Energy balance closure using remote sensing

Another approach to obtaining areal estimates of actual evapotranspiration is to make use of the spatial information in remote sensing images. This was an approach that was pioneered in the UK by Holwill and Stewart (1992) and in the Netherlands by Bastiaanssen et al. (1998, 2005) whose SEBAL program is used on a routine operational basis there. In this approach, multispectral images can be used to identify different classes of vegetation, multidirectional sensing can be used to estimate the effective albedo of the surface under different conditions, and thermal infrared wavelength images can be used to estimate the temperature of the surface. Then, given some information about incoming radiation, air temperatures and wind speed in an area, and a number of parameters, the various flux terms of the energy balance can be estimated so that patterns of evapotranspiration can be derived. The effects of slope angle, aspect and horizon shading can also be taken into account in accounting for the net radiation. Cloudiness can also be determined from thermal and visual wavelength images. In some circumstances it is possible to use active and passive microwave sensing to estimate the near-surface soil moisture content, but microwaves do not penetrate more than a few centimetres into the soil and so this is not always a strong control on the effective water availability in the full profile, which may also be affected by downslope subsurface water flows.

Studies of this type have revealed some very interesting patterns of estimated actual evapotranspiration rates in the landscape (e.g. Fig. 10.10), but studies of the uncertainties associated with the energy balance closure have suggested that the absolute values of Et predicted in this way may be rather uncertain (e.g. Franks et al., 1999). Patterns of wind speed, in particular, which are important in controlling the local

Estimated latent heat fluxes (W/m2)

Estimated latent heat fluxes (W/m2)

100 - 140 Below 100

100 - 140 Below 100

Fig. 10.10 Derived actual evapotranspiration estimates forthe FIFE site in Kansas at 1500 on l5August 1987 (after Franks and Beven, 1999, reproduced by permission of American Geophysical Union).

aerodynamic resistance can be difficult to predict over a complex landscape, especially in mountain areas.

0 0

Post a comment