CHARACTERIZING SPATIAL AND TEMPORAL VARIATIONS OF ...
CHARACTERIZING SPATIAL AND TEMPORAL VARIATIONS OF SURFACE TEMPERATURE OF LAKE TANA (ETHIOPIA) USING MODIS DATA Zheng Duan1*, W.G.M Bastiaanssen1,2 1
2
Delft University of Technology, Stevinweg 1, 2628 CN Delft, The Netherlands eLEAF Competence Center, Generaal Foulkesweg 28, 6703 BS Wageningen, The Netherlands * Corresponding author. E-mail address: [email protected] ABSTRACT
The information of water surface temperature (WST) is essential for many applications such as the estimation of evaporation from lakes and reservoirs. Generally, there is rarely continuous long-term in-situ measurement of water surface temperature for most lakes especially in developing countries. The freely available satellite data is an alternative or even the only data source for obtaining WST over a given lake. This study characterized the spatial and temporal variations of WST over Lake Tana (Ethiopia) using the Version 5 Terra Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature product, namely, MOD11_L2. The MOD11_L2 data during the year 2006 were selected. Twelve maps of monthly mean WST over Lake Tana were generated. The lake-wide mean WST (L-WST) were further computed by averaging all pixels of WST to construct the time-series of monthly LWST. The spatial variations of monthly WST value can range from 1.73 °C in November to 5.42 °C in July. The LWST ranges from 21.27 °C in December to 24.86 °C in May. Index Terms— MODIS, land surface temperature, lake and reservoir, water temperature, Lake Tana 1. INTRODUCTION Water surface temperature in lakes is an important indicator of lake conditions [1]. The information of lake surface temperature is essential for many applications, such as water quality management, numerical weather prediction [2] and studies involving the energy/heat and water fluxes [3, 4] such as evaporation estimation. Such lake surface temperature data are conventionally obtained through conducting limnological sampling for in-situ measurements. However, in-situ measurements of the water temperature are often limited to several point-based places along the lake shorelines and for a non-continuous and short time period. For many lakes there are even no in-situ measurements of water temperature at all. This is the case in particular for remote lakes and ones located in the developing countries.
The lack of such in-situ measurements is a big obstacle for many studies requiring the lake surface temperature. Recently satellite thermal imagery data have been used to derive the surface temperatures of many lakes [1]. The unique advantages of satellite observation lie in the ability to provide the spatial distributions of lake surface temperature at nominal consistent and continuous temporal intervals. Several satellite-derived surface temperature products (e.g. Moderate Resolution Imaging Spectroradiometer (MODIS) land-surface temperature product) have already been generated and put online for freely downloading and further applications. Furthermore, validation studies have shown that such satellite-derived lake surface temperature has reasonable accuracy with a bias of about 1.5 [1] or 2 k/°C [5] when compared to the in-situ lake surface temperature measured by surface buoys over the lake. Lake Tana is the largest lake in Ethiopia. It is the source of the Blue Nile River and the Blue Nile River contributes more than 60% of total flow into the Nile River at Aswan in Egypt [6]. Hence, Lake Tana is the very important water resources for peoples living alongside Blue Nile River and Nile River. However, there is no continuous long-term insitu measurement of lake surface temperature in Lake Tana, which dampers better understanding of lake state and other studies in which surface temperature is needed. The satellite data appear to be the only data source for obtaining lake surface temperature of Lake Tana. The objective of this study is to use the MODIS data to characterize the spatial and temporal variations of surface temperature of Lake Tana. 2. STUDY AREA Lake Tana (between latitude 11°35′-12°18′ N and longitude 37°00′-37°38′ E) is located in the northwestern highlands of Ethiopia. The surface water area of Lake Tana ranges from 2966 to around 3100 km2 according to the seasonal fluctuation in lake level [7]. Lake Tana has a 385 m shoreline. The maximum depth is 14 m and the mean depth is 9 m. The water storage is about 28 km3. The climate in
