Development of a multi-scale data assimilation system for model-observation integration and climate model evaluation

Description

To improve our understanding and the representation of subgrid processes in climate models, an increasing number of ground-based long-term observing systems have been established. These systems focus on detailed measurements over a domain of a typical climate model grid size. Among them, ARM has been collecting data related to radiation, clouds and precipitation at three primary sites, the Southern Great Plains (SGP), the North Slope of Alaska (NSA), and the Tropical West Pacific (TWP), for approximately 20 years. A well-established approach to use ARM-like measurements in climate model evaluation is jointly using the single column models (SCMs), cloud resolving models (CRMs), and/or large eddy simulations (LESs). To enhance the effectiveness of this approach, we have developed multi-scale data assimilation (MS-DA) system on top of the NCEP Gridpoint Statistical Interpolation (GSI) system and implemented in the Weather Research and Forecasting (WRF) model at the cloud resolving resolution (WRF-CRM) over the SGP site. It is demonstrated that the MS-DA effectively assimilates the dense ARM in-situ observations and high-resolution satellite data, thus significantly reduces uncertainties in the pure WRF-CRM simulation. We have used the WRF-CRM simulation constrained by the MS-DA to derive multi-scale forcing that is used to drive SCMs, CRMs, and LESs, expand the large-scale forcing parameters with hydrometeor forcing being included that are not provided in the existing continuous forcing product, and characterize dependency of large-scale forcing on domain-size representing the grid-size of global models, sub-grid processes, and cloud-regimes.