The Stability Analysis Of EOF Modes Of Summer Rainfall In China And Statistical Downscaling Forecast

Summer rainfall forecast is a great challenge in the seasonal climate prediction in China.The current forecast skill of summer rainfall in China is pooer in contrast to the regionalforecast skill score. Therefore, better understanding of the spatial-temporal variation ofrainfall anomalies and its associated monsoon circulation will enhance the mechanism ofrainfall anomalies and improve the seasonal forecast skill scores. In this paper, based on the160station-observed precipitations, NCEP/NCAR reanalyzed datasets and CFSv2.0modeldatasets, the characteristics and temporal stability of the first four EOF modes of Chinasummer rainfall, including the effect of circulations and circulation-interaction to each EOFmode, have been analyzed. The key circulation indices related to each principal component(PC) of EOF mode were identified. In addition, the author analyzed the potential precipitationforecast skill of CFSv2.0model. The main results are summarized as follows:(1) It is found that when one year sample was randomly removed, the leading four EOFmodes suggest a higher time stability. The EOF-based potential forecast skill is mainly inSouth China, the33-year-averaged anomaly correlation coefficient (ACC) between theidealized forecast and the observed rainfall anomalies is about0.6. In contrast, the stability ofthe leading EOF modes for the precipitation anomaly percentage is greatly affected by theextreme climate events, but when we artificially removed this effect, the stability of the firstthree EOF modes had been improved, and its potential forecast skills appear in most parts ofChina, the averaged ACC between the potential forecast and observed rainfall anomalypercentage is0.48. The stability of EOF mode exhibits an obvious decrease when the forecastperiod is extended, especially for the last three EOF modes, which suggests that forecastingskill scores would decrease when we applied the EOF method to predicate the future rainfallanomalies in two years advance.(2) The correlation between circulation fields and the first four EOF modes of rainfallanomalies are found notable and stable, and the identified circulation indices, includingpre-winter and summer counterparts have shown the better potential predictors for PCs. Thetemporal correlation coefficients (TCC) between the observed and forecasted rainfallanomalies in most stations are greater than zero, and the averaged ACC is about0.29. Incontrast, in the TCC of downscaled by those summer indices have more skill scores, and the avergaed ACC is about0.34.(3) China summer rainfall changes with three dimensional East Asian summer monsooncirculations, and they show a very stable relationship. It is found that interannual variability ofrainfall is mainly affected by the wind anomalies at200hPa and850hPa indicating thecombined effect of interaction between the lower and higher layer monsoonal circulation.Where, the the western North Pacific Subtropical High (WNPSH), South Asian High (SAH)and the Mongolia Cyclone (MC), change synchronously and constitute a ‘gear coupling’wheel, which possibly resulted in the diversity of summer rainfall anomalies in China.(4) CFSv2.0has shown a rough predictability of summer rainfall anomalies in China, butwith a pooer skill scores. The model-forecasted leading modes of rainfall have baises, andrelationship between the model circulation and rainfall anomalies is very different from thatof observation. It suggests that the statistical downscaling method of rainfall is not successfulfor this model, where the TCC of rainfall is lower, and the averaged ACC is less than0.1.