Modeling sensible heat flux for vegetated surfaces through an optimized surface aerodynamic temperature approach
Date
2019
Authors
Costa Filho, Edson, author
Chavez, Jose L., advisor
Ham, Jay M., committee member
Venayagamoorthy, Karan, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Agricultural water management advancements rely on improved methods to accurately determine crop water use. Crop evapotranspiration modeling based on the surface energy balance depends on the accurate estimation of all incoming and outgoing heat fluxes at the surface level. This thesis particularly goal is to improve sensible heat flux estimates for row crops through an optimized aerodynamic surface temperature (To) approach based on remote sensing and weather data. Empirical linear and non-linear To models were developed based on percent cover, surface temperature, air temperature, and a new variable named turbulent mixing row resistance using data collected at the USDA-ARS Research Farm located in Greeley (CO). The experiment took place in two sub-surface drip irrigation corn fields with different irrigation water management practices in 2017-2018. Sensible heat flux were measured using LAS, eddy covariance, aerodynamic profile, and Bowen ratio methods. Remote sensing data were measured on-site using a radiometer. The fields were considered a point in space. Data from Aimes (IA) and Rocky Ford (CO) were used to assess proposed model performances under different locations and in comparison to published To models. The results have indicated that the optimized linear To models performed better than the non-linear and published models approaches, indicating that the introduction of percent cover and the new variable has provided reliable results under different data sets. The linear proposed To approaches improved sensible heat flux estimation, on average, in 33 % and 28 % for the deficit and fully irrigated field at LIRF in comparison to the sensible heat based on published To models. Sensible heat flux modeling results were better for the modeling approaches considering the empirical linear To model than the non-linear approaches for all three data set tested.