A new global meso-beta-scale (40 km) downscaled climatography: Improving decision making for societal and ecological applications

NCAR’s Research Applications Laboratory has developed a new global climatography for emergency management applications and renewable energy applications. The dataset consists of 21 years of hourly, global dynamically-downscaled reanalyses (40 km grid increment) from 1985 to 2005. The downscaling was performed with NCAR’s Climate Four Dimensional Data Assimilation System (ClimoFDDA), which was originally based upon the Penn State/NCAR MM5 model, and is now based on NCAR’s WRF Model. For the 21-year climatography presented herein, ClimoFDDA used global-scale data from the NCEP-DOE Reanalysis, standard surface and upper observations, and satellite-derived estimates of winds, temperature, and humidity to downscale the global climate information to meso-beta-scale, while fully accounting for topographic variations and surface characteristics (e.g., crop lands, lakes, dense urban areas, etc.). The presenter will describe the motivation, technical approach, and characteristics of this unique dataset. He will then show several examples of this dataset's utility for greatly enhancing decision making for renewable energy applications and emergency management applications, including: (a) assessing the interannual variability of wind power generation potential, (b) evaluating weather extremes (worst and best case scenarios), (c) relative frequency (return period) of winds at “operationally significant” thresholds, (d) wind climatographies at remote locations (distant from observation stations), (e) “regime-appropriate” uncertainty estimates that account for the variability of weather patterns, and (f) the diurnal constancy of winds at any point on the globe. This re-analysis dataset provides a new opportunity to address basic questions for an array of socioeconomic and ecological applications that were previously out of reach due to the coarse-scale representation (both spatial and temporal) of previous datasets.