Tham khảo tài liệu 'evapotranspiration remote sensing and modeling part 17', khoa học tự nhiên, công nghệ sinh học phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | Operational Remote Sensing of ET and Challenges 469 Except for the LSM applications none of the listed energy balance methods in and of themselves go beyond the creation of a snapshot of ET for the specific satellite image date. Large periods of time exist between snapshots when evaporative demands and water availability from wetting events cause ET to vary widely necessitating the coupling of hydrologically based surface process models to fill in the gaps. The surface process models employed in between satellite image dates can be as simple as a daily soil-surface evaporation model based on a crop coefficient approach for example the FAO-56 model of Allen et al. 1998 or can involve more complex plant-air-water models such as SWAT Arnold et al. 1994 SWAP van Dam 2000 HYDRUS Simunek et al. 2008 Daisy Abrahamsen and Hansen 2000 etc. that are run on hourly to daily timesteps. Problems with use of absolute surface temperature Error in surface temperature Ts retrievals from many satellite systems can range from 3 - 5 K Kalma et al. 2008 due to uncertainty in atmospheric attenuation and sourcing surface emissivity view angle and shadowing. Hook and Prata 2001 suggested that finely tuned Ts retrievals from modern satellites could be as accurate as K. Because near surface temperature gradients used in energy balance models are often on the order of only 1 to 5 K even this amount of error coupled with large uncertainties in the air temperature fields makes the use of models based on differences in absolute estimates of surface and air temperature unwieldy. Cleugh et al. 2007 summarized challenges in using near surface temperature gradients dT based on absolute estimates of Ts and air temperature Tair attributing uncertainties and biases to error in Ts and Tair uncertainties in surface emissivity differences between radiometrically derived Ts and the aerodynamically equivalent Ts required as a sourcing endpoint to dT. The most critical factor in the physically based
