REFINING COMPONENTS OF SATELLITE BASED SURFACE ENERGY BALANCE MODELS FOR FORESTS AND STEEP TERRAIN

Satellite based surface energy balance models are being used in an increasing number of water management application in the western US. These models generally determine evapotranspiration (ET) at Landsat satellite scale utilizing the thermal band of Landsat.  Modeling Evapotranspiration at High Resolution with Internalized Calibration (METRIC) (Allen and others, 2007) is one of these models. METRIC uses the procedure Calibration using Inverse Modeling at Extreme Conditions (CIMEC) to derive a unique calibration for each image. For the operational use of METRIC it is often necessary to employ a variety of refinements during the image processing to account for the roughness characteristics of land use types other than agriculture and to account for changes in elevation. Other tuning parameters and the rationale behind them have been described by Allen and others (2008).

The momentum roughness length (z_om) is often estimated in METRIC for tall and potentially sparse vegetation such as forests, riparian areas or orchards using a z_om function developed by Perrier (1982) that is based on LAI and tree canopy architecture:

where the stand height h for  forest vegetation is estimated as 2.5 times the Leaf Area Index (LAI), which results  in a maximum tree height of 15 m, when LAI = 6, but this can be adjusted. The parameter “a” is an adjustment factor for the LAI distribution within the canopy with a = (2f) for f ≥ 0.5 and a = (2(1-f))^-1 for f < 0.5. The factor f is the proportion of LAI lying above h/2, i.e., f = 0.3, 0.5 and 0.7 indicates sparsely topped canopy, uniform canopy and top heavy canopy, respectively.

Air temperature generally decreases by 6.5 to 10 °C for each 1000 m elevation increase under neutral atmospheric conditions. Because the surface temperature is in strong equilibrium with the air temperature, a similar decrease in surface temperature can usually be observed. During the CIMEC procedure of METRIC, a relationship is established between the near-surface temperature difference (dT) and the surface temperature. The surface temperature is then adjusted to a common reference elevation for accurate prediction of dT. Otherwise, the surface temperature of areas at higher elevation can be misinterpreted to be low because of evaporative cooling, rather than due to elevation increase. A ‘delapsed' (artificial) surface temperature map is created for purposes of estimating dT. Elevation data are provided by a digital elevation map. For areas having steep mountains a dual-stage delapse correction is used, with differentiated lapse rates for flat terrain and for mountainous terrain.

References

Allen R.G., Kjaersgaard J., and Garcia M., 2008, Fine-tuning components of inverse-calibrated, thermal-based remote sensing models for evapotranspiration: The 17^th William Pecora Symposium, Denver, Co., November 2008.

Allen R.G., Tasumi M., and Trezza R., 2007, Satellite-based Energy Balance for Mapping Evapotranspiration with Internalized Calibration (METRIC) – Model: 0 J. Irrig. Drain. Engrg., v 133(4), p. 380-394.

Perrier A., 1982, Land Surface Processes: Vegetation, in Eagelson P., ed., Land Surface Processes in Atmospheric General Circulation Models: Cambridge University Press, p. 395 – 448.