model include catchment area, a time series of catchment average rainfall, a time series of potential evaporation, or an annual value and monthly distributions. Optional requirements include seasonal distributions of irrigation water requirements and other water abstractions, as well as time series of upstream inflow, transfer inflow and downstream compensation flow requirements. Figure 2 shows a flow diagram of the main components of the SPATSIM version of the Pitman model. The list of parameters for the rainfall runoff and a brief description is provided in the following table. Parameters RDF AI PI1s PI1w PI2s PI2w AFOR FF PEVAP ZMINs ZMINw Pitman model parameters description Rain Distribution Factor Proportion of impervious area Summer intercept storage(veg1) Winter intercept storage(veg1) Summer intercept storage (Veg2) Winter intercept storage (Veg2) % Area of Veg2 Veg2/Veg1 Pot. Evap. Ratio Annual Pan Evaporation (mm) Summer min.abs.rate (mm/mth) Winter min.abs.rate (mm/mth) 71 / 91

72 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands ZAVE ZMAX ST SL POW FT GW R TL GL AIRR IWR EFFECT Ml/year Mean abs.rate (mm/mth) Maximum abs.rate (mm/mth) Maximum storage capacity No runoff below storage Power : storage runoff curve Runoff rate at ST (mm/mth) Max. groundwater flow (mm/mth) Evaporation storage coefficient Surface runoff time lag (mnths) Groundwater time lag (months) Irrig.area (km^2) Irrig. return flow fraction Effective Rainfall fraction Non Irrig. Direct Demand Table 5 10: Pitman model parameter description Achievements Data sets produced The following figure shows the correspondence between observed and simulated flow for the upper Koshi sub catchment. The coefficient of determination (R 2 ) is 0.86 which shows very good results for the calibration. Figure 10. Model performance for the Koshi basin. Coefficient of determination: R2 = / 91

73 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands The observed vs. simulated monthly flows can be seen in the subsequent figure. The simulated hydrographs show that base flow is slightly overestimated. Further calibration of the model was not done as a very low groundwater recharge rate was used. Figure 11. Calibration results showing observed vs. simulated monthly flow volumes (m3) for the Koshi basin The next figure shows the simulated base flow component for the upper Koshi sub catchment. Over estimation of base flow could be due to groundwater extraction which was not included in the model. It can be seen that the stormflow component of the flows is more important than the base flow component. Figure 12. Baseflow analysis for the Koshi upper sub catchment The calibration results were thought to be acceptable for the model to be applied to the Lhasa basin. The simulated flow volumes (m 3 ) for the Lhasa basin are shown in the Figure below. The inter annual 73 / 91

74 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands variability in the discharges is primarily due to the variability in rainfall. Normally, with any modeling exercise, there is a need for validation. As data from the Lhasa basin is actually measured, it would add to the reliability of the results if the simulated flows could be validated with these data, in case they were officially available for even a limited time frame. Nevertheless, under the present circumstances, where measured data is unavailable, the model estimated monthly flows is very valuable for further studies that depend on information about the total water availability in the catchment. Figure 13. Simulated monthly flow volumes (m3) for the Outlet point of the Lhasa basin Project relevant achievements Predicting river flows in basins where limited data is available is a challenge facing many hydrologists. In this study, the Pitman monthly model was applied to generate flows for the Lhasa basin in China. As flow data was unavailable for the Lhasa basin, the model was first calibrated for the upper Koshi Basin in Nepal and China. The Pitman model successfully predicted flows for the upper Koshi basin (R 2 =0.86). Therefore, the estimated model parameters from the Koshi basin as well as climate data from the Lhasa basin were used to generate flows for the Lhasa basin outlet. The main modelling assumption is that the basin characteristics of the upper Koshi are similar to that of the Lhasa basin. Under present circumstances, where measured data is unavailable, the model estimated monthly flows for the Lhasa basin can be used in further studies in basin water accounting and management Results from the modelling study can be used by various other work packages in the future application of BrahmaTWinn project results: The simulated monthly total discharge can be used to cross check modelling results from the distributed hydrological modelling activity being carried out by Ludwig Maximilians University Munich, Germany The generated flows can be used in water accounting for the Lhasa basin As the basin is still under development, estimates can be used to calculate maximum possible extraction for irrigation, industry and domestic/urban water use Monthly environmental flows requirements can be estimated using the monthly discharges based on hydrological desktop methods 74 / 91

