The Prediction Mechanisms Of Varying Annual Cycle

Skillful predictions of climate variability months in advance bear fundamental importance in social and agricultural planning and disaster management.Due to the unique location of eastern China and its orographic complexity,the temporal and spatial changes of precipitation have obvious interactions at different time scales due to the influence of multiple climate systems.However,the progress of developing more skillful predictions at these timescales has long been hindered by the lack of an objective approach to harness,to the maximum extent,information from sources of predictability.To solve this problem,this paper devises a novel scheme that pushes the reference boundary of seasonal prediction by predicting the spatiotemporal evolution of eastern China precipitation seamlessly over an entire year.The new scheme is founded on the recognition and isolation of “varying annual cycles”,representing distinct spatiotemporal footprints left by modes of climate variability that interfere with the precipitation response to the solar forcing.On this basis,this paper systematically explores the characteristics and its modulation mechanisms of varying annual cycle of precipitation in eastern China.Meanwhile,the correlation relationship of possible external factors affecting precipitation variability in eastern China is comprehensively analyzed,and the main physical factors that play an important role in interannual variability of precipitation annual cycle in eastern China are determined by successive orthogonalization analysis method.In these main factors,this paper specifically explores the reasons for the continuous warming and snow retreat over Tibetan Plateau in recent decades.Finally,this paper systematically evaluates the utility and advantages of the new scheme in predicting the spatiotemporal variation of precipitation annual cycle in eastern China.This novel scheme in principle can be adopted for climate prediction across interannual to decadal timescales.Moreover,it is generically applicable for improving seasonal forecasts over other monsoon regions and for predictions of other critical climate indicators such as surface temperature and Arctic sea-ice extent.The major conclusions are summarized as following:(1)By virtue of being located in the world's strongest monsoon region,spatiotemporal evolution of precipitation in eastern China is strongly influenced by seasonal evolving solar forcing.On the premise that climate variability related to annually evolving solar forcing is more predictable than the internal climate variability,this paper devises an innovative way to isolate the precipitation variability in eastern China associated with annual cycle patterns from the traditionally defined precipitation anomaly field.This scheme separates the definition of time t of variable explicitly into(Julian)days of a year and years and then combine days of a year and spatial coordinates as the domain in which a meteorological variable(such as precipitation)evolves yearly,referred to as the space-day domain.As a result,yearly precipitation fields in the space-day domain naturally contain the information of the repeatable “space-day” evolution that is directly related to solar forcing and non-repeatable “space-day” variation responses indirectly to solar forcing.The results suggest that the interannual variability of annual cycle accounts for nearly 62% of the total variance of precipitation anomalies in eastern China and representing its year-to-year variations of starting date,duration,and intensity of the wet season.Therefore,the year-to-year space-day variation of monsoon precipitation in eastern China is mainly dominated by the interannual variability of annual cycle.(2)Varying annual cycle encapsulate distinct spatiotemporal footprints left by a set of external factors that regulate Asian monsoon systems and collectively describe atmosphere-ocean-land conditions for eastern China precipitation response to the solar forcing.By applying a successive orthogonalization analysis method,we obtain 14 independent predictors from 30 possible external factors that have been used in the literature for predicting spring/summer precipitation anomalies in some specific regions over eastern China.Collectively,the variation of these 14 factors explains about 51.1% variance of the year-to-year variation of precipitation in eastern China associated with varying annual cycle.The first 5 independent factors that have the largest contribution to varying annual cycle are(i)Multivariate El Nino-Southern Oscillation index(ENSO),(ii)Tibetan Plateau snow depth(TPSD),(iii)Quasi Biennial Oscillation index(QBO),(iv)spatial mean SST over North Pacific Ocean(NPO),and(iv)Antarctic Oscillation index(AAO).These factors regulate the yearly response of precipitation in eastern China to the seasonal evolving solar forcing via advancing/delaying starting dates of monsoons,lengthening/shortening of the monsoon rainfall season,and/or strengthening/weakening of precipitation intensity in different latitude bands of eastern China.(3)Snow cover in the Tibetan Plateau is one of the main factors affecting the precipitation annual cycle change in eastern China.The results based on the multiple data and climate feedback analysis method suggest that the Tibetan Plateau shows continuous warming and snow retreat during global warming hiatus period,especially in winter.For the regional average,the radiatively-forced and dynamically-induced temperature changes contribute 60.9% and 39.1% to surface air temperature variation,indicating that radiatively-forced roles dominate the warming over the Tibetan Plateau.Further studies reveal that the snow/ice albedo feedback results in a warming of 2.6?and is the primary contributor to enhanced warming over the Himalayas,which is much greater than the warming induced by other dynamic and radiative factors.Furthermore,the results indicate that snow cover in the Tibetan Plateau affects the temporal and spatial evolution of precipitation by modulating the annual cycle modes2 and 3 in eastern China.In recent decades,the decreased snow cover in the Tibetan Plateau has weakened such modulated relationship.(4)By independently forecasting the dominant spatiotemporal modes and amplitude changes of rainfall associated with varying annual cycle,the new scheme reduces spatiotemporal noise contamination and enhances the utilization of the potential predictability offered by each physical factor.With the predictors describing previous atmosphere-ocean-land conditions,the new scheme is able to predict multi-provincial and successive flood and drought occurrences in the following year with a good prediction skill.The correlation skill in predicting the original year-to-year time series of these amplitudes of annual cycle patterns,ranging from0.61 to 0.76,is very close to the skill assessed in the diagnostic period.The averaged correlation skill of the predicted interannual variability of annual cycle is about 0.58.This paper also indicates that the predicted interannual variability of annual cycle is more skillful in years when the amplitude changes of annual cycle is stronger(e.g.,predicted correlation skill exceeds 0.85 in 2015 and 2018).