Uncertainties of Eddy Covariance ET in Bushland, Texas.

Accurate measurements of field evapotranspiration (ET) are important for reducing costs associated with irrigation management. Among several options, eddy covariance (EC) systems have key advantages, but also limitations mainly related to an assortment of necessary corrections related to the physical limitations of the measurement system, turbulent transport, and flux exchange heterogeneity. The impact of each source of random errors on ET magnitudes can be quantified with statistical approaches, each providing complimentary information. By using uncertainty analysis, we aim to assess how environmental factors affect estimates of ET from multiple EC systems located in the same field. Four towers with EC systems at 4.6 meters above ground level were assembled and placed on the northern half of a 4.7 ha field cropped with sorghum and subsurface drip irrigated. High frequency (20 Hz) data were collected for 81 consecutive days during the 2015 growing season when sorghum was fully developed, and post-processed with three commonly used high-pass filters and two flux averaging lengths. A 2D footprint analysis indicates that the tri-tower configuration allows for only a small amount of source area overlap, so that the wind direction determines which pair of EC systems can provide independent flux measurements in a two-tower pairing. Relative sampling error for each EC system was calculated using the auto- and cross-covariance of the vertical wind and water vapor concentration raw time series, and a weighted average of the flux across the growing season. This error was similar among all towers except the center tower, which notably had a smaller mean flux and a different model gas analyzer. Daytime error estimates are generally only around 5%, but can be larger in a more turbulent regime. This study may provide new insights into using EC systems for water resource managers and micrometeorologists.