Dynamical and Machine Learning Hybrid Seasonal Prediction of Summer Rainfall in China

2021 ◽  
Vol 35 (4) ◽  
pp. 583-593
Author(s):  
Jialin Wang ◽  
Jing Yang ◽  
Hong-Li Ren ◽  
Jinxiao Li ◽  
Qing Bao ◽  
...  
Keyword(s):  
Machine Learning ◽  
Summer Rainfall ◽  
Seasonal Prediction

Machine learning-based dynamical seasonal prediction of summer rainfall in China

2020 ◽  
Author(s):  
Jialin Wang ◽  
Jing Yang ◽  
Hongli Ren ◽  
Jinxiao Li ◽  
Qing Bao ◽  
...  
Keyword(s):  
Machine Learning ◽  
Prediction Method ◽  
Summer Rainfall ◽  
Seasonal Prediction ◽  
Prediction Skill ◽  
Gradient Boosting ◽  
Disaster Reduction ◽  
Best Fitting ◽  
Application Prospects ◽  
Dynamical Seasonal Prediction

<p>The seasonal prediction of summer rainfall is crucial for regional disaster reduction but currently has a low prediction skill. This study developed a machine learning (ML)-based dynamical (MLD) seasonal prediction method for summer rainfall in China based on suitable circulation fields from an operational dynamical prediction model CAS FGOALS-f2. Through choosing optimum hyperparameters for three ML methods to reach the best fitting and the least overfitting, gradient boosting regression trees eventually exhibit the highest prediction skill, obtaining averaged values of 0.33 in the reference training period (1981-2010) and 0.19 in eight individual years (2011-2018) of independent prediction, which significantly improves the previous dynamical prediction skill by more than 300%. Further study suggests that both reducing overfitting and using the best dynamical prediction are imperative in MLD application prospects, which warrants further investigation.</p>

scholarly journals Skillful seasonal prediction of Yangtze river valley summer rainfall

2016 ◽  
Vol 11 (9) ◽  
pp. 094002 ◽  
Author(s):  
Chaofan Li ◽  
Adam A Scaife ◽  
Riyu Lu ◽  
Alberto Arribas ◽  
Anca Brookshaw ◽  
...  
Keyword(s):  
Yangtze River ◽  
Summer Rainfall ◽  
Seasonal Prediction ◽  
Yangtze River Valley ◽  
River Valley

scholarly journals Long-Lead Seasonal Prediction of China Summer Rainfall Using an EOF–PLS Regression-Based Methodology*,+

2016 ◽  
Vol 29 (5) ◽  
pp. 1783-1796 ◽  
Author(s):  
Wen Xing ◽  
Bin Wang ◽  
So-Young Yim
Keyword(s):  
Air Temperature ◽  
Dynamic Models ◽  
Principal Component ◽  
Summer Rainfall ◽  
Seasonal Prediction ◽  
Rainfall Anomaly ◽  
Pls Regression ◽  
Rainfall Prediction ◽  
Forecast Validation ◽  
Forecast Models

Abstract Considerable year-to-year variability of summer rainfall exposes China to threats of frequent droughts and floods. Objective prediction of the summer rainfall anomaly pattern turns out to be very challenging. As shown in the present study, the contemporary state-of-the-art dynamical models’ 1-month-lead prediction of China summer rainfall (CSR) anomalies has insignificant skills. Thus, there is an urgent need to explore other ways to improve CSR prediction. The present study proposes a combined empirical orthogonal function (EOF)–partial least squares (PLS) regression method to offer a potential long-lead objective prediction of spatial distribution of CSR anomalies. The essence of the methodology is to use PLS regression to predict the principal component (PC) of the first five leading EOF modes of CSR. The preceding December–January mean surface temperature field [ST; i.e., SST over ocean and 2-m air temperature (T2m) over land] is selected as the predictor field for all five PCs because SST and snow cover, which is reflected by 2-m air temperature, are the most important factors that affect CSR and because the correlation between each mode and ST during winter is higher than in spring. The 4-month-lead forecast models are established by using the data from 1979 to 2004. A 9-yr independent forward-rolling prediction is made for the latest 9 yr (2005–13) as a strict forecast validation. The pattern correlation coefficient skill (0.32) between the observed and the 4-month-lead predicted patterns during the independent forecast period of 2005–13 is significantly higher than the dynamic models’ 1-month-lead hindcast skill (0.04), which indicates that the EOF–PLS regression is a useful tool for improving the current seasonal rainfall prediction. Issues related to the EOF–PLS method are also discussed.