75 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands 5.6 Lessons learned By making use of different existing sources of public character and those provided by partners (especially ICIMOD), as well as applying expert opinion based geostatistical approaches for characterising locations with a potential for supporting wetlands (e.g. hydrological data, inclination information from the DTM) locations with wetlands located within the test sub basins could be determine. On a general level the vulnerability of wetlands could be derived from hydrological information. When trying to (i) locate more precisely the borders of wetlands and their seasonal variation, (ii) describe wetland vulnerability with respect to the water sources supporting the individual wetland types, (iii) select and identify the wetland ecotones and their associated ecosystem services, (iv) determine the importance of individual wetlands for sustaining a reasonable level of livelihood qualities for the surrounding population as well as its impact on different wetland types and locations, and (v) focus on the possible changes in wetland status due to climate change effects, the following lessons were learnt in the light of IWRM: As first prerequisite much more detailed information is needed on the local water regime of wetlands in several regions of the three test regions and, of course, in the whole UBRB in total. This set of information must be supplemented by a sufficiently dense net of recording stations in the rivers floodplains, unless all approaches directed in functional understanding of wetland reaction to hydrological and climate change induced changes cannot lead to object oriented actions. Additionally, in floodplains among the most important wetland types in the UBRB, even in higher altitudes much more detailed elevation models are needed to fully describe seasonal and inter annual effects of river floods, and groundwater, as the main water sources of this type of wetlands. Subtle elevation differences determine especially the existence, development or deterioration of ecotones, a major source of biological diversity and ecosystem services, as well as transitional zones of enhanced environmental vulnerability and a focus of population pressure. The importance of wetlands as a source of livelihood services and qualities is often not appreciated well enough by the local/regional population to find a priority order in their visions of population and land use development. When conflicts arise due to limitations in anthropogenic landuse types and strategies in rural and area planning, the reasonable control of population growth and increase in ecological footprint is usually the very last strategy to consider, as compared to technical at hand solutions or/and the enhanced sacrifice of ecotonal landscape elements and wetland areas. The final conclusion lies in the actual, real world integration of all relevant policy levels, decision making parties and stakeholders, necessary external/internal mediation processes, and all relevant scientific and economic expertise needed to save and sustain wetlands and their services throughout the UBRB, presently densely and less densely populated areas alike, despite the considerable difficulties to be expected. Local areas in the Assam reach are prone to groundwater vulnerability due to their hydrogeological characteristics. Detailed studies are needed to make policy makers and stakeholders aware of that threat and to take strategies for future amendment of that critical situation. 75 / 91

76 BRAHMATWINN > Assessment of the Natural Environment > Assessments of wetlands The study on discharge modelling of the Lhasa River demonstrates that the Pitman monthly model can be successfully applied to catchments in the study area. Although, there is some uncertainty in the generated flows for the Lhasa basin as validation was restricted by non available data, the modelling outputs are still very valuable in estimating total water availability for the catchment. The modelling results can be improved if measured values are available for the input parameters needed for the model. In addition to this specific application the results can trigger a process of assessing wetland vulnerability, when combined with remote sensing information on floodplain extension, dating from known periods of high discharge. This integrated information would forward sustainable policies directed to wetland conservation in the high altitude areas of the Lhasa catchment, as well as in other regions of the Upper Brahmaputra Basin. 76 / 91

77 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis 6 Deliverable 3.6 Hydrological systems analysis and delineation of Hydrological Response Units (HRU) 6.1 Lead partner and partners involved FSU (LMU, ZGIS, UniOs, all Asian partners) 6.2 Rationale and main outcome HRU delineations for all reference catchments within UBRB and UDRB. 6.3 Methodology Input data Summarizing the input data can be divided into two types of data, the data which describe the morphology of the basin and the spatial datasets which contain the pedo geological information of the area. Hydrological / morphological data DEM Slope Aspect Other geospatial information Land use / land cover Geology Soil Flow direction Flow accumulation River network Delineation of sub basins Table 6 1: Input data sets used Methods applied The hydrological system analysis is the base for the delineation of the Hydrological response units (HRUs). HRUs are treated as model entities in the process oriented hydrological models PRMS and J2000. Initially, the hydrological system analysis studies the interactions of the landscape parameters soil, water, vegetation and climate in order to understand the system response to rainfall and therefore the generation of the respective hydrological response. There are different concepts available to represent entities showing a similar or equal system response. Following some concepts were described. The concept of Aggregated Simulation Area (ASA) was applied the Semi distributed Land Use based Runoff Processes model (SLURP). This concept involves aggregating simulation areas which are heterogeneous in their land cover and elevation. These ASA works as a kind of sub catchment and require of contributing runoff to a stream channel. The Representative Elementary Area (REA) defines minimal areas in which the spatial heterogeneity of hydrological variables like infiltration, evapotranspiration and runoff are unimportant. The distribution of these variables within the areas is represented by a probability function. 77 / 91

