1 Impacts of climate change on the hydrological processes in the Mekong River Hui Lu & Wei Wang Tsinghua University
Outline Backgrounds Method and data Model calibration Temperature and rainfall trends Runoff trends Summary Acknowledgement 2
Backgrounds climate change will likely change rainfall amounts and patterns and the frequency and extent of extreme weather events Evaluation of the hydrological impacts of climate change become a research hotspot As changes in weather patterns are being felt across the Lower Mekong Basin, the impacts of climate change have become a topic of strong public interest The IPCC 5th Assessment Report has been released this year 3
Method and data Build a distribute model in Mekong Hydrological model validation Bias correction for GCM simulation Using downscale GCM outputs to force the model validation period data 1998 2007 TRMM bias correction spin up simulate 1990 1994 1995 2004 projection spin up simulation 2001 2010 2011 2040 2056 2060 2061 2070 2086 2089 2090 2099 4
Model and data We used a geomorphology based hydrological model (GBHM) A physical distributed hydrological model a flow-interval hillslope scheme is applied to simplify the complex two-dimensional water kinematics to one-dimensional 5
Method and data Data for building the digital basin DEM data, Land use and land cover from the USGS Soil type and soil depth data are from FAO NDVI from MODIS 6
Method and data Data for driving the model Calibration data: remote sensing rainfall in grid TRMM3b42 Future projection driven by output of 5 GCM in 4 RCP: 0.25 downscaling daily data from Max Planck Institute GFDL-ESM2M HadGEM2-ES IPSL-CM5a-LR MIROC-ESM-CHEM NorESM1-M 7
Model calibration Calibration by TRMM (1998) Daily Runoff m^3/s 2.53.5 x 104 x 104 3.5 x 104 Luang Stung Prabang Treng Chiang Mukdahan Sean 16000 3.5 x 104 obs obs obs simobs sim Pakse sim 3 3 3 2 2.5 2.5 2.5 1.5 2 2 2 1.5 nash=0.86 1.5 1.5 nash=0.57 nash=0.82 nash=0.77 6000 nash=0.63 1 1 1 4000 0.5 0.5 0.5 2000 0.5 Chiang Sean Luang Prabang Mukdahan Pakse Stung Tren 0 0 0 50 50 50 100 100 150 150 200 200 250 250 300 300 350 350 400 400 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Day of year 8
Model calibration Model performance in validation (1999-2007) 3.5 104 4 x 104 obs obs 15000 obs sim sim Pakse obs sim obs sim 3.5 sim 3 5 7000 3 2.5 6000 4 2.5 2 5000 2 3 4000 1.5 1.5 3000 5000 1 2 1 2000 0.5 1.5 Monthly Runoff m^3/s 6 x 104 Chiang Mukdahan Sean Stung Treng Luang Prabang nash=0.89 nash=0.86 nash=0.81 nash=0.86 nash=0.88 Chiang Sean Luang Prabang Mukdahan Pakse Stung Treng 0 0 0 0 1998 1998 1999 1999 2000 2000 2001 2001 2002 2002 2003 2003 2004 2004 2005 2005 2006 2006 2007 2007 1998 1999 2000 2001 2002 2003 2004 2005 2006 1998 1999 2000 2001 2002 2003 2004 1998 1999 2000 2001 2002 2003 2004 2005 2006 Year 9
