Combination satellit and in-situ data for the determination of evapotranspiration over heterogeneous landscape of the Tibetan Plateau

Key Laboratory of Tibetan Environment Changes and Land Surface Processes, In titute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing 100101, China Qomolangma Station for Atmospheric Environmental Observation and Research, Chinese Academy of Sciences, Dingri 858200, Tibet, China University of Chinese Academy of Sciences, Beijing 100049, China Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou 730000, China School of Earth and Space Sciences, University of Science and Technology of China, Hefei 230026, China


Figures Back Close
Full

Abstract
In this study, a new parameterization method based on MODIS (Moderate Resolution Imaging Spectroradiometer) data, AVHRR (Advanced Very High-Resolution Radiometer) data and in-situ data is constructed and tested for deriving the regional evaporative fraction (EF) over heterogeneous landscape.As a case study, the methodology was applied to the Tibetan Plateau area.Eight images of MODIS data (17 January 2003, 14 April 2003, 23 July 2003and 16 October 2003;30 January 2007, 15 April 2007, 1 August 2007and 25 October 2007) and four images of AVHRR data (17 January 2003, 14 April 2003, 23 July 2003and 16 October 2003) were used in this study for the comparison among winter, spring, summer and autumn and the annual variation analysis.The derived results were also validated by using the "ground truth" measured in the stations of the Tibetan Observation and Research Platform (TORP) and the CAMP/Tibet (CEOP (Coordinated Enhanced Observing Period) Asia-Australia Monsoon Project (CAMP) on the Tibetan Plateau).The results show that the derived EF in four different seasons over the Tibetan Plateau area is in good accordance with the land surface status.The EF show a wide range due to the strong contrast of surface features over the Tibetan Plateau.Also, the estimated EF is in good agreement with the ground measurements, and their absolute percent difference (APD) is less than 10 % in the validation sites.The results from AVHRR were also in agreement with MODIS, with the latter usually displaying a higher level of accuracy.It is therefore concluded that the proposed methodology is successful for the retrieval of EF using the MODIS data, AVHRR data and in-situ data over the Tibetan Plateau area, and the MODIS data is the better one and it should be used widely for the evapotranspiration (ET) research over this region.

Introduction
As the most prominent and complicated terrain on the globe, the Tibetan Plateau, with an elevation of more than 4000 m on average above mean sea leave (msl) makes up approximately one fourth of the land area of China (Fig. 1).Due to its topographic character, the plateau surface absorbs a large amount of solar radiation energy, and undergoes dramatic seasonal changes of surface heat and water fluxes (e.g.Yanai et al., 1992;Ye and Wu, 1998;Hsu and Liu, 2003;Sato and Kimura, 2007).Evapotranspiration (ET) between the land surface and atmosphere of the Tibetan Plateau play an important role in the Asian Monsoon system, which in turn are major components of the energy and water cycles of the global climate system.Some interesting detailed studies concerning the ET (or latent heat flux) have been reported in the different sites and land surface types over the Tibetan Plateau in the past years (e.g.Tanaka et al., 2001;Ma et al., 2005;Li et al., 2006;Zhong et al., 2009).These researches were, however, on point-level or a local-patch-level.Remote sensing from satellites however offers the possibility to derive ET regional (or areal) distribution over heterogeneous land surface in combination with sparse field experimental stations.And the regional ET distributions have been reported over the Tibetan Plateau in the past years (e.g.Ma et al., 2003;Ma et al., 2009), but the results were still in a meso-scale area.To understand the effect of the Tibetan Plateau on the climatic change over China, East Asia, and even the globally, the regional distribution of ET distribution of over whole Tibetan Plateau must been determined.Our purpose in this study is to estimate the regional distribution of ET and its seasonal variation over the whole Tibetan Plateau with the aid of MODIS data, AVHRR data and in-situ data.It is because the ET from land is essential for understanding climate dynamics and ecosystem productivity (e.g.Churkina et al., 1999) and it also has applications in areas such as water resource management.We will introduce "evaporative fraction (EF)" as an index for ET after Shuttleworth et al. (1989).EF is defined Introduction

