Results from the study of the development of physical-empirical models for the prediction of spring South Korean temperatures, i.e., establishing asynchronous physically plausible relationships between spring South Korean temperatures, predictands, preceding climate variables, and predictors, and developing statistical tools for the prediction based on the established relationships, are presented in this report. There are different possible predictors for each spring months, with a transition from extratropical influences dominating the predictions for March and April to tropical influences dominating the predictions for May. The developed models based on multiple regression analysis with a flexible construction of physically plausible predictors have proven their efficiency in tests performed on simulated real-time forecasts for 34 springs (1983–2016). The temporal correlation coefficient (TCC) and mean square skill score (MSSS) of the monthly forecasts vary within 0.47–0.56 and 0.17–0.42, respectively. The ranked probability skill score (RPSS) values of the corresponding probabilistic forecasts are within 0.15–0.30. The obtained skill implies that the Asia-Pacific Economic Cooperation Climate

Center multi-model ensemble (APCC MME) forecasts can be complemented with the forecasts obtained with the developed methods.

1. INTRODUCTION

2. DATA AND METHODOLOGY

2.1 Data

2.2 Methodologies

3. RESEARCH RESULTS

3.1 South Korea Spring Temperature. Peculuarities of the Time Series

3.2 Pre-Prediction Analysis

3.2.1 Lag Correlations in South Korean Winter-Spring Temperature

3.2.2 Mechanisms Providing Lag Correlations

3.3 Prediction of the Spring Temperature Anomalies

3.3.1 March

3.3.1.1 Concurrent Relationships of March Temperature with Circulation and Surface Properties

3.3.1.2 Lag Relationships of March Temperature with Preceding January Circulation and Surface Properties

3.3.1.3 Prediction of March Temperature

3.3.2. April

3.3.2.1 Concurrent Relationships of April Temperature with Circulation and Surface Properties

3.3.2.2 Lag Relationships of April Temperature with Preceding January and February Circulation and Surface Properties

3.3.2.3 Prediction of April Temperature

3.3.3. May

3.3.3.1 Concurrent Relationships of May Temperature with Circulation and Surface Properties

3.3.3.2 Lag Relationships of May Temperature with Preceding March Circulation and Surface Properties

3.3.3.3 Prediction of May Temperature

3.3.4 March-May

3.4 Probabilistic Interpretation

4. CONCLUDING REMARKS

REFERENCE

APPENDIX 1

APPENDIX 2

APPENDIX 3