Alaska TAMDAR and the RTFDDA WRF QC System

The unique and challenging topographical environment along with less than adequate access to data creates a constant challenge for operational forecasting in Alaska, and by extension, storm systems that originate in this region. In 2008, Tropospheric Airborne Meteorological Data Reporting (TAMDAR) sensors were deployed on a fleet of 10 SAAB 340 turboprop aircraft operated by Alaskan-based Peninsula Airways. These sensors collect an array of meteorological information including temperature, humidity, winds aloft, icing and turbulence, in addition to providing location, time and altitude from GPS built into the sensor. This information is then relayed in real-time to a ground-based operations center using a worldwide satellite network. Since 2008, the TAMDAR sensor fleet has doubled and the increase in sensors also allowed for a nearly complete spatial coverage of Alaska. Consequently, the number of available observations has doubled during the last 4 years.

This study further explores the advantages TAMDAR can provide in this region by utilizing an updated RTFDDA WRF Alaska model, significant increase in available TAMDAR data and a longer available timeframe of data. Data-denial studies are employed through the use of the NCAR Advanced Research WRF, the MM5 and the AirDat RTFDDA WRF Alaska. Simulations, withholding and including the additional TAMDAR data, were conducted to investigate the impact of the increased number of observations and the expanded area of coverage of the data field. Using various forecast skill statistics, these results were then validated by comparison against established observation techniques (i.e., RAOBS, ASOS, etc.) for varying domains and model configurations. Furthermore, the differing cases were compared against operational forecasts to determine the accuracy of the forecasts as well as biases and areas for continuing improvement.

This study will re-confirm the continued viability of TAMDAR data as an important supplement to standard ground and radiosonde records in sparsely populated regions. Moreover, the increase in forecast skill for models with the additional TAMDAR data gives encouraging results for future work in improving data poor areas where traditional observing methods are harder to establish and the enhancement of models to produce better forecasts.