Using self-organizing maps (SOMs) to develop a climatology and characterization of the low-level wind field across the Ross Ice Shelf region, Antarctica

Developing a climatology and characterization of the weather in data sparse regions is often a difficult task. The increased presence of numerical weather prediction datasets provides new opportunities for insights into the presence and depiction of weather features. The Antarctic region is one region which is significantly constrained by the lack of observations. The recent increase in operational numerical weather prediction products has provided new resources to develop a climatology of common features across the region. This study uses the method of self-organizing maps (SOMs) to develop an understanding of the low-level wind field across the Ross Ice Shelf region, Antarctica based on archived output from the Antarctic Mesoscale Prediction System (AMPS). SOMs is a method to objectively stratify large volumes of data into a smaller number of recurring patterns on a physically meaning basis. The results of the SOMs analysis identify 20 patterns of the low-level wind field across the Ross Ice Shelf region. An analysis of the seasonality and frequency of occurrence of the 20 patterns provides a better understanding and characterization of the Ross Ice Shelf air stream in Antarctica. The results based on the method of SOMs, although limited by being based on model output, provide new insights into the low-level wind field of Antarctica.