Conclusions References
Tables Figures

Back Close
Full here as: Where H is the sensible heat flux, λE the latent heat flux, R n the net radiation flux and G 0 the soil heat flux.Our goal is not estimation of ET but rather EF in whole Tibetan Plateau area.This is due to two reasons.First, EF is more suitable as an index for surface moisture condition than ET.Because ET itself is a function not only of the land surface conditions (e.g.soil moisture and vegetation) but also surface available energy R n −G 0 (= H+λE ), ET cannot be easily interpreted for soil moisture condition or drought status.On the contrary, EF can be more directly related to these land surface conditions.Secondly, EF is useful for scaling up instantaneous observations to longer time periods.As one knows, a satellite (except for geostationary satellite) observes each land surface very briefly during a day.ET, however, can generally change drastically during a day mainly due to changes in sun angle and cloud coverage.Therefore, even if we can accurately estimate ET at the moment of satellite overpass, it cannot be directly related to daily or daytime average ET.On the contrary, EF is well known to be nearly constant during most of daytime (e.g.Shuttleworth et al., 1989;Sugita and Brutsaert, 1991;Crago, 1996a, b).
We also find that EF is also nearly constant from sunrise to sunset during the clear days in BJ station (31.37 Full  2).Therefore, if we can estimate daytime average R n − G 0 or H + λE , daytime average ET will be estimated from Eq.( 1) by using instantaneous EF derived by satellite.There are some cases for the estimation by using the satellite and in-situ data.A similar spatial variation of broadband albedo and surface temperature was proposed to estimate EF (Su, 1999;Roerink et al., 2000).Jiang and Islam (2001) estimated EF by interpolating the Priestley-Taylor parameter.Venturini et al. (2004) did the comparison of EF estimated from AVHRR and MODIS sensors over South Florida of USA.Verstraeten et al. (2005) estimated EF from NOAA-imagery at satellite overpass time over European forests area.Wang et al. ( 2006) estimated EF from a combination of day and night land surface temperatures and NDVI.But no any study of EF over heterogeneous landscape of whole Tibetan Plateau till now.
The regional distribution of EF over the Tibetan Plateau will be estimated and validated in this study.We first propose the methodology in Sect. 2. The using satellite data and in-situ data for the validation are shown in Sect.3. The application of the methodology to the Tibetan Plateau area is presented in Sect.4, where the distribution of EF are estimated for four different phases, spring, summer, autumn and winter by using eight scenes of MODIS data and four scenes of AVHRR data.Discussions are given in the Sect. 5.

