Design of Buoy Observation Network over the Arctic Ocean

Abstract:

Due to variations of sea ice extents and no land condition, few in-situ observations are available over the Arctic Ocean, which could increase the uncertainty of initial conditions in the model and degrade the forecast accuracy. Under this circumstances, the number of buoy observations were rapidly reduced from 2016 to 2019, which may result in even worse effects on weather forecasts over the Arctic Ocean. To improve the weather forecasts over the Arctic Ocean, the new buoys are need to be deployed over the regions where the newly added buoy data may contribute to the improvement of weather forecasts. Since operating consistent in-situ observation (i.e., buoy) sites is difficult and needs lots of cost, it is important to decide where to deploy the new buoy in advance.

In this study, the optimal buoy observation network was designed to minimize forecast errors of 2 m temperature by using the ensemble sensitivity method. The effect of designed buoy observation network on the reduction of forecast error for 2 m temperature was also evaluated. The analyses and forecast errors of 2 m temperature for summer months (July, August, September, and October) during 2016-2019 from the Global Forecast System (GFS) data were used for ensemble sensitivity calculations.

First, by using the ensemble sensitivity method, total 10 buoy positions that are expected to reduce the forecast error variance were selected in order. In general, the regions with large forecast error variance were selected. The 10 buoy positions were concentrated in 6 areas with duplicated positions. Second, the effect of designed buoy observation network for the forecasts of Arctic environment was evaluated. By assuming the assimilation of buoy observations from firstly selected 6 buoy sites, the variances of 2 m temperature forecast errors were reduced by 3% over the Arctic Ocean.

Acknowledgments

This work was supported by the Korea Polar Research Institute (KOPRI, PN20081) and a National Research Foundation of Korea (NRF) grant funded by the South Korean government (Ministry of Science and ICT) (Grant 2017R1E1A1A03070968).