THE Regional Distributions and seasonal variations of surface heat Fluxes over the Central Tibetan Plateau area
YAOMING MA
Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, 730000,China
Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing 100029, China
Hirohiko Ishikawa
Disaster Prevention Research Institute, KyotoUniversity,Kyoto 611-0011,Japan
Toshio Koike
Department of Civil Engineering, University of Tokyo, Tokyo 113-8656, Japan
Tandong Yao
Institute of Tibetan Plateau Research, ChineseAcademy of Sciences,Beijing 100029, China
In this study, a parameterization method based on LANDDSAT7 ETM data and field observations is described and tested for deriving the regional land surface variables, vegetation variablesand land surface heat fluxes over heterogeneous landscape.As a case study, the method was applied to the experimental area of the CEOP (Coordinated Enhanced Observing Period) Asia-Australia Monsoon Project (CAMP) on the Tibetan Plateau (CAMP/Tibet, 2001-2005), which located at the central Tibetan Plateau. Two scenes of LANDDSAT7 ETM data are used in this study. The scene of June 9, 2002 was selected as a case of summer, and the scene of December 2, 2002 was selected as a case of winter. Using the parameterization method, the reasonableregional distributions of surface reflectance, NDVI, MSAVI (Modified Soil Adjusted Vegetation Index), vegetation coverage, LAI (leaf area index), surface temperature, net radiation flux, soil heat flux, sensible heat flux and latent heat fluxwere determined over the Tibetan Plateau area. The parameterization method is still in a developing stage, further improvement of the method was also discussed.
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. Long-term research on the Tibetan Plateau have shown that the giant prominence exerts thermal effects on the atmosphere, thus greatly influencing circulations over China, Asia and even the globe (Ye and Gao, 1979; Ye, 1981; Ye and Wu, 1998;Yanai et al., 1992; Ma et al., 2002a; Ma and Tsukamoto, 2002b). Due to its topographic character, the plateau surface absorbs a large amount of solar radiation energy (much of which is redistributed by cryospheric processes), and undergoes dramatic seasonal changes of surface heat and water fluxes (Ye and Gao, 1979; Ye and Wu, 1998; Yanai et al., 1992). The lack of quantitative understanding of interactions between the land surface and atmosphere makes it difficult to understand the complete energy and water cycles over the Tibetan Plateau and their effects on the Asian monsoon system with numerical models. Therefore, it has increased the number of land surface processes studies over the Tibetan Plateau in the past 30 years. But the previous experiments were only carried out in a short period and the observational items were limited, and the previous investigations were only in summer period and on some points or local level (Ye and Wu, 1998; Ma et al., 2002a; Ma 2001).
The intensive observation period (IOP) and long-term observation of the CAMP/Tibet (CEOP (Coordinated Enhanced Observing Period) Asia-Australia Monsoon Project (CAMP) on the Tibetan Plateau, 2001-2005)have been done successfully in the past 3 years. A large amount of data has been collected, which is the best data set so far for the study of energy and water cycle over the Tibetan Plateau. It gives us a chance to investigate the energy and water cycle over the Tibetan Plateau in detail.
The study on the regional surface energy partitioning and its inter-monthly variation is of paramount importance over heterogeneous landscape of the Tibetan Plateau and it is also one of the main scientific objectives of the CAMP/Tibet. Therefore, we will try to derive the regional land surface heat fluxes by using the Landsat ETM data and the field observational data in this study. Here, “Regional” land surface heat fluxes are not “aggregated” fluxes (Batchvarova et al., 2001), but surface fluxes fields or surface heat fluxes on each pixel of Landsat ETM data.
Experiment and data
Experiment
The objectivesof CAMP/Tibet are: 1)Quantitative understanding of an entire seasonal hydro-meteorological cycle including winter processes by solving surface energy "imbalance" problems in the Tibetan Plateau; 2) Observation of local circulation and evaluation of its impact on plateau scale water and energy cycle; and 3) Establishment of quantitative observational methods for entire water and energy cycle between land surface and atmosphere by using satellites. To achieve the scientific objectives of CAMP/Tibet, a meso-scale observational network (150×250 km, 91°-92.5°E, 30.7°-33.3°N) were implemented in the central plateau (Ma et al., 2003a): 1) A basic observational station (BJ). A flux measuring tower (20 m, two levels), a Sky Radiometer, a LIDAR system, a Wind Profiler and RASS, a radio sonde system, and 4 AWSs have been set up at this station; (2) AWS networks. Six AWS stations have beendeployed in this area; (3) Soil moisture and soil temperature (SMTMS) networks. Seven SMTMS sites have been deployed in this area; (4) Two deep soil temperature measurements were put in D105 and NaquDS; (5) Anduo PBL station (91°3730E, 32°1428N). A PBL tower (17 m, wind speed, wind direction, air temperature and humidity at three-levels) and radiation observational system have already been continued for 6 years from 1997 and will be continued for another two years till the end of 2005. The enhanced automated observing period (EAOP) and IOP of the CAMP/Tibet have been started from October 1, 2002 and will be continued to September 30, 2004.