Theory and scheme
The general concept and procedure of the methodology are shown in a diagram (Fig. 3).The surface reflectance for short-wave radiation r 0 (x,y) is retrieved from MODIS data by using Zhong's method (Zhong , 2007) and AVHRR data by using the method of Ma et al. (2003) with the atmospheric correction, using land surface and aerological observation data.The land surface temperature T sfc (x,y) is also derived from MODIS data by using the method of Zhong et al. ( 2010 1989) computes the downward short-wave and long-wave radiation at the surface by using satellite data, land surface and aerological observation data and the atmospheric correction (Ma and Tsukamoto, 2002).With these results the regional surface net radiation flux R n (x,y) is determined by using the surface radiation energy budget theorem for land surface.The regional soil heat flux G 0 (x,y) is estimated from R n (x,y) and field observations over the Tibetan Plateau (Ma et al., 2002).It means that based on the field observation data from the key stations over the Tibetan Plateau, a good relationship between G 0 and R n can be found over the area.And the relationship will be used for the determination of G 0 (x,y) from R n (x,y ).The regional sensible heat flux H(x,y) is estimated from T sfc (x,y), surface and aerological data with the aid of so-called "tile approach" (Ma et al., 2010).The procedure to determine R n (x,y ) and H(x,y) in detail will be found in Ma's paper (Ma et al., 2011).The regional latent heat flux λE (x,y ) can be derived as the residual of the energy budget theorem for land surface.If the evaporative fraction (EF) defining Eq.( 1) is used to the MODIS and AVHRR pixel scale, it will be become: EF value is between 0.0 and 1 Λ equaling 0.0 means that the surface is very dry, and no ET from surface.Λ equaling 1.0 means that surface is very wet, and there are maximum ET from the surface.2003, 14 April 2003, 23 July 2003and 16 October 2003) are used here to find which satellite data is better for the determination of EF over heterogeneous landscape of the Tibetan Plateau.More four images of MODIS data in 2007 is used here for the more validations of methodology due to more validation sites set up in 2007 over the Tibetan Plateau (Ma et al., 2008).
The most relevant in-situ data, collected at the key stations over the Tibetan Plateau to support the parameterization of EF and analysis of the MODIS and AVHRR images, consist of surface radiation budget components, surface radiation temperature, surface reflectance, vertical profiles of air temperature, humidity, wind speed and direction measured at the PBL towers, Wind Profiler and RASS, radiosonde and tethersonde, turbulent fluxes measured by eddy-correlation technique, soil heat flux, soil temperature profiles, soil moisture profiles, and the vegetation state.The key stations in the TORP are including BJ station (31.37 • N, 91.90 Full shown here due to their key position in the procedure of determination of EF (Fig. 3).
The R n and EF distribution maps are based on 2875 × 1487 pixels with a size about 1 × 1 km 2 .The derived R n and EF can be validated by the field measurements.In-situ data observed in nine stations of Haibei, Maqu, D105, Amdo, NPAM, BJ, NAMOR, QOMS and SETS in the TORP (Ma et al., 2008) are used for the validation in 2007 and four stations of D105, NPAM, ANNI and BJ of the CAMP/Tibet ( Ma et al., 2005) are used for the validation in 2003.In Fig. 6, the derived results are validated against the measured values in the stations.The absolute percent difference (APD) was quantitatively measures the difference between the derived results (H derived(i ) ) and measured value (H measured(i ) ) here, and H measured(i ) . (3) The results show that: (1) the derived R n and EF in four different months over the Tibetan Plateau area are in good accordance with the land surface status.The Tibetan Plateau includes variety of land surfaces such as a large area of grassy marshland, some desertification grass-land areas, sparse grass-Gobi, sparseness meadow, many small rivers and lakes, snow (glacier) mountains, forest, and farmland etc.Therefore, these derived parameters show a wide range due to the strong contrast of surface fea- The mean EF from MODIS data over the Tibetan Plateau area is increasing from January to April and August, then decreasing from October (Fig. 5b and Fig. 5c).They are 0.275, 0.415, 0.567 and 0.406 for 2007 and 0.271, 0.342, 0.569 and 0.347 for 2003.
The mean EF from AVHRR data over the Tibetan Plateau area is also increasing from January to April and July, then decreasing from October.They are 0.267, 0.335, 0.502, and 0.331(Fig.5a).( 4) Because of the land surface cover and property in spring (April) is much complex (there are ice, snow, seasonal and long-living permafrost, grass land and lakes etc existing in this month), therefore the EFs distribution in this month is also complicated (Fig. 5).( 5) The most derived regional EFs with APD less than 10.0 % at validation sites in the Tibetan Plateau are in good agreement with field measurements, except EFs derived from AVHRR in ANNI (APD = 13.0 %) and D105 (APD = 11.0 %) in 23 July 2003 (Fig. 6).The reason is that the radiation transportation processes, using MODTRAN model, land surface and aero-logical data and the process of atmospheric boundary layer, such as using "tile approach" and the measuring and calculating accurately of wind speed u, air temperature T a and specific humidity q at the reference height, zero-plane displacement d 0 , aerodynamic roughness length z 0m and thermodynamic roughness length z 0h , the excess resistance for heat transportation kB −1 etc. in each tile were considered in more detail in our method (Fig. 3).It is pointed that our proposed parameterization methodology for EF is reasonable, and it can be used over the Tibetan Plateau area.( 6) The retrieval EF and net radiation from MODIS are supe-Introduction

Conclusions References
Tables Figures

Back Close
Full rior to those from AVHRR even the same methodology are used .It means that the derived values from MODIS are more closed to the "ground truth" measured in the validation stations (Fig. 6).Many reasons might account for this, such as the image quality itself, the different Split Window Algorithms (SWAs), or the different algorithms for water vapor.It is therefore concluded that the MODIS data is the better one and it should be used widely for the determination of surface heat fluxes and ET over the Tibetan Plateau area.

Conclusions
Concluding remarks In this study, the regional distributions of evaporative fraction (EF) over heterogeneous landscape of the Tibetan Plateau are derived with the aid of MODIS data, AVHRR data and the in-situ data.And the MODIS data is the better one than AVHRR data for the determination of EF over heterogeneous landscape.Compared with the field measurements, the proposed EF has been proved to be a better index to getting related evapotranspiration (ET) over heterogeneous landscape.
Regionalization the evapotranspiration (ET) over heterogeneous landscape is not an easy issue.The parameterization methodology presented in this research is still in developing stage due to only a single set of values at a specific time of specific day are used in this research.To reach more accurate regional ET and seasonal and annual variations of evapotranspiration over the Tibetan Plateau area, more field observations, more accurate radiation transfer models to determine the surface reflectance and surface temperature, more MODIS data, and other satellites such as ASTER (Advanced Spaceborne Thermal Emission and Reflection radiometer), Landsat-5 TM, Landsat-7 ETM, GMS (Geo-stationary Meteorological Satellite), and ATSR (Along Track Scanning Radiometer) have to be used.These research works will be done in the next step.Introduction

