A sensitivity study of long-wave and short-wave radiation schemes in the WRF model for air quality applications in western Argentina

CELESTE MULENA; RAFAEL PEDRO FERNÁNDEZ; DAVID G. ALLENDE; S. ENRIQUE PULIAFITO; GABRIELA LAKKIS

San Juan

Congreso; XXVII Reunión Científica de la Asociación Argentina de Geofísicos y Geodestas; 2014

Asociación Argentina de Geofísicos y Geodestas (AAGG)

Solar radiation is the primary natural energy source of the Earth, and plays an irreplaceable role in the energy balance of the Earth-Atmosphere system. The Earth ́s radiative budget describes the net flow of energy in the form of absorption, dispersion and re-emission of the radiation within different atmosphere layers. Radiation also affects surface temperature, the dynamical transport and is the main driver of photochemical reactions. As a key meteorological component, it also determines the regional air quality by influencing formation and destruction of several atmospheric pollutants. The radiative transfer modeling often assumes several approximations in weather simulations. The impact of these formulations on the accuracy of the irradiance predictions and the overall weather prediction is still subject of research. Because uncertainties in meteorological models highly affect chemical transport of pollutant gases and particles, accurate information on spatial and temporal variability of radiation is neededfor reliable air quality prediction systems. In this study we used the Weather Research and Forecasting (WRF) model (Michalakes et al 2004) to examine the sensitivity of the main short-wave and long-wave parameterizations available for use in high resolution nested domains. The experiments were configured to run during January 2013 with three domains using two-way nesting strategy. A large simulation period was selected to include a wide range of atmospheric conditions affecting the short-wave and long-wave radiative transfer calculations (clean-air, synoptic and convective clouds, aerosols, etc.). All WRF sensitivities were driven by identical initial and boundary conditions from the Global Tropospheric Analyses database (NCEP) with 0.5° × 0.5° spatial resolution and temporal resolution of 6 h (UCAR, 2002). A 3:1 grid ratio between inner and outer nests centered on 32.8°S, 68.8°W was used, including an innermost domain size of approximately 84×88 km 2 and a horizontal resolution of 4 km. In order to determine the individual impact of each radiative transfer scheme, only the short-wave or the long-wave parameterizations were modified at a time between the different simulations, remaining unaltered the remaining setup options. The vertical structure of the model is divided into 60 terrain-following hydrostatic pressure levels with the top level being located at 50 hPa.A model evaluation methodology considering (1) surface measurements and (2) satellite data was used. Surface measurement (surface total radiation and surface temperature) were compared to simulation results on an hourly basis. Additionally, model Outgoing Lonwave Radiation (OLR) at the top of the atmosphere was compared with daily Moderate-Resolution Imaging Spectroradiometer (MODIS) data. Results indicate that WRF model is more sensitive to changes in the long-wave radiation scheme considered. The statistics selected to drive the evaluation were the Root Mean Standard Error (RMSE) and the Mean Bias Error (MBE).