Assessing the impact of end‐member selection on the accuracy of satellite‐based spatial variability models for actual evapotranspiration estimation

Abstract

[1] This study examines the impact of end-member (i.e., hot and cold extremes) selection on the performance and mechanisms of error propagation in satellite-based spatial variability models for estimating actual evapotranspiration, using the triangle, surface energy balance algorithm for land (SEBAL), and mapping evapotranspiration with high resolution and internalized calibration (METRIC) models. These models were applied to the soil moisture-atmosphere coupling experiment site in central Iowa on two Landsat Thematic Mapper/Enhanced Thematic Mapper Plus acquisition dates in 2002. Evaporative fraction (EF, defined as the ratio of latent heat flux to availability energy) estimates from the three models at field and watershed scales were examined using varying end-members. Results show that the end-members fundamentally determine the magnitudes of EF retrievals at both field and watershed scales. The hot and cold extremes exercise a similar impact on the discrepancy between the EF estimates and the ground-based measurements, i.e., given a hot (cold) extreme, the EF estimates tend to increase with increasing temperature of cold (hot) extreme, and decrease with decreasing temperature of cold (hot) extreme. The coefficient of determination between the EF estimates and the ground-based measurements depends principally on the capability of remotely sensed surface temperature (T s ) to capture EF (i.e., depending on the correlation between T s and EF measurements), being slightly influenced by the end-members. Varying the end-members does not substantially affect the standard deviation and skewness of the EF frequency distributions from the same model at the watershed scale. However, different models generate markedly different EF frequency distributions due to differing model physics, especially the limiting edges of EF defined in the remotely sensed vegetation fraction (f c ) and T s space. In general, the end-members cannot be properly determined because (1) they do not necessarily exist within a scene, varying with the spatial extent, resolution, and quality of satellite images being used and/or (2) different operators can select different end-members. Furthermore, the limiting edge of EF ¼ 0 in the f c -T s space varies with the model, with SEBAL-type models having inherently an increasing curvilinear limiting edge of EF ¼ 0 with f c . The spatial variability models therefore require careful calibration in order to deduce reasonable EFlimiting edges and then confine the magnitudes of EF estimates.Citation: Long, D., and V. P. Singh (2013), Assessing the impact of end-member selection on the accuracy of satellite-based spatial variability models for actual evapotranspiration estimation, Water Resour.

Evapotranspiration Estimation with Remote Sensing and Various Surface Energy Balance Algorithms—A Review

“…From the above review, it is found that SEBAL/METRIC models had high potential for successful ET estimates in the semi-arid US by comparing the derived ET with lysimeter observations [99]. Nevertheless, in the recently conducted study, the context-dependency of SEBAL, METRIC, and triangular models has been noticed [92], which has already been discussed in details in earlier section. Although the reviewed methods show an enough potential and viability for evapotranspiartion estimation on a regional scale, these methods have several drawbacks as well, which limit their applicability and accuracy to some extent.…”

“…Identifying the dry pixels is the most important aspect in SEBAL, while the wet pixels are frequently spotted at a location of well watered areas or over a relatively large, calm water surface. However, a recent study [92] indicates that the SEBAL model tends to be context-dependent, i.e., wet/dry pixels (edges) required to trigger SEBAL may not necessarily exist within a specific extent of an image.…”

“…Besides its several advantages, it has several drawbacks as well. Major disadvantages of this method are that (1) Subjective specifications of representative hot/dry and wet/cool pixels within the image are required [92,93] to determine model parameters a and b. The resulting H flux and ET estimates from SEBAL can vary with differing extreme pixels selected by the operator, domain size, and spatial resolution of satellite sensors [94]; (2) Over mountainous regions, some adjustments are required based on a digital elevation model for T s and u to account for the lapse rate [95,96]; (3) Estimated H is greatly affected by the errors in surface-air temperature differences or surface temperatures measurements; and (4) Ignoring the effect of radiometer viewing angle, can cause variation in T s by several degrees for some images.…”

“…Once surface temperature, T s , and dT are calculated corresponding to hot and cold conditions, the linear relationship as indicated in Equation (18) is defined. However, the context-dependency of SEBAL, METRIC, and triangular models has been indicated in a recently conducted study [92]. They indicated that the wet/dry pixels (edges) required to trigger these models may not necessarily exist within a specific extent of an image.…”

“…For instance, one would not be able to select a hot pixel from a large homogeneous forest. Also, there is no other alternative for the SEBAL/METRIC models to automatic selection of extreme pixels from images with varying extents, spatial resolutions, and clouds [94,100], Furthermore, even though the extremes can be properly selected from relatively large images that probably entail hot and cold extremes reflecting surface conditions after cloud and terrain effects are favorably reduced/removed, the SEBAL-type algorithms appear to be limited in providing reasonable ET patterns due mostly to constant coefficients "c" and "d" in the SEBAL algorithm that do not accommodate the effect of variations in fractional vegetation cover on ET extremes [92,93].…”