1、SENSIBLE HEAT FLUX ESTIMATION USING SURFACE ENERGY BALANCE SYSTEM (SEBS), MODIS PRODUCTS, AND NCEP REANALYSIS DATA,Yuanyuan Wanga, Xiang Lia,ba, National Satellite Meteorological Center, China Meteorological Administrationb, Nanjing University of Information Science & Technology,1. INTRODUCTION,2. M
2、ETHODOLOGY,3. DATA,4. RESULTS AND ANALYSIS,5. DISCUSSION AND CONCLUSION,OUTLINE:,1. INTRODUCTION,This paper used the Surface Energy Balance System (SEBS) to estimate the regional sensible heat flux, fully using the advantages of temporal and spatial resolutions of MODIS data and the height of the Pl
3、anetary Boundary Layer of NCEP reanalysis data. In addition, a new approach which deriving the value of roughness length of momentum transport was being used in this paper to improve the accuracy of calculation. Based on the LAS continuously measurements at Arou site and sensible heat predictions da
4、ta provided by NCEP, regional sensible heat flux in Arou area estimated by MODIS data and NCEP data extended over several months were validated. The results demonstrated that running SEBS model with NCEP meteorological data was feasible.And a sensitivity analysis for , and has been performed in orde
5、r to investigate how the variable affect the result of regional sensible heat flux calculated by SEBS.,To estimate the components of energy balance Conventional techniques : point measurements ( point scales ) Recently : Remote sensing ( regional scales ) Validation of estimated sensible heat flux C
6、onventional techniques : point measurements ( point scales ) i.e. Bowen ratio system, Eddy correlation system Recently : Large Aperture Scintillometers (LAS) ( regional scales ),1. INTRODUCTION,1. INTRODUCTION,The roughness length of momentum can be derived from wind and temperature profiles in the
7、past.disadvantage :a. sensitive to measurement errors ;b. rejecting a large number of valuable datasets under non-neutral conditions .Since is physically related to the underlying surface and not sensitive to the diurnal variation of atmospheric stability, a new approach to derive the value of propo
8、sed by Kun Yang(2003). can be considered constant over a short period, In this paper, the short periods of time means 30 days (1 month).,SEBS (Surface Energy Balance System) was developed by Z.Su to estimate land surface fluxes using remotely sensed data and available meteorological observations.,2.
9、 METHODOLOGY,Fig.1 Single-source Model,Fig.2 The structure SEBS (Wang et al., 2008),1. INTRODUCTION,2. METHODOLOGY,Surface energy balance equation in its instantaneous form is expressed as : Where: is the net radiation, is the soil heat flux, is the turbulent sensible heat flux, and is the turbulent
10、 latent heat flux. the soil heat flux is estimated as:Where : and are the proportions of for full cover and bare soil, fixed as 0.05 and 0.315 respectively. The fractional vegetation cover weights between limiting cases.,2. METHODOLOGY,Innovations of SEBS:(1)Following the full canopy only model of C
11、houdhury and Monteith (1988), a bare soil surface of Brutsaert (1982), SEBS describes the parameterization method to interaction between vegetation and bare soil surface.Then, the roughness length for heat transfer can be derived by:,2. METHODOLOGY,(2) In order to derive the actual sensible heat flu
12、x H , use is made of the similarity theory. Where, is the potential temperature at the surface， is the potential temperature at PBL . Definding the reference height:If the reference height z_pblhst(the height of Atmospheric Surface Layer)，BAS set of equation applied；otherwise z_pblhst，MOS does.,2. M
13、ETHODOLOGY,(3) Considering energy balance at limiting cases, then the derived H is further subjected to constraints in the range set by the sensible heat flux at the wet limit Hwet, and at dry limit Hdry in SEBS.Under the dry-limit, the latent heat becomes zero due to the limitation of soil moisture
14、, and the sensible heat flux is at its maximum value. orUnder the wet-limit, where the evaporation takes place at potential rate, (i.e. wet the evaporation is only limited by the available energy under the given surface and atmospheric conditions), the sensible heat flux takes its minimum value.,LAS
15、 measurements :Arou County, east of Qinghai province covered with grasslands and with an altitude about 3000 m The LAS made measurements along a path between transmitter (380324.3N, 1002816.4E) and receiver (380218.1N, 1002725.9E) with distance of 2390 m. Fig.3 The location of LAS on MODIS pixels;Wh
