Climate prediction of dust weather frequency over northern China based on sea-ice cover and vegetation variability

Seasonal climate predictions of spring (March‒April‒May) dust weather frequency (DWF) over North China (DWFNC) are conducted based on a previous-summer (June–July–August) normalized difference vegetation index in North China (NDVINC), winter (December–January–February) sea-ice cover index over the Barents Sea (SICBS), and winter Antarctic Oscillation index (AAOI). The year-to-year increment approach is applied to improve the prediction skill. Two statistical prediction schemes—statistical models based on year-to-year-increment-form predictors (SM-DY) and anomaly-form predictors (SM-A)—are applied based on NDVINC, SICBS, and AAOI. The results show that the prediction model using the year-to-year increment approach performs much better in predicting DWFNC, with the correlation coefficient between the average DWFNC and the cross-validated results of SM-DY (SM-A) being 0.80 (0.68) during 1983–2016. A hybrid dynamical–statistical prediction model (HM-DY) is constructed based on NDVINC, SICBS, and a spring 850-hPa geopotential height index, derived from the second version of the NCEP Climate Forecast System. Results show that HM-DY has comparable prediction skill with SM-DY. Both SM-DY and HM-DY are extended to hindcast DWF over the 245 stations in the whole of northern China, indicating comparably high skill. The results show that NDVINC and SICBS account for large variances of the dust climate over northern China. In particular, NDVINC and SICBS can enhance 64% of stations in North China in their prediction of dust climate.