this region is characterized by a mean annual precipitation 1395 mm with a rainy season (June-September) and a dry season (October-March) [8].The air temperature is characterized by a large diurnal but small seasonal variability, and the mean annual air temperature is 20 °C [9]. 3. DATASETS AND METHODS MODIS land surface temperature (LST) level 2, 1-km nominal resolution data were used in this study. MODIS LST product is generated from two MODIS thermal bands 31(10.780-11.280 µm) and 32 (11.770-12.270 µm) using a split-window algorithm designed for a range of land cover types including inland water surfaces/bodies [1, 10]. The Version 5 Terra MODIS product MOD11_L2 (i.e. Level 2 5-min land surface temperature and Emissivity -1km) were obtained through NASA's Goddard Space Flight Center LAADS Web, available at http://ladsweb.nascom.nasa.gov/data/. LAADS Web can provide several practical post-processing services such as data reformatting (e.g. converting to GeoTIFF), reporjection, geographical subsetting and mosaicing. The rectangle box with latitude of 11.4-12.4°N and longitude of 36.8-37.8°E was defined to spatially search the data covering the whole Lake Tana. The year 2006 was selected and a total of 889 images were found. The nominal acquisition times of these 889 images covering Lake Tana are about 7-9 am UTC during daytime or 7-9 pm UTC during nighttime. Several post-processing services provided by LAAD Web were used: (1) reprojecting to UTM Zone 37N with resample type as Nearest; (2) converting data to the GeoTIFF file format. The calibration coefficients (scale factor and add offset) were used to convert the initial pixel values in MOD11_L2 product to the corresponding real physical values (e.g. LST value in the unit of Kelvin). The calibration coefficients are given in the MODIS LST products Users’ Guide [10]. It should be noted that there can be a cold bias introduced to the satellite-derived surface temperature when at higher view angles, thus [1] only used MODIS LST data with satellite zenith angles less than 56° to mitigate the possible cold bias effect. Similarly, we also excluded the LST pixel values with satellite zenith angles greater than 56° for further analysis in this study. In addition, the quality control (QC) data included in MOD11_L2 were also used to select the usefulness of the corresponding LST data. Only LST pixel with a value of QC equal to 0 was used for further analysis. It should be noted that the value of 0 for QC means good data quality only when this pixel has a valid LST value according to the users’ guide [10]. Hence, the further analysis of LST value was conducted to remove pixels with invalid values. Lake Tana has seasonal fluctuation in water levels correspondingly leading to variations in surface areas and
shorelines. Therefore, the shoreline of Lake Tana at a low water level is better to be generated firstly to mask the selected MODIS LST products. This mask can help us to extract all MODIS temperature values which are solely from water surface. The lowest water level in Lake Tana was observed in June 2003, and the most coincident Landsat TM image identified was acquired on May 18, 2003. This image has already been processed using a Modified Normalized Difference Water Index (MNDWI) to delineate the shoreline of Lake Tana representing the situation at a relatively low level in our previous study [7]. Initial analysis showed that there were still several land or land-contaminated surface temperature pixels included when using this shoreline as the mask. Hence, we made an inner 2-km (corresponding to two MODIS LST pixels) wide buffer zone of the shoreline derived from TM image. Based on this buffer zone, a smaller shoreline of Lake Tana was generated and used as the final mask to extract MODIS temperature data. For the whole year 2006, all processed images were collected and overlaid for a given month (e.g. January); then the pixel-bypixel mean value was computed to obtain a map of monthly mean water surface temperature (WST) for this month. The histogram of the computed monthly mean WST map was checked, and the pixels with abnormally low and high WST values were removed. The WST values in unit of Kelvin were converted to the unit of °C by subtracting a constant of 273.15. These processing procedures were repeated for each individual month. Finally, twelve maps of monthly mean WST were generated for the year 2006 to study spatial and temporal variations of WST of Lake Tana. 4. RESULTS AND DISCUSSION All twelve monthly mean WST maps of Lake Tana for the year 2006 are shown in Figure 1. The scattered white pixels within WST maps refer to the unreliable temperature values which were eliminated through histogram analysis. There are relatively large spatial variations in WST, the difference between the minimum and maximum WST value can range from 1.73 °C in November to 5.42 °C in July (see the legend bar included in Figure 1). The spatial pattern of WST also varies from month to month. In general, higher WST occurs near the shoreline of lake while lower WST often occurs in the center and northwest part of Lake Tana. The lake-wide mean temperature was computed for each month by averaging all pixels within Lake Tana. Figure 2 shows time-series of monthly lake-wide mean WST (L-WST) over Lake Tana for the year 2006. The L-WST ranges from 21.27 °C occurred in December to 24.86 °C in May. The temporal range is 3.59 °C. There was little difference in LWST observed in dry months January-March. From April, the L-WST started to fluctuate. During the rainy periods
Figure 1. MODIS-derived monthly mean water surface temperature of Lake Tana for the year 2006. (June-September), there were large variations in L-WST. This large variation could be due to the effect of rain and corresponding large river inflow into Lake Tana. The LWST values in July and September were about 1.5-2.0 ℃ higher than June and August. There was a consistent decreasing trend in L-WST between October and December.