station GCM bias bias_r CORR nash RMSE Model calibration GFDL -108.779-0.04 0.474-0.427 2490.549 Had -287.295-0.105 0.823 0.652 1230.533 Chiang Sean IPSL 90.892 0.033 0.783 0.516 1449.881 MIROC 44.172 0.016 0.793 0.57 1367.943 Bias correction by history data Nor 135.205 0.049 0.774 0.486 1494.461 GFDL -529.076-0.063 0.669 0.336 6510.556 Use the history data of GCMs to test the GCMs Mukdahan output in 1995-2004 Had -1052.019-0.125 0.816 0.647 4745.339 IPSL 354.59 0.042 0.759 0.474 5797.509 MIROC 387.416 0.046 0.835 0.663 4639.135 Calculate the ratio of simulated runoff with observation: b Nor 116.308 0.014 0.822 0.651 4720.134 GFDL -642.115-0.062 0.704 0.421 7646.033 Had -1296.422-0.126 0.796 0.615 6235.71 Pakse Use b to IPSL correct 205.337 the projected 0.02 runoff 0.776 to 0.529 remove 6897.488 the errors brought MIROC by 87.19 GCMs 0.008 and hydrological 0.848 0.708 model 5429.155 Nor -322.535-0.031 0.836 0.689 5604.019 GFDL -209.218-0.015 0.781 0.526 9594.309 Had -1458.809-0.103 0.78 0.583 8994.594 Stung Treng IPSL 715.313 0.05 0.801 0.552 9321.214 MIROC 222.555 0.016 0.866 0.735 7168.04 10 Nor -597.948-0.042 0.855 0.715 7433.296
Temperature and rainfall trends 28 27 26 rcp2.6 rcp2.6 GFDL rcp2.6 HadESM 30 rcp2.6 IPSL 27 temperature The temperature will rise projected by all GCMs rcp2.6 MIROC rcp2.6 NorGEM 28 26 rcp4.5 rcp4.5 GFDL rcp4.5 HadESM rcp4.5 IPSL rcp4.5 MIROC rcp4.5 NorGEM 25 24 23 29 28 25 GFDL rcp2 GFDL rcp4 24 GFDL rcp6 GFDL rcp8 IPSL rcp2 IPSL rcp4 23 IPSL rcp6 IPSL rcp8 22 21 20 28 27 26 25 24 23 22 21 HadESM rcp2 HadESM rcp4 22 HadESM rcp6 HadESM rcp8 27 MIROC rcp2 MIROC rcp4 21 MIROC rcp6 MIROC rcp8 NorGEM rcp2 NorGEM rcp4 20 NorGEM rcp6 NorGEM rcp8 2000s 2010s 2020s 2030s 26 2060s 2090s rcp6.0 25 rcp6.0 GFDL rcp6.0 HadESM rcp6.0 IPSL rcp6.0 MIROC rcp6.0 NorGEM 24 23 22 29 28 27 rcp8.5 rcp8.5 GFDL rcp8.5 HadESM rcp8.5 IPSL rcp8.5 MIROC rcp8.5 NorGEM 26 25 24 23 22 21 20 22000s nd Mekong Climate 2010s Change 2020sForum2030s 2060s 2090s 11 20
Temperature and rainfall trends 1800 1750 1700 1650 1600 1550 1500 1450 1400 1350 1300 1800 1750 1700 1650 rcp2.6 2000 1750 The GFDL HadGEM rainfall IPSL trends MIROC NorESM show great disagreement rainfall with 1700 1900 1650 each other 1800 1550 The NorESM increase as 1500 time going 1700 1450 1600 The GFDL and IPSL fluctuate 1400 dramatically 1500 1300 The HadGEM and MIROC first 2000s decease 2010s 2020s then 2030s increase 2060s 2090s 1400 rcp6.0 rianfall (mm) 1800 1600 1350 1800 rcp4.5 GFDL HadGEM IPSL MIROC NorESM rcp8.5 1300 1750 GFDL HadGEM IPSL MIROC NorESM GFDL rcp2 GFDL rcp4gfdl GFDL HadGEM rcp6 IPSLGFDL rcp8 MIROC IPSL NorESM rcp2 1700 1200 IPSL rcp4 1650 IPSL rcp6 IPSL rcp8 HadESM rcp2 HadESM rcp4 1600 1550 1500 1450 1400 1100 1000 HadESM rcp6 1600 HadESM rcp8 MIROC rcp2 MIROC rcp4 MIROC rcp6 1550 MIROC rcp8 1500 NorGEM rcp2 NorGEM rcp4 NorGEM rcp6 NorGEM rcp8 2000s 2010s 1450 2020s 2030s 2060s 2090s 1400 1350 1350 1300 2000s 2 nd 2010s 2020s 2030s Mekong Climate Change Forum 2060s 2090s 12 1300
Runoff trends-chiang Saen 4000 3800 3600 3400 GFDL G2 G4 G6 G8 4500 4000 HadGEM H2 H4 H6 H8 3400 3200 3000 IPSL 3200 3000 2800 3500 3000 2800 2600 2600 2400 2200 2000 2500 2000 2400 2200 2000 I2 I4 I6 I8 3600 3400 MIROC 3800 3600 NorESM 3800 3600 average 3200 3400 3400 3000 2800 2600 2400 3200 3000 2800 2600 2400 3200 3000 2800 2600 2400 2200 2000 M2 M4 M6 M8 2200 2000 N2 N4 N6 N8 2200 2000 rcp2.6 rcp4.5 rcp6.0 rcp8.5 13