78 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis The concept of the Hydrological Response Units was introduced as model entities in the Precipitation Runoff Modelling System (PRMS). They are characterized as homogenous areas with respect to their hydrological response or as model entities that show a hydrological similarity. HRUs are defined as distributed spatial model entities which respond similar in their hydrological process dynamics to a given input and can be seen as part of the hydrological system analysis. Based on the results of a hydrological system analysis the GIS datasets are reclassified, aggregated and overlaid in a step bystep procedure. To ensure the hydrological flow between HRUs and the stream network, the flow routing is determined (JAMS). The attributes for each HRU are given by the class (combined information) of the datasets. After the analysis of the driving forces of the system the delineation of the HRUs takes place. In the first steps the different datasets will be resampled to the same spatial resolution and will be combined to new classes. After that the classes will be reclassified to eliminate unrealistic classes and to conserve substantial spatial classes that they cannot be joined to other subclasses. In our case the land use class urban area will not be divided into different slopes or expositions. Following the created classes will be separated into single areas, the HRUs. The attributes are contained in the corresponding class information. The HRUs are delineated based on detailed hydrological system analyses in a GIS. To delineate hydrological response units different steps are necessary. In a first step the data have to be corrected and the morphological information has to be delineated from the DEM, the data will be classified knowledge based into hydrological important classes and the nominal scaled datasets will be reclassified into classes with hydrological relevance. The reclassified datasets were combined in the GIS by using overlay functions. Finally a set of many different small areas emerges. These areas have similar characteristics within their border, but they are different to their neighbouring areas. After the final step of eliminating the smallest polygons and joining the attributes, the final hydrological response units are delineated. In the next passages in a detailed description the HRU delineation is shown for different basins within the BRAHMATWINN project. Before the elimination of small polygons, classes were defined which were not eliminated. These are small classes which have a significant impact on hydrological processes. For this reason these classes, e.g. urban areas have to be maintained. Figure 14. Schema of HRU delineation Preprocessing of the data. To use the overlay function it is fundamental that the data fit together. Each pixel of a dataset one has to lie exactly on the pixel of the dataset two. This requires that all datasets underlie the same projection and that they have the same resolution. The digital elevation 78 / 91

