中文摘要：

The year-to-year increment prediction approach proposed by was applied to forecast the annual number of tropical cyclones (TCs) making landfall over China.The year-to-year increase or decrease in the number of land-falling TCs (LTCs) was first predicted to yield a net number of LTCs between successive years.The statistical prediction scheme for the year-to-year increment of annual LTCs was developed based on data collected from 1977 to 2007,which includes five predictors associated with high latitude circulations in both Hemispheres and the circulation over the local,tropical western North Pacific Ocean.The model shows reasonably high predictive ability,with an average root mean square error (RMSE) of 1.09,a mean absolute error (MAE) of 0.9,and a correlation coefficient between the predicted and observed annual number of LTCs of 0.86,accounting for 74% of the total variance.The cross-validation test further demonstrated the high predictive ability of the model,with an RMSE value of 1.4,an MAE value of 1.2,and a correlation coefficient of 0.74 during this period.

英文摘要：

The year-to-year increment prediction approach proposed by was applied to forecast the annual number of tropical cyclones (TCs) making landfall over China. The year-to-year increase or decrease in the number of land-falling TCs (LTCs) was first predicted to yield a net number of LTCs between successive years. The statistical prediction scheme for the year-to-year increment of annual LTCs was developed based on data collected from 1977 to 2007, which includes five predictors associated with high latitude circulations in both Hemispheres and the circulation over the local, tropical western North Pacific Ocean. The model shows reasonably high predictive ability, with an average root mean square error (RMSE) of 1.09, a mean absolute error (MAE) of 0.9, and a correlation coefficient between the predicted and observed annual number of LTCs of 0.86, accounting for 74% of the total variance. The cross-validation test further demonstrated the high predictive ability of the model, with an RMSE value of 1.4, an MAE value of 1.2, and a correlation coefficient of 0.74 during this period.