Runoff trends-mukdahan 11500 GFDL 11500 HadGEM 11500 IPSL I2 I4 I6 I8 10500 10500 10500 9500 9500 9500 8500 8500 8500 7500 7000 G2 G4 G6 G8 2010s 2010s 2020s 2030s 2060s 2090s 7500 7000 H2 H4 H6 H8 2010s 2010s 2020s 2030s 2060s 2090s 7500 7000 2010s 2010s 2020s 2030s 2060s 2090s MIROC M2 M4 M6 M8 NorESM N2 N4 N6 N8 rcp2.6 rcp6.0 average rcp4.5 rcp8.5 7000 7000 7000 6000 2010s 2010s 2020s 2030s 2060s 2090s 6000 2010s 2010s 2020s 2030s 2060s 2090s 6000 2010s 2010s 2020s 2030s 2060s 2090s 14
Runoff trends-pakse GFDL HadGEM IPSL G2 G4 G6 G8 MIROC H2 H4 H6 H8 NorESM 11500 I2 I4 I6 I8 average 10500 9500 M2 M4 M6 M8 N2 N4 N6 N8 8500 rcp2.6 rcp6.0 rcp4.5 rcp8.5 2010s 2010s 2020s 2030s 2060s 2090s 15
Runoff trends-stung Treng 20000 1 GFDL 20000 1 HadGEM 20000 1 IPSL I2 I4 I6 I8 1 1 1 17000 17000 17000 16000 16000 16000 15000 15000 15000 G2 G4 G6 G8 H2 H4 H6 H8 20000 1 1 17000 MIROC M2 M4 M6 M8 20000 1 1 17000 NorESM N2 N4 N6 N8 1 17000 16000 rcp2.6 rcp6.0 average rcp4.5 rcp8.5 16000 16000 15000 15000 15000 2010s 2010s 2020s 2030s 2060s 2090s 16
Runoff trends- in 2090s Runoff increase biggest in RCP4.5 Chiang Saen increase biggest for most scenarios Stung Treng increase smallest for all scenarios Chiang Sean Mukdahan Pakse Stung Treng RCP2.6 6.7% 11.1% 8.7% 6.5% RCP4.5 24.1% 18.1% 14.2% 12.5% RCP6.0 8.7% 7.2% 7.7% 7.6% RCP8.5 29.1% 10.0% 10.3% 8.4% 17
Runoff trends-inter-annual variations 900 800 700 600 Chiang Sean average of std There has a trend of inter-annual 2800 variations increasing especially in 2300Chiang Saen 3300 1800 Mukdahan average of std 500 1300 400 800 300 rcp2.6 rcp4.5 rcp6.0 rcp8.5 300 rcp2.6 rcp4.5 rcp6.0 rcp8.5 3800 3300 Pakse average of std 4300 3800 Stung Treng average of std 2800 3300 2300 1800 2800 2300 1800 1300 1300 800 300 rcp2.6 rcp4.5 rcp6.0 rcp8.5 800 300 18 rcp2.6 rcp4.5 rcp6.0 rcp8.5
Summary It is obviously that the average of temperature in the basin will increase in the future For the precipitation, the GCMs shows great disagreement from each other For the runoff, it depends largely on precipitation trends, but it will increase on average The inter-annual variations will increase especially in Chiang Saen, the discharge in Chiang Saen increase biggest for most scenarios the upper Mekong is more vulnerable to climate change 19
Acknowledgement This work was jointly supported by the National 863 Program of China (2012AA12A309), the National Natural Science Foundation of China (51109111, 41371328 and 51190092), Tsinghua University Initiative Research Program (2011081132), and the Beijing Higher Education Young Elite Teacher Project (YETP0132). Thank the International Scientific & Technical Data Mirror Site, Computer Network Information Center, CAS for RS data and MRC for situ data. Thank the High Performance Computing (HPC), Tsinghua University for the computation support. 20
21 Thank you luhui@tsinghua.edu.cn