Data
Landsat-7 Thematic Mapper (TM) providea spectral radiance in seven narrow bands, with a spatial resolution of about 30×30m2 for three visible bands (Band 1, 2, 3),three near infrared bands (Band4, 5, 7), and 60×60m2 for the thermal infrared band 6. Two ETM images used in this paperare at 10:00(local time), June 2, 2002(summer) and December 2, 2002(winter) over the CAMP/Tibet.
The most relevant data, collected at the CAMP/Tibet surface stations to support the parameterization of land surface heat fluxes and analysis of ETM 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, radio-sonde, turbulent fluxes measured by eddy-correlation technique, soil heat flux, soil temperature profiles, soil moisture profiles, and the vegetation state etc..
Methodology
The general concept of the methodology is shown in adiagram (see Figure 1). The surface reflectance for short-wave radiation (r0) is retrieved from Landsat ETM data with the atmospheric correction by a four-stream radiative transfer assumption for atmospheric correction in solar spectral bands (Verhoef 1997) using aerological observation data. The land surface temperature (Tsfc) is also derived from Landsat ETM data and aerological observation data (Ma et al., 2003b). The radiative transfer model MODTRAN (Berk et al. 1989) computethe downward short wave and long- wave radiation at the surface. With these results the surface net radiation (Rn) is determined. The soil heat flux (G0) is estimated from Rn, Tsfc , r0 and MSAVI (Qi et al.1994) which is also derived from Landsat ETM data. The sensible heat flux (H) is estimated from Tsfc, surface and aerological data with the aid of so called‘blending height’approach (Mason 1988).
Figure 1. Diagram of parameterization procedure by combiningLandsat ETM data with field observations.
Net radiation
The regional net radiation flux can be derived from:
(1)
where r0(x,y),issurface reflectance.It can be derived from Landsat ETM databy using a four- stream radiative transfer assumption for atmospheric correction in solar spectral bands (Vehoef 1997; Ma 2001). Surface temperature Tsfc(x,y) in Eq.(1) can be derived from Landsat ETM band-6 (10.2-12.5m) spectral radiance (Ma et al., 2003b). The incoming short-wave radiation flux and incoming long-wave radiation flux K(x,y) and L(x,y)in Eq.(1) can be derived from radiative transfer model MODTRAN(Ma 2001).Surface emissivity 0(x,y) will be determined by Valor and Caselles’s method(Valor and Caselles 1997) .
Soil heat flux
The regional soil heat flux G0(x,y)will be determined by using a parameterization based onMSAVI (Modified Soil Adjusted Vegetation Index, Qi et al. 1994)
(2)
where the constants a, b, c, d and eare determined by using the field data observed at the CAMP/Tibet observation stations;is a daily mean reflectance value, i.e.
where (4)
where r3 and r4 are the band reflectance of Landsat ETM Band-3 and Band-4 on the land surface.
Sensible heat flux
The regional distribution of sensible heat flux can be estimated from
(5)
To simulate sensible heat flux on a large scale, a straightforward method is to scale-up or aggregate the regional sensible flux by a weighted average of the contributions from different surface elements, based on the principle of flux conservation. A method of “blending height”is proposed to derive the regional sensible heat flux density in this study. If the local-scale advection is comparatively small during theperiod of Landsat ETM observation taking place, the development of convection boundary layer may adjust the surface-disorganized variability at “blending height”, where the atmospheric characteristics become proximately independent of the horizontal position. Based on this approach,the regional sensible heat flux density H(x,y) can bedescribed as
(6)
where ZB is blending height, uB is wind speedat the blending height.ZBand uB can be determined by using field measurements or numerical models. In this study, they will be determined with the aid of field measurements of radiosonde Ta(x,y) in equation (6) is the regional distribution of air temperature at the reference height.An improved interpolation method (Ma et al., 2002b)is used to derive theregional distribution of air temperature here.The effective aerodynamic roughness length Z0m(x,y)in Eq.(6) over the CAMP/Tibet area including the effect of topography and low vegetation (e.g. grass),it can be determined by the Taylor’s model (Tayloret al. 1989). Raupach’s method (Raupach 1994)will be used to derive the zero-plane displacementd0(x,y) in equation (6) over the CAMP/Tibet areas, i.e.
