Effect of the Cycling WRF-3DVAR Data Assimilation of the Ship-borne Arctic Radiosonde Sounding on the Simulation of the Intense Arctic Cyclone in mid-August 2016

Abstract

The Arctic Ocean is more accessible to human activities as sea-ice declines. As a result, Arctic severe weather prediction is becoming practically more important. Previous studies have demonstrated that the cyclone activity tends to enhance seasonally during summer in the central Arctic Ocean and, during the last several decades, the entrance of extratropical cyclones has increased toward the Arctic [1-3]. One representative case was a violent, long-lived Arctic cyclone in early August of 2012, with record lifetime minimum central pressure of 966 hPa [4]. In mid-August 2016, another atypically intense Arctic cyclone was observed, which was comparable to the 2012 cyclone. Originating from Scandinavia, it explosively developed over the Eurasian sector of the central Arctic Ocean, reaching its lifetime minimum central pressure of 968 hPa on 16 August. Thereafter it was weakening for three days while moving eastward and finally dissipated by merging into another developing cyclone from Alaska. In this study, we perform WRF forecasts of this 2016 cyclone’s life cycle with the initial (00Z 10 August) and boundary conditions from the NCEP GFS forecast fields. In addition, extra Arctic radiosonde sounding data over the Chukchi and East Siberian seas are available, since we produced on the icebreaker Araon. For an experimental purpose, we test the assimilation impact of those extra sounding data using the WRF-3DVAR. Three model experiments have been set up: 1) a control run without extra data assimilation (DA_NO), 2) an experiment with one-time assimilation on 00Z 10 August (DA_10), and 3) an experiment with 24-hr cycling assimilation from 00Z 10 through 00Z 13 August. Comparing the central pressure evolution among the three experiments, DA_NO and DA_10 simulated the cyclone’s life cycle in a very similar way, whereas DA_13 shows a notable discrepancy on the simulation of the weakening phase from 16 August. In DA_13, the weakening rate is even more rapid, compared with the reanalysis data: the central pressure increased to the weakest peak of 979 hPa within 36-hr (c.f., from 968 hPa to 984 hPa within 48-hr in the reanalysis data). The reason of this difference is that an anticylonic increment generated by the cycling sounding data assimilation grows and spreads to the central Arctic Ocean and excessively affects the cyclone weakening while it passes around the International Dateline, located at the closest distance from the Araon. Therefore, the extra Araon Arctic sounding data could affect the cyclone’s central pressure forecast around the time of its lifetime peak period.