Abstract

NCAR's Data Assimilation Research Testbed (DART) has generated an 80 member ensemble, global, atmospheric reanalysis with the Community Atmosphere Model v6 configuration of NCAR's Community Earth System Model (CESM), the same configuration used for CMIP6 contributions from this state of the art m... Show moreNCAR's Data Assimilation Research Testbed (DART) has generated an 80 member ensemble, global, atmospheric reanalysis with the Community Atmosphere Model v6 configuration of NCAR's Community Earth System Model (CESM), the same configuration used for CMIP6 contributions from this state of the art model. In addition to the atmospheric model reanalysis states, this project created realistic, high frequency, ensemble, atmospheric forcing spanning 2011-2019 for use by all other components of CESM. The forcing is in native CESM format, ready for immediate use in hindcast or data assimilation experiments. Reanalyses are a blend of the information in observations and numerical hindcasts, statistically accounting for the uncertainties in each and providing more information about the atmosphere than either alone. The atmospheric ensemble average is saved 4 times/day, so it includes a wide range of actual atmospheric behavior ranging from the daily cycle to regional fronts and storms to multiyear variability like ENSO. This time series provides insights into the climate. This CAM6 configuration includes active land (CLM5), sea ice (CICE5), and river runoff (MOSART) models. The ensemble of "restart" files of all active models are saved weekly. These provide insights into the hindcast model constrained by several million observations per day. In addition to atmospheric forcing of surface models, these data sets have many uses: Provide real world initial conditions for CESM component model ensemble hindcasts. The ensemble spread in the initial conditions is an accurate measurement of the actual uncertainty. These can be used in DA experiments using CESM's surface models as the hindcast model. Evaluate the Community Land Model's response to actual atmospheric forcing at sub-daily frequency. Identify CAM6 biases through direct comparison to observations. Provide easily accessible opportunities for publications, theses, and undergraduate student projects. Compare this reanalysis ensemble with several CESM free running ensembles. Evaluate the performance of an ensemble Kalman filter DA system. Explore DA algorithm choices. Show less