(7)
where h(x, y) is the height of vegetation, and Cd1 is a free parameter (Raupach 1994). In otherwords, zero-plane displacement d0(x,y) can be derived when LAI and the vegetation height are determined over two areas. kB-1(x,y) in Eq.(6) will be determined by using the relationship between kB-1(x,y) and Ts(x,y)( Ma et al., 2002a).h(x,y) and m(x,y) in Eq.(6) are the integrated stability functions. They can be determined by using the models of Paulson (Paulson 1970) and Webb (Webb 1970).
Latent heat flux
The regional latent heat fluxE(x,y)will be derived as the residual of the energy budget theorem for land surface, i.e.
(8)
Cases study and validation
Using the method proposed in last Section, the distribution maps of land surface heat fluxes over the CAMP/Tibet area are derived (color maps were omitted here). The frequency distributions of surface temperature and land surface fluxes over the CAMP/Tibet area are shown in Figure 2. The derived land surface heat fluxes are validated by field measurements.Since it isdifficult to determine where the exact locations of the experimental sites are, the values of a 5×5 pixel rectangle, surrounding the determined Universal Transfer Macerator (UTM) coordinate, are compared with the field measurements. The mean absolute percent difference (MAPD) can quantitatively measure the difference between the derived results (Hderived(i)) and measured values( Hmeasured(i)) as
. (9)
It is seen that: (1) the derived land surface variables (land surface reflectance and surface temperature), vegetation variables (NDVI, MSAVI, vegetation coverage Pvand Leaf Area Index LAI) and land surface heat fluxes (net radiation flux Rn, soil heat flux G0, sensible heat flux H and latent heat flux E).These parameters show a wide range of variations due to the strong contrast of surface features in the study area; (2) not only on June 9, but also on December 2, the derived net radiation flux, soil heat flux and sensible heat flux were close to the field measurements. The difference between the derived results and the field observation MAPD was less than 10%;
Figure 2. Frequency distribution of surface temperature and land surface heat fluxes for the CAMP/Tibet area.10:00(LST)
Figure 2. Frequency distribution of surface temperature and land surface heat fluxes for the CAMP/Tibet area.10:00(LST) (Continued)
(3) the derived land surface temperature, net radiation flux, soil heat flux, sensible heat flux and latent heat flux in summer (9 June 2002) are much higher than they are in winter (2 December 2002); (4) during the experimental periods, the derived net radiation flux was larger than that in the HEIFE area (Ma et al., 2002b) due to the high altitude (the higher value of downward short-wave radiation) and land surface coverage of grassy marshland (the lower value of the upward long-wave radiation) in this area. For example, the regional average value of net radiation flux was470 W/m2over the HEIFE area in 9 July 1991 and that was 600 W/m2 over the CAMP/Tibet areain 9 June 2002;(5) The derived regional soil heat fluxes based on MSAVI were reasonable in different months in this area with MAPD less than 10%; (6) the results derived from Landsat ETM are comparable to the results derived from NOAA/AVHRR over this area in June (Ma et al., 2003c).
Concluding remarks
Using Landsat ETM data and field observations,the regional distributions and inter-monthly variations of land surface heat fluxes (net radiation, soil heat flux and sensible heat flux) over heterogeneous area of the CAMP/Tibet are derived. The results are in good agreement with field observations. The approach of deriving regional latent heat flux as the residual of the energy budget may not be a good method due to the measured“imbalance” of energy and the strong advection over the study area (Ma et al., 2003a). Future improvements are to be made to derive more accurate regional latent heat flux over such areas. It is also worth trying SEBS (Su 2002) to derived regional land surface heat fluxes, especially for latent heat flux.
All the results in this paper are gotten from high elevation area, the Tibetan Plateau. In order to extend them and the parameterization method presented here to a broader perspective, the results gotten in the Tibetan Plateau have to be compared to them over similar landscape types, i.e. arctic, sub-arctic and alpine etc., and our parameterization method should also be used to the similar landscape types. All these research will be done in the coming days.