Figure 2. MODIS-derived monthly lake-wide mean water surface temperature of Lake Tana for the year 2006.
5. CONCLUSIONS The lack of continuous long-term in-situ measurements of water surface temperature (WST) is a big obstacle for better understanding of Lake Tana (Ethiopia) which is the important water resources for countries alongside the Blue Nile River and Nile River. Therefore, using freely available and continuous satellite data could be the only way to characterizing the spatial and temporal variations of water surface WST of Lake Tana. In the study, twelve maps of monthly mean WST of Lake Tana were derived from Version 5 Terra MODIS Land Surface Temperature product (i.e., MOD11_L2) for the year 2006. The spatial variations of monthly WST value can range from 1.73 °C in November to 5.42 °C in July. In general, higher WST often occurs near the shoreline of lake while lower WST often occurs in the center and northwest part of Lake Tana. The annual cycle of lake-wide mean WST (L-WST) exhibits a small variation (3.59 °C) with L-WST ranging from 21.27 °C in December to 24.86 °C in May. 6. ACKNOWLEDGMENT The author Zheng Duan would like to thank China Scholarship Council (CSC) for providing financial support for a PhD study at Delft University of Technology (TU
Delft), The Netherlands. Zheng Duan also acknowledges financial supports from the Lamminga Fonds (TU Delft) and IGARSS 2013 Travel Grant to attend the IGARSS 2013 in Australia. 7. REFERENCES [1] E. T. Crosman and J. D. Horel, "MODIS-derived surface temperature of the Great Salt Lake," Remote Sensing of Environment, vol. 113, pp. 73-81, Jan 15 2009. [2] D. Oesch, J. M. Jaquet, R. Klaus, and P. Schenker, "Multi-scale thermal pattern monitoring of a large lake (Lake Geneva) using a multi-sensor approach," International Journal of Remote Sensing, vol. 29, pp. 5785-5808, 2008. [3] E. H. Alcantara, J. L. Stech, J. A. Lorenzzetti, M. P. Bonnet, X. Casamitjana, A. T. Assireu, and E. M. L. D. Novo, "Remote sensing of water surface temperature and heat flux over a tropical hydroelectric reservoir," Remote Sensing of Environment, vol. 114, pp. 2651-2665, Nov 15 2010. [4] Z. K. Xing, D. A. Fong, K. M. Tan, E. Y. M. Lo, and S. G. Monismith, "Water and heat budgets of a shallow tropical reservoir," Water Resources Research, vol. 48, Jun 26 2012. [5] D. C. Oesch, J. M. Jaquet, A. Hauser, and S. Wunderle, "Lake surface water temperature retrieval using advanced very high resolution radiometer and Moderate Resolution Imaging Spectroradiometer data: Validation and feasibility study," Journal of Geophysical Research-Oceans, vol. 110, Dec 15 2005. [6] S. Uhlenbrook, Y. Mohamed, and A. S. Gragne, "Analyzing catchment behavior through catchment modeling in the Gilgel Abay, Upper Blue Nile River Basin, Ethiopia," Hydrology and Earth System Sciences, vol. 14, pp. 2153-2165, 2010. [7] Z. Duan and W. G. M. Bastiaanssen, "Estimating water volume variations in lakes and reservoirs from four operational satellite altimetry databases and satellite imagery data," Remote Sensing of Environment, vol. 134, pp. 403–416, Jul 2013. [8] Z. Duan and W. G. M. Bastiaanssen, "First results from Version 7 TRMM 3B43 precipitation product in combination with a new downscaling-calibration procedure," Remote Sensing of Environment, vol. 131, pp. 1-13, Apr 15 2013. [9] S. G. Setegn, D. Rayner, A. M. Melesse, B. Dargahi, and R. Srinivasan, "Impact of climate change on the hydroclimatology of Lake Tana Basin, Ethiopia," Water Resources Research, vol. 47, Apr 19 2011. [10] Z. Wan, "Collection-5 MODIS Land Surface Temperature products Users’ Guide", available at: http://www.icess.ucsb.edu/modis/LstUsrGuide/MODIS_LST_prod ucts_Users_guide_C5.pdf, 2007.
