Using Eddy Covariance and Flux Partitioning to Assess Basal, Soil, and Stress Coefficients for Crop Evapotranspiration Models.

Current approaches to scheduling crop irrigation using reference evapotranspiration (ET₀) recommend using a dual-coefficient approach using basal (K_cb) and soil (K_e) coefficients along with a stress coefficient (K_s) to model crop evapotranspiration (ET_c), [e.g. ET_c=(K_s*K_cb+K_e)*ET₀].  However, independently determining K_s, K_cb, and K_e from the combined single coefficient (K_c) is challenging, particularly for K_e, and a new method is needed to more rapidly determine crop coefficients for novel cultivars and cultivation practices.  In this study, we partition Eddy Covariance ET observations into evaporation and transpiration components using correlation structure analysis at multiple irrigated agricultural sites with minimal growing season precipitation.  These include a C4 photosynthetic-pathway species (sugarcane – Sacharum officinarum L.) and a C3 pathway species (peach - Prunus persica) under sub-surface drip and furrow irrigation, respectively.  Both sites showed high overall K_c consistent with their height (>4 m) and full canopy cover.  However, the results showed differences in K_e, with the drip-irrigated sugarcane having a K_e that was ~30% of the peach K_e.  There was no significant relationship (r²<0.05) between root zone soil moisture in sugarcane, and observed K_cb*K_s, indicating that there was no stress (K_s=1).  Comparison of shallow soil moisture and soil diurnal temperature variation with K_e in Peach showed a significant relationship that could be used to constrain model estimates of K_e.  Partitioning of K_c into K_cb and K_e resulted in a stronger regression (r=0.65) between the Normalized Differential Vegetation Index (NDVI) and K_cb in sugarcane than between NDVI and K_c (r=0.41).  The results indicate the strong potential for correlation structure flux partitioning to improve crop ET coefficient determination and its relationship with both satellite and soil observations.  However, further work on this method is needed to reduce the uncertainty in the partitioning and to assess the validity of partitioned fluxes and coefficients against independent evaporation and transpiration observations.