16、ere, T is the transmitter of LAS located on MODIS pixel; R is the receiver of LAS located on MODIS pixel.,3. DATA,MODIS products and preprocessing ALBEDO EMISSIVITYNCEP data and preprocessing,3. DATA, Previous research using empirical relationship : a new approach to derive the value of proposed by
17、Kun Yang(2003) is being used in this paper. According to this, can be considered constant over a short period, since is physically related to the underlying surface and not sensitive to the diurnal variation of atmospheric stability. the specific values of from May to September in 2011 are as follow
18、s (Table 1. ):Table.2 values being used in this paper,3. DATA,4. RESULTS AND ANALYSIS,4.1 Comparison between Sensible heat from SEBS and LASFig.2 Comparsion between SEBS-predicted sensible heat flux and LAS observation from Jul. to Sept.,4. RESULTS AND ANALYSIS,4.1 Comparison between Sensible heat f
19、rom SEBS and LASTable.3 Comparsion between SEBS-predicted sensible heat flux and LAS observation from Jul. to Sept.,4. RESULTS AND ANALYSIS,4.2 Comparison between Sensible heat from SEBS and NCEPAs for means and standard deviation, SEBS outputs showed higher values, suggesting SEBS overestimated sen
20、sible heat with more fluctuations compared to LAS measurements (Table.4). Table.4 Statistics of sensible heat (w/m2) from LAS observation, SEBS-predicted and NCEP sensible heat flux data,4. RESULTS AND ANALYSIS,4.2 Comparison between Sensible heat from SEBS and NCEPTable.5 The Root Mean Square Error
21、 (RMSE), Relative Root Mean Square Error (RRMSE) and Correlation Coefficient (r) of SEBS-predicted sensible heat flux and NCEP sensible heat flux data,4. RESULTS AND ANALYSIS,From July to September, the disparity between SEBS results and LAS measurements was smaller. When results from May and June w
22、ere taken into account, the disparity increased. This was probably related to the vegetation condition. Before July, the surfaces are nearly bare soil with sparse vegetation; From July to the end of August, the surfaces are partially covered by growing grasses. After September the surfaces are cover
23、ed by mature grasses . The better Hs estimation from July to September suggests SEBS is more applicable for dense vegetation.,4. SENSITIVITY ANALYSIS,4.1 SENSITIVITY ANALYSIS According to the sensible heat flux defined by equation in SEBS So we performed a sensitivity analysis on three variables, wh
24、ich are temperature difference between ground surface and reference height ( ), wind speed at PBL( ), and surface roughness for momentum transport ( ).Three typical dates (respectively are 21,June, 11,Aug. and 22,Sept.) were chosen for sensitive analysis. For each date, one parameter was varied and
25、others were fixed. Fig 3-5 showed the results.,Fig 3. Sensitivity of sensible heat flux(H) when varying from 0.01m to 0.4m,4. SENSITIVITY ANALYSIS,Fig 4. Sensitivity of sensible heat flux(H) when varying from 2k to 34k,4. SENSITIVITY ANALYSIS,Fig 5. Sensitivity of sensible heat flux(H) when varying
26、u_pbl from 1 m/s to 20 m/s,4. SENSITIVITY ANALYSIS,Sensitivity analysis showed , and all influenced sensible heat strongly. However, the influence disappeared when sensible heat reached the maximum value under dry limit. Besides, the relationship between and sensible heat flux was linear, while for
27、other two parameters, the relationship was non-linear.,4. SENSITIVITY ANALYSIS,5. DISCUSSION AND CONCLUSION, Although NCEP meteorological data is on 1x1 degree grids, it can still be used with meso-scale remote sensing data to get high-quality sensible heat results given the strong correlation betwe
28、en NCEP and SEBS. NCEP data may not be appropriate for geostationary satellite data to calculate sensible heat at morning or night time when the height of PBL is small. However, LAS measurements were line-averaged over 3km and integrated over 30 minutes. SEBS model outputs were instantaneous and pix
29、el-averaged. The mismatch could be another source of error. To get more accurate sensible heat with SEBS model, local parameterization scheme on roughness length of momentum, and higher resolution meteorological information maybe needed. More in-depth researches are forthcoming in the future.,Thank You !,
