Upscaling of evapotranspiration fluxes from instantaneous to daytime scales for thermal remote sensing applications

Authors: Cammalleri, Carmelo; Anderson, Martha; Kustas, William - Bill

Submitted to: Hydrology and Earth System Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/19/2014
Publication Date: 3/19/2014
Publication URL: http://handle.nal.usda.gov/10113/59952
Citation: 3Cammalleri, C.N., Anderson, M.C., Kustas, W.P. 2014. Upscaling of evapotranspiration fluxes from instantaneous to daytime scales for thermal remote sensing applications. Hydrology and Earth System Sciences.

Interpretive Summary: Satellite imagery can be used to map evapotranspiration (ET) or water use over large areas. This capability is beneficial for improved water management in agricultural systems. However, most satellites pass over a target area on the land surface on a daily or less frequent basis providing only instantaneous snapshots of water use at one time of day. To be useful for management, these snapshot estimates must be upscaled to daily total water use. In this study we investigate several mehtods for upscaling instantaneous ET to daily total ET using ground-based measurements of water loss. A statistical approach is taken to account for expected errors in the water use estimates, to ensure that these errors do not lead to false conclusions about an optimal approach. The analysis determined that an instantaneous water use is most reliably upscaled using hourly measurements of solar radiation - in other words, changes in water use during the day follow changes in sunshine amount and both curves are reasonably similar. Using solar radiation as an upscaling metric yielded the lowest errors in daily evapotranspiration seasonally and regionally, and appears to be a robust methodology for obtaining daily water use estimates from satellite imagery. This methodology will ultimately be incorporated in decision suport tools for improved water management of agricultura systems.

Technical Abstract: Four upscaling methods for estimating daytime evapotranspiration (ET) from single time-of-day snapshots, as commonly retrieved using remote sensing, were compared. These methods are based on the assumption of self-preservation of the ratio between ET and a given reference variable over the daytime hours. The analysis was performed using eddy covariance data collected at 12 AmeriFlux towers, sampling a range in climate and landcover. The choice of energy budget closure method significantly impacted performance using different scaling fluxes. Therefore, a statistical evaluation approach was adopted to better account for the inherent uncertainty in eddy covariance ET fluxes. Overall, this approach suggests that at-surface solar radiation is the most robust reference variable amongst those tested, due to high accuracy of upscaled fluxes and absence of systematic biases. Top-of-atmosphere irradiance was also tested and proved to be reliable under near clear-sky conditions, but tended to overestimate the observed daytime ET during cloudy days. Use of reference ET as a scaling flux did not perform as well as the solar radiation method, but resulted in a similar reduced seasonality in error magnitude. Finally, the commonly-used evaporative fraction method returned satisfactory results only in July and August, and tended to underestimate the observations from November to January at the flux sites studied.