Conclusions References
Tables Figures

Tibetan Plateau
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | et al., 2008) (see Fig.
) and AVHRR data by using the method of Ma et al. (2003) with the atmospheric correction, using land surface and aerological observation data.The radiative transfer model MODTRAN (Berk et al., Discussion Paper | Discussion Paper | Discussion Paper |

3
Satellite data and field observation data Eight images of MODIS data (17 January 2003, 14 April 2003, 23 July 2003 and 16 October 2003; 30 January 2007, 15 April 2007, 1 August 2007 and 25 October 2007) and four images of AVHRR data (17 January 2003, 14 April 2003, 23 July 2003 and 16 October 2003) was chosen in this study for the comparison among winter, spring, summer and autumn.Same day images of AVHRR and MODIS data in 2003 (17 Discussion Paper | Discussion Paper | Discussion Paper |

Figure 4
Figure4and Fig.5shows the distribution maps of net radiation flux R n and evaporative fraction EF over the Tibetan Plateau area.The distribution maps of net radiation flux are shown here due to their key position in the procedure of determination of EF (Fig.3).The R n and EF distribution maps are based on 2875 × 1487 pixels with a size about tures over whole Tibetan Plateau.(2) The derived pixel value (Fig. 5) of EFs in summer (1 August 2007 and 23 July 2003) and autumn (25 October 2007 and 16 October 2003) are higher than that in winter( 30 January 2007 and 17 January 2003) and spring (15 April 2007 and 14 April 2003).EFs in summer are mostly from 0.70 to 1.0, EFs in autumn are mostly around 0.60, EFs in spring are mostly around 0.50, EFs in winter are mostly around 0.30 (some of the EFs are 1.0 in the distribution maps indicating cloud-Discussion Paper | Discussion Paper | Discussion Paper |covering).The differences between 2007 image and 2003 image seem to be large is due to day-to-day variability.It means that there are much more evapotranspiration in summer and autumn than it in winter and spring in the Tibetan Plateau area.The reason is that most the land surface is wet and covered by green grass and growing vegetations in summer and autumn and it is dry and most the mountain ranges is covered by snow and ice during winter and spring on the Tibetan Plateau area.In other words, sensible heat and latent heat fluxes play different roles in the partition of net radiation flux in different month in the Tibetan Plateau: sensible heat flux plays the main role in winter and spring and latent heat flux plays the main role in summer and autumn.(3) Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | EU-FP7 projects of "CORE-CLIMAX"(313085) and "CEOP-AEGIS"(212921).The in-situ data of Haibei station and Maqu station was come from Yingnian Li, Xinqaun Zhao and Yu Zhang.Discussion Paper | Discussion Paper | Discussion Paper | Wang, K., Li, Z., and Cribb, M.: Estimation of evaporative fraction from a combination of day and night land surface temperatures and NDVI: a new method to determine the Priestley-Taylor parameter, Remote Sens. Environ.,102,293-305, 2006.Yanai, M., Li, C., and Song, Z.: Seasonal heating of the Tibetan Plateau and its effects on the evolution of the Asian summer monsoon, J. Meteorol.Soc.Jpn., 70, 319-351, 1992.Ye, D. and Wu, G.: The role of the heat source of the Tibetan Plateau in the general circulation, Fig.1The location and landscape of the Tibetan Plateau.

Fig. 1 .
Fig. 1.The location and landscape of the Tibetan Plateau.
. Zhong, L.: Measurement and satellite remote sensing of land surface characteristic parameters over the Tibetan Plateau area, Ph.D. Thesis, Graduate University of Chinese Academy of Sciences, Beijing, China, 124 pp., 2007.Zhong, L., Ma, Y., Su, Z., Lu, L., Ma, W., and Lu, Y.: Land-atmosphere energy transfer and surface boundary layer characteristics in the Rongbu Valley on the northern slope of Mt.Everest, Arct.Antarct.Alp.Res., 41, 396-405, 2009.Zhong, L., Ma, Y., Salama Suhyb, Mhd., and Su, Z.: Assessment of vegetation dynamics and their response to variations in precipitation and temperature in the Tibetan Plateau, Clim.