Parametric sensitivity analysis of precipitation at global and local scales in the Community Atmosphere Model CAM5

Abstract

We investigate the sensitivity of precipitation characteristics (mean, extreme, and diurnal cycle) to a set of uncertain parameters that influence the qualitative and quantitative behavior of cloud and aerosol processes in the Community Atmosphere Model (CAM5). We adopt both the Latin hypercube and QuasiMonte Carlo sampling approaches to effectively explore the high-dimensional parameter space and then conduct two large sets of simulations. One set consists of 1100 simulations (cloud ensemble) perturbing 22 parameters related to cloud physics and convection, and the other set consists of 256 simulations (aerosol ensemble) focusing on 16 parameters related to aerosols and cloud microphysics. In the cloud ensemble, six parameters having the greatest influences on the global mean precipitation are identified, three of which (related to the deep convection scheme) are the primary contributors to the total variance of the phase and amplitude of the precipitation diurnal cycle over land. The extreme precipitation characteristics are sensitive to a fewer number of parameters. Precipitation does not always respond monotonically to parameter change. The influence of individual parameters does not depend on the sampling approaches or concomitant parameters selected. Generally, the Generalized Linear Model is able to explain more of the parametric sensitivity of global precipitation than local or regional features. The total explained variance for precipitation is primarily due to contributions from the individual parameters (75-90% in total). The total variance shows a significant seasonal variability in midlatitude continental regions, but very small in tropical continental regions.

Validity of parameter optimization in improving MJO simulation and prediction using the sub-seasonal to seasonal forecast model of Beijing Climate Center

“…It is assumed that a large part of the convection and precipitation biases in climate models may come from the cloud and convection parameterization schemes that include many uncertain parameters (e.g., Jackson et al 2008;Yan et al 2014;Qian et al 2015;Yang et al 2015). Especially, significant sensitivities of MJO simulation to deep and shallow convection schemes and relevant parameters have been shown by previous studies.…”

Can nudging be used to quantify model sensitivities in precipitation and cloud forcing?

“…TAU has been shown by various studies to have substantial impact on the precipitation characteristics including the global mean precipitation rate [e.g., Yang et al, 2013;Qian et al, 2015], the extreme precipitation events [Williamson, 2013;Qian et al, 2015], and the amplitude of diurnal cycle . Rhminl is often tuned to adjust the cloud amount and the energy Journal of Advances in Modeling Earth Systems 10.1002/2016MS000659 balance of the global atmosphere, and also has been shown by many studies to have important impact on the global mean precipitation rate [e.g., Qian et al, 2015].…”

“…In free-running simulations, we first ran the control simulation (F_def), with the default value of TAU (1 h) and rhminl (0.8875). We then changed TAU to 8 h or 0.5 h, the high and low ends of the range of TAU identified in Qian et al [2015], in simulations of F_TAU_H or F_TAU_L, respectively. Likewise, we perturbed TAU to 8 and 0.5 h in nudged UVT and nudged UV simulations as well (N_UVT_H, N_UVT_L, N_UV_H, and N_UV_L).…”

“…However, such long simulations can be computationally expensive, or even impractical, for high-resolution models or when many experiments are needed. One such case is the uncertainty quantification (UQ) studies that usually require several hundred or even thousands of simulations [Murphy et al, 2004;Carslaw et al, 2013;Jackson et al, 2008;Zhang et al, 2012;Yang et al, 2012Yang et al, , 2013Yan et al, 2015;Qian et al, 2015]. Optimizing the simulation strategy to reduce the computational cost but meanwhile retain an accurate characterization of model sensitivity is therefore important for the calibration of models designed for future climate prediction.…”

SUPECA kinetics for scaling redox reactions in networks of mixed substrates and consumers and an example application to aerobic soil respiration

“…This problem has been rectified in ECA kinetics (Tang and Riley, 2013a), which was shown to predict much more accurate parametric sensitivity than Monod kinetics when compared with analytical solutions (Tang, 2015). Since the success of all model calibrations relies on the accuracy of modeled response variables' sensitivity to model parameters (e.g., Wang et al, 2001;Williams et al, 2005;Tang and Zhuang, 2009;van Werkhoven et al, 2009;Qian et al, 2015), and plant-microbial competitions of nutrients often occur under high consumer abundances with respect to their substrates (as corroborated by the nitrogen and phosphorus limitations that are commonly observed in natural ecosystems; e.g., Vitousek et al, 2010), developing robust biogeochemical models requires substrate kinetics that gives accurate parametric sensitivities under a wide range of parameter values.…”