2
Delft University of Technology, Stevinweg 1, 2628 CN Delft, The Netherlands eLEAF Competence Center, Generaal Foulkesweg 28, 6703 BS Wageningen, The Netherlands * Corresponding author. E-mail address: [email protected] ABSTRACT
The information of water surface temperature (WST) is essential for many applications such as the estimation of evaporation from lakes and reservoirs. Generally, there is rarely continuous long-term in-situ measurement of water surface temperature for most lakes especially in developing countries. The freely available satellite data is an alternative or even the only data source for obtaining WST over a given lake. This study characterized the spatial and temporal variations of WST over Lake Tana (Ethiopia) using the Version 5 Terra Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature product, namely, MOD11_L2. The MOD11_L2 data during the year 2006 were selected. Twelve maps of monthly mean WST over Lake Tana were generated. The lake-wide mean WST (L-WST) were further computed by averaging all pixels of WST to construct the time-series of monthly LWST. The spatial variations of monthly WST value can range from 1.73 °C in November to 5.42 °C in July. The LWST ranges from 21.27 °C in December to 24.86 °C in May. Index Terms— MODIS, land surface temperature, lake and reservoir, water temperature, Lake Tana 1. INTRODUCTION Water surface temperature in lakes is an important indicator of lake conditions [1]. The information of lake surface temperature is essential for many applications, such as water quality management, numerical weather prediction [2] and studies involving the energy/heat and water fluxes [3, 4] such as evaporation estimation. Such lake surface temperature data are conventionally obtained through conducting limnological sampling for in-situ measurements. However, in-situ measurements of the water temperature are often limited to several point-based places along the lake shorelines and for a non-continuous and short time period. For many lakes there are even no in-situ measurements of water temperature at all. This is the case in particular for remote lakes and ones located in the developing countries.
The lack of such in-situ measurements is a big obstacle for many studies requiring the lake surface temperature. Recently satellite thermal imagery data have been used to derive the surface temperatures of many lakes [1]. The unique advantages of satellite observation lie in the ability to provide the spatial distributions of lake surface temperature at nominal consistent and continuous temporal intervals. Several satellite-derived surface temperature products (e.g. Moderate Resolution Imaging Spectroradiometer (MODIS) land-surface temperature product) have already been generated and put online for freely downloading and further applications. Furthermore, validation studies have shown that such satellite-derived lake surface temperature has reasonable accuracy with a bias of about 1.5 [1] or 2 k/°C [5] when compared to the in-situ lake surface temperature measured by surface buoys over the lake. Lake Tana is the largest lake in Ethiopia. It is the source of the Blue Nile River and the Blue Nile River contributes more than 60% of total flow into the Nile River at Aswan in Egypt [6]. Hence, Lake Tana is the very important water resources for peoples living alongside Blue Nile River and Nile River. However, there is no continuous long-term insitu measurement of lake surface temperature in Lake Tana, which dampers better understanding of lake state and other studies in which surface temperature is needed. The satellite data appear to be the only data source for obtaining lake surface temperature of Lake Tana. The objective of this study is to use the MODIS data to characterize the spatial and temporal variations of surface temperature of Lake Tana. 2. STUDY AREA Lake Tana (between latitude 11°35′-12°18′ N and longitude 37°00′-37°38′ E) is located in the northwestern highlands of Ethiopia. The surface water area of Lake Tana ranges from 2966 to around 3100 km2 according to the seasonal fluctuation in lake level [7]. Lake Tana has a 385 m shoreline. The maximum depth is 14 m and the mean depth is 9 m. The water storage is about 28 km3. The climate in
this region is characterized by a mean annual precipitation 1395 mm with a rainy season (June-September) and a dry season (October-March) [8].The air temperature is characterized by a large diurnal but small seasonal variability, and the mean annual air temperature is 20 °C [9]. 