Impacts of Background Albedo Dataset on WRF Model Simulations

For many years, the default monthly background albedo used by the Weather Research and Forecasting (WRF) model and WRF Preprocessing System (WPS) has been based on relatively old and low-resolution data derived from Advanced Very High Resolution Radiometer (AVHRR) observations. In addition to being relatively coarse, the original AVHRR background albedo dataset contains some suspicious artifacts.

A new background albedo dataset based on NASA Moderate Resolution Imaging Spectroradiometer (MODIS) observations is scheduled for release with WRF/WPS v4.0 in mid 2018. This MODIS background albedo dataset is at a much finer resolution than the AVHRR dataset, is free of obvious artifacts, and has albedo values of 0.1–0.3 higher in arid regions with dry, sandy soils, such as in the western U.S. and Mexico. In very isolated instances, like White Sands, New Mexico, the new data represents a doubling of albedo value over the AVHRR dataset.

Here we examine the impacts that the original (AVHRR) and new (MODIS) background albedo dataset have on WRF simulations. We use the NOAA High-Resolution Rapid Refresh (HRRR) model domain and drive WRF with the real-time HRRR initial and boundary condition input files obtained from NOAA/Global Systems Division. Sensitivity experiments using once-daily WRF simulations were performed for one week in Dec 2016 and one week in Jul 2017. In winter, when we expect differences in background albedo to have the least impact, differences in lowest model level temperature in many locations in the southwestern U.S. reach 1–3 K within a few hours after sunrise, with the MODIS experiment being generally cooler due to higher albedo. In summer, as expected, even more substantial differences between the two experiments in both lowest model layer temperature and precipitation are observed, which become even more pronounced by day 2 of the simulations.