Development and evaluation of a mosaic approach in the WRF‐Noah framework

Abstract

[1] The current Weather Research and Forecasting (WRF)-Noah modeling framework considers only the dominant land cover type within each grid cell, which here is referred to as the "dominant" approach. In order to assess the impact of subgrid-scale variability in land cover composition, a mosaic/tiling approach (hereafter the "mosaic" approach) is implemented into the coupled WRF-Noah modeling system. In the mosaic approach, a certain number (N) of tiles, each representing a land cover category, is considered within each grid cell. WRF simulations of a clear sky day and a rainfall period over a heterogeneous urban/suburban setting show that the two approaches generate differences in the surface energy balance, land surface temperature, near-surface states, boundary layer growth, as well as rainfall distribution. Evaluation against a variety of observational data (including surface flux measurements, the MODIS land surface temperature product, and radar rainfall estimates) indicates that, compared to the dominant approach, the mosaic approach has a better performance. In addition, WRF-simulated results with the mosaic approach are less sensitive to the spatial resolution of the grid: Larger differences are observed in simulations of different resolutions with the dominant approach. The effect of increasing the number of tiles (N) on the WRF-simulated results is also examined. When N increases from 1 (i.e., the dominant approach) to 15, changes in the ground heat flux, sensible heat flux, surface temperature, and 2 m air temperature are more significant during nighttime. Changes in the 2 m specific humidity are more significant during daytime, and changes in the boundary layer height are most prominent during the morning and afternoon transitional periods.

A comprehensive assessment of land surface-atmosphere interactions in a WRF/Urban modeling system for Indianapolis, IN

“…Giorgi and Avissar (1997) summarized that sub-grid aggregation, which was accomplished using various methodologies, can affect surface energy fluxes, soil dynamics, and snow pack dynamics performance in models. Li et al (2013) developed a mosaic/tiling approach that integrated sub-grid land surface type data into a WRF simulation for the Washington D.C. -Baltimore corridor. The integration of sub-grid data either improved or had a negligible effect on many meteorological variables (such as surface energy fluxes, surface air temperatures, and rainfall patterns); for example, the sensible heat fluxes and latent heat fluxes were changed by ~2 0 W m -2 and ~1 00 W m -2 respectively when using a mosaic/tiling approach.…”

Multiscale decomposition for heterogeneous land‐atmosphere systems

“…The length scale we defined in this paper is based on the patches imbedded in the processes, and it obviously is about the spatial extent. Scaling heterogeneous land surface with spatial extent of patches is physically interpretable and has been widely accepted [e.g., Avissar and Pielke , ; Lynn et al ., ; Giorgi and Avissar , ; Dai et al ., ; Ament and Simmer , ; Li et al ., ]. While for the representation of atmospheric turbulence, there have been a variety of statistical methods; among them, the two most commonly used are the Fourier transform and the wavelet transform.…”

“…An important approach for representing subgrid land heterogeneity effects is the “mosaic of patches” approach [ Avissar and Pielke , ], which assumes that the land surface consists of mosaic of patches that independently interact with the atmosphere through their surface fluxes. Due to its simplicity and practicality, the mosaic of patches approach has been widely adopted in the atmospheric modeling community [e.g., Avissar , ; Lynn et al ., ; Giorgi and Avissar , ; Dai et al ., ; Ament and Simmer , ; Li et al ., ]. However, this approach does not count heterogeneity scales but describes heterogeneity merely in terms of probability density function (PDF).…”

Effect of Land-Use Change on the Urban Heat Island in the Fukuoka–Kitakyushu Metropolitan Area, Japan

“…To make the NLNI land-use datasets suitable for the WRF LSM, the WRF Preprocessing System (WPS) was modified [16]. The NLNI land-use datasets were published in 1976, 1987, 1991, 1997, 2006 and 2009.…”

“…The surface temperature and the sea surface temperature, for setting the lower boundary conditions, were obtained from the NCEP FNL data. In this study, the following schemes were applied for all domains and all cases: microphysics: Thompson scheme [17]; long-wave radiation: Rapid Radiative Transfer Model (RRTM) scheme [18]; short-wave radiation: Goddard scheme [19]; surface layer: Eta similarity scheme based on Monin-Obukhov similarity [20]; land surface: Noah LSM model [21] with mosaic option [16]; and planetary boundary layer: Mellor-Yamada-Janjic scheme [20]. Grell 3D ensemble cumulus parameterization [22] was applied only for domain 1.…”