Acknowledgement
This research is under the auspices of the Innovation projects of ChineseAcademy of Science (KZCX3-SW-339 and KZCX3-SW-329) and the cooperative research project between Cold and Arid Regions Environmental and Engineering Research Institute, ChineseAcademy of Sciences and Mitsubishi Heavy Industries, Ltd, Japan.Some parts of this study were conducted as cooperative research works at Kyoto University, Japan.The first author would like also to acknowledgeDr.Zhongbo Su, Dr. Zhaoliang Li, Profs.M.Menenti and Dr. Jun Wen for their help in the procedure of the paper. The authors thank all the participants from China and Japan in the field observation of the GAME/Tibet and the CAMP/Tibet.
REFERENCES
[1]Batchvarova E, Gryning S E and Hasager C B, “Regional fluxes of momentum and heat over a sub-artic landscape during late winter”. Bound.-Layer Meteor, Vol.99, (2001), pp489-507.
[2]BerkA, Bernstein LS and Robertson DC, “MODTRAN, A moderate resolution model for LOTRAN 7”, GL-TR-89-0122, (1989).
[3]MaY M, “Parameterization of land surface heat flux densities over inhomogeneous landscape by combining satellite remote sensing with field observations”, P.h.D Thesis, Okayama University, Japan, (2001), pp1-195.
[4]MaY M, Tsukamoto O, Wang J M, Ishikawa H and Tamagawa I, “Analysis of aerodynamic and thermodynamic parameters over the grassy marshland surface of Tibetan Plateau”,Progress in Natural Sci, Vol.12, No.1, (2002a), pp36-40.
[5]MaY M, Tsukamoto O, Ishikawa H, Su Z B, Menenti M, WangJ M and Wen J, “Determination of Regional land surface heat flux densities over heterogeneous landscape of HEIFE integrating satellite remote sensing with field observations”,J. Meteor. Soc. Japan, Vol.80, No.3, (2002b), pp485-501.
[6]MaY M, Su Z B, Koike T, Yao T D, Ishikawa H, Ueno K and Menenti M, “On measuring and remote sensing surface energy partitioning over the Tibetan Plateau -from GAME/TIBET to CAMP/TIBET”, Physics and Chemistry of the Earth, Vol.28 No.1-3, (2003a), pp63-74.
[7]MaY M,WangJ M, Huang R H, WeiG A, Menenti M, Su Z B, HuZ Y, Gao F and Wen J, “Remote sensing parameterization of land surface heat fluxes over arid and semi-arid areas”,Advances in Atmospheric Sciences, Vol.20, No.4, (2003b),pp530-539.
[8]MaY M, Ishikawa H, Tsukamoto O, Menenti M, Su Z B, Yao T D,Koike T and Yasunari T, “Regionalization of surface flux densities over heterogeneous landscape of Tibetan Plateau using satellite remote sensing data”,J. Meteor. Soc. Japan, Vol.81, No.2, (2003c), pp 277-293.
[9]Mason P, “The formation of areally averaged roughness lengths”, Quart.J.Roy-Meteor.Soc.Vol.114, (1988), pp399-420.
[10]Paulson C A, “The mathematic representation of wind speed and temperature profiles in the unstable atmospheric surface layer”, J. Meteor. Soc. Japan. Vol.9, (1970), pp 856-861.
[11]Qi J G, Chehbouni A, Huete A R, Kerr Y H and Sorooshian S, “A ModifiedSoil Adjusted Vegetation Index”,Remote Sens. Environ , Vol.48, (1994), pp119-126
[12]RaupachMR, “Simplified expressions for vegetation roughness length and zero-plane displacements as functions of canopy height and area index”,Bound.-Layer Meteor. Vol.71, (1994), pp 211-216.
[13]Su Z B, “The Surface Energy Balance System (SEBS) for estimation of turbulent heat fluxes”, Hydrology and Earth System Sciences, Vol.6, No.1, (2002), pp85-99.
[14]TaylorPA, Sykes RI and MasonPJ, “On the parameterization of drag over small scale topography in neutrally stratified boundary flow”,Bound.-Layer Meteor.Vol.48, (1989), pp 409-422.
[15]ValorE and CasellesV, “Mapping land surface emissivity from NDVI: application to European, African, and South American areas”,Remote SensingEnviron. Vol.57, (1989),pp167-184.
[16]Verhoef W, “Theory of radiative transfer models applied in optical remote sensing of vegetation canopies”, PhD thesis, Remote Sensing Department of National Aerospace Laboratory, The Netherlands, (1997).