The Common Land Model

Abstract

Scientists from several institutions and with different research backgrounds have worked together to develop a prototype modular land model for weather forecasting and climate studies. This model is now available for public use and further development.C limate and weather forecasting models require the energy, water, and momentum fluxes across the land-atmosphere interface to be specified. Various land surface parameterizations (LSPs), ranging from the simple bucket-type LSP in the 1960s to the current soil-vegetation-atmosphere interactive LSP, have been developed in the past four decades to calculate these fluxes. The Project for Intercomparison of Land Surface Parameterization Schemes (PILPS) has demonstrated that, even with the same atmospheric forcing data and similar land surface parameters, different LSPs still give significantly different surface fluxes and soil wetness, partly because of the differences in the formulations of individual processes and coding architectures in participant models . On the other hand, most LSPs share many common components, suggesting the need to develop a publicly available common land model with a modular structure that could facilitate the exploration of new issues, less repetition of past efforts, and sharing of improvements and refinements contributed by different groups.The Common Land Model (CLM) effort dates back to the mid-1990s and has evolved through various workshops and e-mail correspondence. The initial motivation was to provide a framework for a truly community-developed land component of the National Center for Atmospheric Research (NCAR) Community Climate System Model (CCSM). Interest in applying CLM came from the Goddard Space Flight Center (GSFC) Data Assimilation Office (DAO), which was implementing the Mosaic model (Koster and Suarez 1992), and the Center for Ocean-Land-Atmosphere Studies (COLA) scientists, who were revising their Simplified Simple Biosphere Model (SSiB) (Xue et al. 1991). We also established ties to groups performing carbon cycle and ecological modeling.In developing CLM, we attempted to combine the best features of three existing successful and relatively

Multi-objective parameter optimization of common land model using adaptive surrogate modeling

“…The Common Land Model proposed by Dai et al (2003Dai et al ( , 2004 and Ji and Dai (2010) is one of the most widely used LSM in the world. It combines the advantages of the LSM (Bonan, 1996), the biosphere-atmosphere transfer scheme (BATS) (Dickinson et al, 1993) and the Institute of Atmospheric Physics LSM (IAP94) (Dai and Zeng, 1997;.…”

“…Parameter optimization can improve the model performance if the model physics are correct, but is helpless if the model structure is inconsistent with the true underlying physical processes. Although CoLM's performance of simulating frozen soil and snow cover has been evaluated in the experiment in Valdai, Russia (Dai et al, 2003), the situation of Heihe in China can be very different. For instance, in CoLM the soil depth is set to 2.86 m globally, but actually the soil depth varies in different places.…”

“…This framework involves the following steps: (1) a parameter screening step using global sensitivity analysis to identify the most sensitive parameters to be included in the optimization; (2) surrogate modeling that can emulate the response surface of the dynamic system model to the change in parameter values; (3) an adaptive sampling strategy to improve the efficiency of the surrogate model construction; and (4) a multi-objective optimization step to optimize the most sensitive parameters of the dynamic system model. In this paper, we will illustrate this parametric uncertainty quantification framework with the Common Land Model (CoLM), a widely used, physically based LSM developed by Dai et al (2003Dai et al ( , 2004 and Ji and Dai (2010). The work related to parameter screening and surrogate modeling-based optimization (adaptive surrogate model-based optimization strategy -ASMO) method for single objective has already been published Wang et al, 2014).…”

Estimating groundwater storage changes in the Mississippi River basin (USA) using GRACE

“…Soil moisture and snow water equivalent (SWE) were simulated by the Noah (Ek et al 2003), Common Land Model version 2 (CLM2; Dai et al 2003), and Mosaic (Koster and Suarez 1996) land surface models driven by the Global Land Data Assimilation System (GLDAS; Rodell et al 2004a). GLDAS ingests satellite-and ground-based observational data products in order to generate optimal fields of land surface states (e.g., soil moisture, snow, surface temperature) and fluxes (e.g., evapotranspiration, ground heat flux).…”

A non-linear and stochastic response surface method for Bayesian estimation of uncertainty in soil moisture simulation from a land surface model

“…Schulz and Beven, 2003;Schulz et al, 2001;Franks et al, 1998;Franks and Beven, 1997). A more realistic Bayesian assessment of uncertainty requires the application of GLUE to physically complex operational LSMs such as Common Land Model (CLM; Dai et al, 2003), NOAH-LSM (Pan and Mahrt, 1987), BATS (Dickinson et al, 1986) or SiB (Sellers et al, 1986). Uncertainty assessment of these models are important because, despite their physical complexity, they nevertheless suffer from parameter equifinality where a wide range of parameter sets exhibit equally acceptable simulations against data available.…”