3. DATASETS AND METHODS MODIS land surface temperature (LST) level 2, 1-km nominal resolution data were used in this study. MODIS LST product is generated from two MODIS thermal bands 31(10.780-11.280 µm) and 32 (11.770-12.270 µm) using a split-window algorithm designed for a range of land cover types including inland water surfaces/bodies [1, 10]. The Version 5 Terra MODIS product MOD11_L2 (i.e. Level 2 5-min land surface temperature and Emissivity -1km) were obtained through NASA's Goddard Space Flight Center LAADS Web, available at http://ladsweb.nascom.nasa.gov/data/. LAADS Web can provide several practical post-processing services such as data reformatting (e.g. converting to GeoTIFF), reporjection, geographical subsetting and mosaicing. The rectangle box with latitude of 11.4-12.4°N and longitude of 36.8-37.8°E was defined to spatially search the data covering the whole Lake Tana. The year 2006 was selected and a total of 889 images were found. The nominal acquisition times of these 889 images covering Lake Tana are about 7-9 am UTC during daytime or 7-9 pm UTC during nighttime. Several post-processing services provided by LAAD Web were used: (1) reprojecting to UTM Zone 37N with resample type as Nearest; (2) converting data to the GeoTIFF file format. The calibration coefficients (scale factor and add offset) were used to convert the initial pixel values in MOD11_L2 product to the corresponding real physical values (e.g. LST value in the unit of Kelvin). The calibration coefficients are given in the MODIS LST products Users’ Guide [10]. It should be noted that there can be a cold bias introduced to the satellite-derived surface temperature when at higher view angles, thus [1] only used MODIS LST data with satellite zenith angles less than 56° to mitigate the possible cold bias effect. Similarly, we also excluded the LST pixel values with satellite zenith angles greater than 56° for further analysis in this study. In addition, the quality control (QC) data included in MOD11_L2 were also used to select the usefulness of the corresponding LST data. Only LST pixel with a value of QC equal to 0 was used for further analysis. It should be noted that the value of 0 for QC means good data quality only when this pixel has a valid LST value according to the users’ guide [10]. Hence, the further analysis of LST value was conducted to remove pixels with invalid values. Lake Tana has seasonal fluctuation in water levels correspondingly leading to variations in surface areas and
shorelines. Therefore, the shoreline of Lake Tana at a low water level is better to be generated firstly to mask the selected MODIS LST products. This mask can help us to extract all MODIS temperature values which are solely from water surface. The lowest water level in Lake Tana was observed in June 2003, and the most coincident Landsat TM image identified was acquired on May 18, 2003. This image has already been processed using a Modified Normalized Difference Water Index (MNDWI) to delineate the shoreline of Lake Tana representing the situation at a relatively low level in our previous study [7]. Initial analysis showed that there were still several land or land-contaminated surface temperature pixels included when using this shoreline as the mask. Hence, we made an inner 2-km (corresponding to two MODIS LST pixels) wide buffer zone of the shoreline derived from TM image. Based on this buffer zone, a smaller shoreline of Lake Tana was generated and used as the final mask to extract MODIS temperature data. For the whole year 2006, all processed images were collected and overlaid for a given month (e.g. January); then the pixel-bypixel mean value was computed to obtain a map of monthly mean water surface temperature (WST) for this month. The histogram of the computed monthly mean WST map was checked, and the pixels with abnormally low and high WST values were removed. The WST values in unit of Kelvin were converted to the unit of °C by subtracting a constant of 273.15. These processing procedures were repeated for each individual month. Finally, twelve maps of monthly mean WST were generated for the year 2006 to study spatial and temporal variations of WST of Lake Tana. 4. RESULTS AND DISCUSSION All twelve monthly mean WST maps of Lake Tana for the year 2006 are shown in Figure 1. The scattered white pixels within WST maps refer to the unreliable temperature values which were eliminated through histogram analysis. There are relatively large spatial variations in WST, the difference between the minimum and maximum WST value can range from 1.73 °C in November to 5.42 °C in July (see the legend bar included in Figure 1). The spatial pattern of WST also varies from month to month. In general, higher WST occurs near the shoreline of lake while lower WST often occurs in the center and northwest part of Lake Tana. The lake-wide mean temperature was computed for each month by averaging all pixels within Lake Tana. Figure 2 shows time-series of monthly lake-wide mean WST (L-WST) over Lake Tana for the year 2006. The L-WST ranges from 21.27 °C occurred in December to 24.86 °C in May. The temporal range is 3.59 °C. There was little difference in LWST observed in dry months January-March. From April, the L-WST started to fluctuate. During the rainy periods
Figure 1. MODIS-derived monthly mean water surface temperature of Lake Tana for the year 2006. (June-September), there were large variations in L-WST. This large variation could be due to the effect of rain and corresponding large river inflow into Lake Tana. The LWST values in July and September were about 1.5-2.0 ℃ higher than June and August. There was a consistent decreasing trend in L-WST between October and December.
