Stone content influence on land surface model simulation of soil moisture and evapotranspiration at Reynolds Creek Watershed

Authors: PARAJULI, KSHITIJ - Utah State University; ZHAO, LIN - Utah State University; JONES, SCOTT - Utah State University; TARBOTON, DAVID - Utah State University; HIPPS, LAWRENCE - Utah State University; TORRES-RUA, ALFONSO - Utah State University; SADEGHI, MORTEZA - Utah State University; Flerchinger, Gerald; ROCKHOLD, MARK - Pacific Northwest National Laboratory

Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/23/2020
Publication Date: 8/21/2020
Citation: Parajuli, K., Zhao, L., Jones, S., Tarboton, D., Hipps, L., Torres-Rua, A., Sadeghi, M., Flerchinger, G.N., Rockhold, M.L. 2020. Stone content influence on land surface model simulation of soil moisture and evapotranspiration at Reynolds Creek Watershed. Journal of Hydrometeorology. 21(8):1889-1904. https://doi.org/10.1175/JHM-D-19-0075.1.
DOI: https://doi.org/10.1175/JHM-D-19-0075.1

Interpretive Summary: There has been considerable advancement in spatiotemporal resolution of remote sensing and ground-based measurements leading to better performance of land surface models in simulating water and energy fluxes from the land surface. However, limitations remain for simulating water- and energy- fluxes from soil due to inadequate representation of subsurface processes and soil parameters. In this study, we investigate the performance of the Noah-Multiphysics (Noah-MP) land surface model for simulating soil moisture and evapotranspiration under various soil parameterizations. A comprehensive field data set including soil property measurements within the soil profile, micrometeorological data and soil moisture from different depths were obtained from the Lower Sheep sub-catchment within the Reynolds Creek Experimental Watershed in Southwestern Idaho. Data were employed to drive Noah-MP and assess the simulation results. We evaluated the performance of Noah-MP considering four different scenarios: (i) homogeneous soil profile with default hydraulic parameters; (ii) 5-layer soil profile with default hydraulic parameters for each layer; (iii) 5-layer soil profile with hydraulic parameters from field observations for each layer; and (iv) 5-layer soil profile with hydraulic parameters from field observations modified for stones for each layer. Additional simulations were performed using the HYDRUS-1D numerical model employing more detailed representation of soil hydraulic functions. These included: (i) neglecting the presence of stones and (ii) considering the effect of stone content. Each experiment was forced with the same set of initial conditions, atmospheric forcing and vegetation parameters. The best simulation fit to measured soil moisture was obtained with the HYDRUS-1D numerical model. Significant improvement in the Noah-MP soil moisture simulation was achieved using the improved soil parameters. The Noah-MP model informed of stone content impact and using detailed soil properties provided the better estimation of evapotranspiration as compared to that with eddy covariance measurements. We conclude that improvement in representation of soil properties along with stone content information can substantially improve the ability of land surface models to simulate soil water flow and boundary fluxes.

Technical Abstract: There has been considerable advancement in spatiotemporal resolution of remote sensing and ground-based measurements leading to better performance of land surface models in simulating water and energy fluxes from the land surface. However, limitations remain for simulating water- and energy- fluxes from soil due to inadequate representation of subsurface processes and soil parameters. In this study, we investigate the performance of the Noah-Multiphysics (Noah-MP) land surface model for simulating soil moisture and evapotranspiration under various soil parameterizations. A comprehensive field data set including soil property measurements within the soil profile, micrometeorological data and soil moisture from different depths were obtained from the Lower Sheep sub-catchment within the Reynolds Creek Experimental Watershed in Southwestern Idaho. Data were employed to drive Noah-MP and assess the simulation results. We evaluated the performance of Noah-MP considering four different scenarios: (i) homogeneous soil profile with default hydraulic parameters; (ii) 5-layer soil profile with default hydraulic parameters for each layer; (iii) 5-layer soil profile with hydraulic parameters from field observations for each layer; and (iv) 5-layer soil profile with hydraulic parameters from field observations modified for stones for each layer. Additional simulations were performed using the HYDRUS-1D numerical model employing more detailed representation of soil hydraulic functions. These included: (i) neglecting the presence of stones and (ii) considering the effect of stone content. Each experiment was forced with the same set of initial conditions, atmospheric forcing and vegetation parameters. The best simulation fit to measured soil moisture was obtained with the HYDRUS-1D numerical model. Significant improvement in the Noah-MP soil moisture simulation was achieved using the improved soil parameters. The Noah-MP model informed of stone content impact and using detailed soil properties provided the better estimation of evapotranspiration as compared to that with eddy covariance measurements. We conclude that improvement in representation of soil properties along with stone content information can substantially improve the ability of land surface models to simulate soil water flow and boundary fluxes.