79 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis model has to be checked for errors and if necessary corrected. The derivates of the DEM slope, aspect, flow direction, flow accumulation and stream network have to be calculated using standard methods which are integrated in the ArcGIS. Reclassify datasets. The input data can be divided into two groups of data, metric scaled data like the slope, aspect and DEM and the nominal scaled information like land cover, geology and soil. In a first step the metric scaled data will be reclassified into classes, for example the aspect information is reclassified in the three classes north, south and west/east. By this way the metric scaled data will be converted into a nominal scale. The class range will be selected according to the hydrological relevance, so it can be neglected if a hill slope has an east or a west aspect because the radiation is closely equal. The spatial information land cover, geology and soil will be reclassified into classes which have a specific hydrological response. In this step similar information will be summarized in one class. The for the HRU delineation used classes will be defined for each basin and test area separately. The classes are based on the basin heterogeneity and the distinctions of the basins. Combination of datasets. Based on the results of the hydrological system analysis the GIS datasets are reclassified, aggregated and overlaid in a step by step procedure. The combination of two datasets can be done by two methods. The fiction is that we will need double digit number to cover all entities of each dataset. In the next step all datasets will be added together and the value presents the characteristics of the dataset. The final number (GRIDCODE) contains the corresponding class information of each layer. The combination of the datasets will be done in a step by step action. In a first step two layers will be combined. This new dataset will be analyzed how the data are distributed and small classes, which have an extent less than one percentage of the whole basin will be eliminated by adding to a related larger class. In the next iteration the next dataset will be added to this reclassified dataset, the same analyzes will be done, but the area threshold will be 0.5 percentage of the basin. The threshold can be chosen knowledge based on basis of catchment characteristics. Separate groups into single shapes. In the steps before all datasets were combined and reclassified to a dataset containing the class information. In this step the classes will divided into single areas with a separate ID for each unit. The used function in ArcGIS is the regiongroup command. The output of separating single areas will be a shape file. Eliminating smallest polygons. Splitter polygons will be eliminated in the last step of the HRU delineation. Polygons with an area less than 4 pixels so called splitter polygons have a low hydrological importance. Additionally these small polygons cause high computing times during a hydrological model run. When the smallest polygons are eliminated, the result of a hydrological model will not change, because they have no effect on the final result. The elimination will be done stepwise. In a first step all polygons less than an area of 1 pixel will be eliminated. In the next step the polygons with an area less than 2 pixels will be eliminated. The elimination process will join the areas which shall be eliminated to the neighbor polygon with the longest border. Despite they are small in size polygons with the land use class wetland and urban area were not eliminated because they have a high importance for the hydrological system and have to be preserved. HRU delineation in the Salzach river reference catchment: The HRU delineation was done in a 90 m resolution. 79 / 91

80 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis The DGM shows the lowest areas and highest peaks in the basin. The highest elevations exist in the southern part. The northern part is characterized as a lower basin. In a first step of the HRU delineation the layers slope and landuse were combined. In respect to the generalization process, land use classes are defined which have to be maintained because of their high hydrological relevance. These classes were not aggregated: 1000, 6000, 7000, 8000 and Agriculture areas can be found especially in the northern lower parts of the basin and along the river streams. Forest covers large parts in the middle part and the high mountain range is particularly covered with grassland, ice and snow. The soil information was used from the Hydrological Atlas Austria (HAÖ) and the FAO soil information. The northern part is mainly covered by Cambisols. The middle part is characterized by Rendzina and Immature soil. The southern part shows large areas covered by Podsol, but also Cambisol and Immature soil are existing. Map 6.1. Soil type map of the Salzach river reference catchment HRU delineation in the Lech river reference catchment: The HRU delineation was done within a 250m resolution. The 90m SRTM DGM was resampled to 250m by using the cubic convolution method. The landuse map shows that the land use classes are widely spread within the river basin. Agriculture dominates more in the southern area, while grassland dominates in the northern part. Bare soil is only detectable in the south. Forests can be found in the entire basin. The soil information was made available by the FAO/ WRB World Reference Base for Soil Resources in a 1000m resolution. The soil information was resampled to a resolution of 250m with nearest neighbor method. The soil map shows that major part of the basin and 80 / 91

81 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis the entire southern area is covered by Leptosols. Cambisols and Luvisols exist northern part; Fluvisol in the central region. only in the HRU delineation in the Lhasaa river reference catchment: The Lhasa river basin is a glacier fed catchment located on the Tibetan Plateau. The topographic gradient reaches from 3.500m up to 7.000m. First step of HRU delineation was the preparation of the original data. All dataa were resampled in a resolution of 500m. In a second step reclassified property layers were combined in a specific order to keep the hydrological functionn and to take care that the HRUs are characterized by their homogeneity and behavior in respect to hydrological processes. In a final step small splitter polygons weree aggregated to minimize the absolute number of HRUs. The slope layer was reclassified in threee classes. Each class represents one runofff component. So an area with a slope smaller than 5 degrees contributes to groundwater r recharge. In comparison on an area with a slope of more than 15 degrees most of the runofff will flow on the surface. 16% 62% 22% >15 Figure 15. Slope class distribution in the Lhasa river catchment Most areas have steep slopes of more than 15 degrees. The figure shows also, that the riverr stream and the trench in the north west are significant represented in the class with zero or minimal slope. The aspect information is delineated from the SRTM DGM. In a result the exposition of slopess is made visible. This exposition influences different climate parameters like radiation and temperature. Slopes with a southern exposition receivee a stronger radiation, temperature e are higher and following evaporation and snow melt processes will be accelerated in comparison to north exposed slopes. Different investigations have shown, that the classes east and west can be summarized, because they receivee the same radiation over one day and have thereby same impact on hydrological processes like evaporation and snow melt. Most of the catchment is covered with grassland. The river stream and neighboringg agriculture areas become obviously, also the snow and ice fields in the high mountain ranges in the north west and the eastern area. The soil information was made available by the Chinese project partners. 7 soil types could be identified in the Lhasa catchment. Becausee of similar properties, (see table soil properties) 81 / 91