Figure 2. MODIS-derived monthly lake-wide mean water surface temperature of Lake Tana for the year 2006.
5. CONCLUSIONS The lack of continuous long-term in-situ measurements of water surface temperature (WST) is a big obstacle for better understanding of Lake Tana (Ethiopia) which is the important water resources for countries alongside the Blue Nile River and Nile River. Therefore, using freely available and continuous satellite data could be the only way to characterizing the spatial and temporal variations of water surface WST of Lake Tana. In the study, twelve maps of monthly mean WST of Lake Tana were derived from Version 5 Terra MODIS Land Surface Temperature product (i.e., MOD11_L2) for the year 2006. The spatial variations of monthly WST value can range from 1.73 °C in November to 5.42 °C in July. In general, higher WST often occurs near the shoreline of lake while lower WST often occurs in the center and northwest part of Lake Tana. The annual cycle of lake-wide mean WST (L-WST) exhibits a small variation (3.59 °C) with L-WST ranging from 21.27 °C in December to 24.86 °C in May. 6. ACKNOWLEDGMENT The author Zheng Duan would like to thank China Scholarship Council (CSC) for providing financial support for a PhD study at Delft University of Technology (TU
Delft), The Netherlands. Zheng Duan also acknowledges financial supports from the Lamminga Fonds (TU Delft) and IGARSS 2013 Travel Grant to attend the IGARSS 2013 in Australia. 7. REFERENCES [1] E. T. Crosman and J. D. Horel, "MODIS-derived surface temperature of the Great Salt Lake," Remote Sensing of Environment, vol. 113, pp. 73-81, Jan 15 2009. [2] D. Oesch, J. M. Jaquet, R. Klaus, and P. Schenker, "Multi-scale thermal pattern monitoring of a large lake (Lake Geneva) using a multi-sensor approach," International Journal of Remote Sensing, vol. 29, pp. 5785-5808, 2008. [3] E. H. Alcantara, J. L. Stech, J. A. Lorenzzetti, M. P. Bonnet, X. Casamitjana, A. T. Assireu, and E. M. L. D. Novo, "Remote sensing of water surface temperature and heat flux over a tropical hydroelectric reservoir," Remote Sensing of Environment, vol. 114, pp. 2651-2665, Nov 15 2010. [4] Z. K. Xing, D. A. Fong, K. M. Tan, E. Y. M. Lo, and S. G. Monismith, "Water and heat budgets of a shallow tropical reservoir," Water Resources Research, vol. 48, Jun 26 2012. [5] D. C. Oesch, J. M. Jaquet, A. Hauser, and S. Wunderle, "Lake surface water temperature retrieval using advanced very high resolution radiometer and Moderate Resolution Imaging Spectroradiometer data: Validation and feasibility study," Journal of Geophysical Research-Oceans, vol. 110, Dec 15 2005. [6] S. Uhlenbrook, Y. Mohamed, and A. S. Gragne, "Analyzing catchment behavior through catchment modeling in the Gilgel Abay, Upper Blue Nile River Basin, Ethiopia," Hydrology and Earth System Sciences, vol. 14, pp. 2153-2165, 2010. [7] Z. Duan and W. G. M. Bastiaanssen, "Estimating water volume variations in lakes and reservoirs from four operational satellite altimetry databases and satellite imagery data," Remote Sensing of Environment, vol. 134, pp. 403–416, Jul 2013. [8] Z. Duan and W. G. M. Bastiaanssen, "First results from Version 7 TRMM 3B43 precipitation product in combination with a new downscaling-calibration procedure," Remote Sensing of Environment, vol. 131, pp. 1-13, Apr 15 2013. [9] S. G. Setegn, D. Rayner, A. M. Melesse, B. Dargahi, and R. Srinivasan, "Impact of climate change on the hydroclimatology of Lake Tana Basin, Ethiopia," Water Resources Research, vol. 47, Apr 19 2011. [10] Z. Wan, "Collection-5 MODIS Land Surface Temperature products Users’ Guide", available at: http://www.icess.ucsb.edu/modis/LstUsrGuide/MODIS_LST_prod ucts_Users_guide_C5.pdf, 2007.
