Quantifying parameter sensitivity, interaction, and transferability in hydrologically enhanced versions of the Noah land surface model over transition zones during the warm season

Abstract

[1] We use sensitivity analysis to identify the parameters that are most responsible for controlling land surface model (LSM) simulations and to understand complex parameter interactions in three versions of the Noah LSM: the standard version (STD), a version enhanced with a simple groundwater module (GW), and version augmented by a dynamic phenology module (DV). We use warm season, high-frequency, near-surface states and turbulent fluxes collected over nine sites in the U.S. Southern Great Plains. We quantify changes in the pattern of sensitive parameters, the amount and nature of the interaction between parameters, and the covariance structure of the distribution of behavioral parameter sets. Using Sobol 0 's total and first-order sensitivity indexes, we show that few parameters directly control the variance of the model response. Significant parameter interaction occurs. Optimal parameter values differ between models, and the relationships between parameters also change. GW decreases unwarranted parameter interaction and appears to improve model realism, especially at wetter study sites. DV increases parameter interaction and decreases identifiability, implying it is overparameterized and/or underconstrained. At a wet site, GW has two functional modes: one that mimics STD and a second in which GW improves model function by decoupling direct evaporation and base flow. Unsupervised classification of the posterior distributions of behavioral parameter sets cannot group similar sites based solely on soil or vegetation type, helping to explain why transferability between sites and models is not straightforward. Our results suggest that the a priori assignment of parameters should also consider the climatic conditions of a study location.

Exploring the impact of forcing error characteristics on physically based snow simulations within a global sensitivity analysis framework

“…The goal of sensitivity analysis is to determine which input factors (F and θ ) are most important to specific outputs (Y) (Matott et al, 2009). Sensitivity analyses often focus more on the model parameter array (θ ) than on the forcing matrix (Foglia et al, 2009;Herman et al, 2013;Li et al, 2013;Nossent et al, 2011;Rakovec et al, 2014;Rosero et al, 2010;Rosolem et al, 2012;Tang et al, 2007;van Werkhoven et al, 2008). However, recent analyses have considered other input factors and sources of uncertainty (e.g., Baroni and Tarantola, 2014;Schoups and Hopmans, 2006).…”

“…3.2.3) as among the most important considerations. While it is common practice in sensitivity analysis to assume a uniform distribution when sampling model parameters (e.g., Campolongo et al, 2011;Rosero et al, 2010), this may fail to represent the real distribution of errors in meteorological forcing data, as the uniform distribution implies that extreme and small biases are equally probable. It is more likely that real error distributions more closely resemble non-uniform distributions, with higher probability of smaller biases and lower probability of more extreme biases (e.g., normal distributions).…”

Effective soil moisture estimate and its uncertainty using multimodel simulation based on Bayesian Model Averaging

“…Feasible ranges of the parameters (i.e., search spaces in a modified-microGA) for each model were defined based on literature related to the model parameter sensitivity and to accommodate a diversity of soils ranging from clay to sandy loam Liu et al, 2004;Ines and Mohanty, 2008;Rosero et al, 2010;Shin et al, 2012].…”

“…In this study, we adapted the uncoupled mode as a one-dimensional (1-D), physically based model for estimating the soil moisture dynamics at field scales. Noah LSM calculates the total evapotranspiration by summing the direct evaporation from top soil layer, canopy evaporation, and potential Penman-Monteith transpiration [Rosero et al, 2010]. The model has typically four soil layers with the thicknesses of 10, 30, 60, and 100 cm (total soil depth of 200 cm), but we changed top soil layer depth to 5 cm (while maintaining the same total root zone depth) to be compared to the soil moisture observation (top 5 cm) in this study.…”

A new approach to identify the sensitivity and importance of physical parameters combination within numerical models using the Lund–Potsdam–Jena (LPJ) model as an example

“…The interaction among parameters was ignored during the process (Jackson et al 2003;Bastidas et al 2006). Other methods have also been employed to explore the sensitivity of certain physical parameters, such as the adjoint method, factorial experimentation, and the Multi-Objective Generalized Sensitivity Analysis (MOGSA) method (Henderson-Sellers 1992;Wang et al 2001;Rayner et al 2005;Bastidas et al 2006;Rosero et al 2010). The adjoint method is based on linear approximation, and while it may be valid for small parameter errors over short time periods, it is less applicable to large parameter errors and long integration times.…”