82 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis the soil classes 40 to 10 were summarized as well as the class 50 was added to 60. This reclassification limits the total number of HRU classes and calculation expenses. Map 6.2. Reclassified soil information in the Lhasa river catchment HRU delineation in the Wang Chu river reference catchment: The SRTM in its original resolution of 90m serves as basis for the delineation of the flow accumulation, slope, aspect etc. Following the data for the HRU delineation are represented and the process of the delineation is described. The SRTM is available in a 90m resolution. This DEM was corrected from errors, sinks were filled etc. The information for slope and aspect was delineated from the SRTM in 90m resolution and was then resampled to 250m resolution. The digital elevation model shows the topographic conditions of the Wang Chu catchment. The lowest areas with about 140 m above sea level are in the south, while the highest elevations up to 6700m were reached in the north. The green valley s visualize the stream network in the catchment. The slope information, delineated from the 90m SRTM was reclassified in three classes: low slopes until 5, middle spread and steep slopes. The major part of the catchment is characterized by steep slopes with more than 15. The percentage of the class 0 5 is very low, only the stream network is weakly visible. Knowledge based reclassification of MODIS land use data has given 10 land use classes. The classes and their percentages are listed below. Largest parts of the area are covered by deciduous forest and coniferous forest. Nearly 20% are covered by grassland. Snow and glacier have only a small percentage on the land use distribution in the Wang Chu catchment. 82 / 91

83 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis Agriculture is widely spread along the river streams. Snow and glacier fields can be found in the high mountain range in the north. Forests dominate nearly the whole catchment without the high mountain range; here the area is mostly covered with grassland. The aspect information is calculated from the SRTM. The table below shows the classes. The figure shows the distribution of the classes within the Wang Chu catchment. The class distribution is well balanced. The soil information was made available by the FAO. The soil information was reclassified in 3 classes, based upon the soil properties. By this way Leptosols and Regosols were summarized in one class, as well as Acrisols, Alisols and Plinthosols. The division of soil typed is very unique. While the Cambisols are only present at the southern tip of the basin, the class 96 (Acrisols) covers the whole southern part and the class 33 (Leptosols and Regosols) the entire northern part of the basin. HRU delineation in the Assam Brahmaputra reference catchment: The topographic gradient is much lower than in the previous regions. Highest elevations are reached in the south with a maximal altitude of 2000m. Greatest areas of Assam are low, because it is a plain area characterized by the wide spread Brahmaputra River. In consequence of the topographic gradient, the slope was reclassified in the two classes 0 2 and <2. The landuse data were extracted from the MODIS dataset and reclassified from original 16 classes in 9 classes in the map below. Large parts of the area are characterized due to arable land. Forests dominate the southern area along the middle mountain ranges. 83 / 91

84 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis 6.4 Achievements Map showcase Map 6.3. HRU delineation in the Upper Danube River Basin 84 / 91

85 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis Map 6.4. HRUs in the Salzach reference catchment Map 6.5. HRUs in the Lech reference catchment 85 / 91

86 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis Map 6.6. HRUs in the Lhasa river reference catchment Map 6.7. HRUs in the Wang Chu reference catchment 86 / 91

87 BRAHMATWINN > Assessment of the Natural Environment > Hydrological systems analysis Map 6.8. HRUs in the Brahmaputra in Assam reference catchment Data sets produced Data set Remarks HRUs for the Lhasa river reference catchment HRUs for the Wang Chu river reference catchment HRUs for the Brahmaputra in Assam reference catchment HRUs for the Salzach river reference catchment HRUs for the Lech river reference catchment Table 6 2: Data sets produced in Project relevant achievements The following table lists the number of HRUs delineated and the number of classes produced. Reference catchment # of HRUs # of classes Salzach Lech Lhasa Wang Chu n/a Assam n/a 87 / 91