1171 
Fig. 3: accuracy of soil moisture estimation from 
ERS-1 a° backscattering coefficient. 
4. COMBINE USE OF RADAR AND THERMAL DATA 
4.1. Theory and method 
Fig. 4: general scheme of the approach. 
The letters (a), (b) and (c) refer to 
the three models described below. 
4.1.1 Sensible heat flux model: as shown in Fig. 4 (a), the sensible heat flux is estimated here according to a 
two-layer formulation model as proposed by Shuttleworth and Wallace (1985), since the classical one layer 
approach is not designed to take into account aerodynamic exchanges between soil and vegetation over sparse 
canopy unless some extra-modelling (kB-1 method, Prevot et al., 1993a). This model is based on a system of 
temperatures and resistances between soil, vegetation and air mass, controlling sensible heat fluxes between 
each component: 
T-T T-T 
— - + — - 
1 r a r o 
1 + -2-+-2- 
r r. 
(W/m 2 ) 
(5) 
where a, c and s indices correspond respectively to air, canopy and soil temperatures and T and r stand for 
temperature and aerodynamic resistance. In addition to these input variables, aerodynamic resistances 
calculation needs wind speed values and some classical vegetation properties concerning roughness estimates: 
mean height, LAI and fraction cover. 
4.1.2 Soil temperature model: {Fig. 4 part (b)) numerous studies demonstrated the strong dependence of the 
actual to potential soil evaporation ratio (E/Ep) upon soil surface resistance or more simply upon top surface 
soil moisture (Deardorff, 1977; Chanzy et al., 1993). Moreover, soil temperature is linearly related to this ratio