Distinct Principal Modes of Early and Late Summer Rainfall Anomalies in East Asia*

2009 ◽  
Vol 22 (13) ◽  
pp. 3864-3875 ◽  
Author(s):  
Bin Wang ◽  
Jian Liu ◽  
Jing Yang ◽  
Tianjun Zhou ◽  
Zhiwei Wu
Keyword(s):  
East Asia ◽  
Summer Monsoon ◽  
Land Surface ◽  
Southern Oscillation ◽  
Southern China ◽  
Late Summer ◽  
Summer Rainfall ◽  
Seasonal Prediction ◽  
Early Summer ◽  
Rainfall Anomalies

Abstract The current seasonal prediction of East Asia (EA) summer monsoon deals with June–July–August (JJA) mean anomalies. This study shows that the EA summer monsoon may be divided into early summer [May–June (MJ)] and late summer [July–August (JA)] and exhibits remarkable differences in mean state between MJ and JA. This study reveals that the principal modes of interannual precipitation variability have distinct spatial and temporal structures during the early and late summer. These principal modes can be categorized as either El Niño–Southern Oscillation (ENSO) related or non-ENSO related. During the period of 1979–2007, ENSO-related modes explain 35% of MJ variance and 45% of JA variance, and non-ENSO-related modes account for 25% of MJ variance and 20% of JA variance. For ENSO-related variance, about two-thirds are associated with ENSO decaying phases, and one-third is associated with ENSO developing phases. The ENSO-related MJ modes generally concur with rapid decay or early development of ENSO episodes, and the opposite tends to apply to ENSO-related JA modes. The non-ENSO MJ mode is preceded by anomalous land surface temperatures over southern China during the previous March and April. The non-ENSO JA mode is preceded by lasting equatorial western Pacific (the Niño-4 region) warming from the previous winter through late summer. The results suggest that 1) prediction of bimonthly (MJ) and (JA) anomalies may be useful, 2) accurate prediction of the detailed evolution of ENSO is critical for prediction of ENSO-related bimonthly rainfall anomalies over East Asia, and 3) non-ENSO-related modes are of paramount importance during ENSO neutral years. Further establishment of the physical linkages between the non-ENSO modes and their corresponding precursors may provide additional sources for EA summer monsoon prediction.

scholarly journals The Predictability of Rainfall over the Greater Horn of Africa. Part II: Prediction of Monthly Rainfall during the Long Rains

2015 ◽  
Vol 16 (5) ◽  
pp. 2001-2012 ◽  
Author(s):  
Sharon E. Nicholson
Keyword(s):  
Summer Rainfall ◽  
Seasonal Prediction ◽  
Meridional Wind ◽  
Lead Times ◽  
Horn Of Africa ◽  
Sea Level Pressure ◽  
Boreal Spring ◽  
Predictive Skill ◽  
Wind Sea ◽  
Greater Horn Of Africa

Abstract Seasonal prediction of the boreal spring rains in the Greater Horn of Africa has been notoriously challenging. Predictability is markedly lower than during the autumnal rainy season. Part I of this article explored predictability at the seasonal scale, using multiple linear regression. However, the three months of the boreal spring season are clearly different climatologically and with respect to the prevailing atmospheric circulation and controls on interannual variability. For that reason, the current study follows up on Part I by examining the predictability of the three months individually. The current study utilizes 1- and 2-month lead times and the results are evaluated via cross validation. This approach provided improved skill for April and May in the equatorial rainfall region, but not for March in this region and not for the region with predominantly summer rainfall. Overall, the best predictors are shown to be atmospheric variables, most often zonal and meridional wind. Sea surface temperatures and sea level pressure provided little predictive skill.

scholarly journals Forecasting northern Australian summer rainfall bursts using a seasonal prediction system

2021 ◽  
Author(s):  
Tim Cowan ◽  
Matthew C. Wheeler ◽  
S. Sharmila ◽  
Sugata Narsey ◽  
Catherine de Burgh-Day
Keyword(s):  
Summer Rainfall ◽  
Seasonal Prediction ◽  
Burst Activity ◽  
Northern Coast ◽  
Business Decisions ◽  
Prototype Development ◽  
Predictive Skill ◽  
Northern Regions ◽  
Australian Community ◽  
User Friendly

AbstractRainfall bursts are relatively short-lived events that typically occur over consecutive days, up to a week. Northern Australian industries like sugar farming and beef are highly sensitive to burst activity, yet little is known about the multi-week prediction of bursts. This study evaluates summer (December to March) bursts over northern Australia in observations and multi-week hindcasts from the Bureau of Meteorology’s multi-week to seasonal system, ACCESS-S1 (Australian Community Climate and Earth-System Simulator, Seasonal version 1). The main objective is to test ACCESS-S1’s skill to confidently predict tropical burst activity, defined as rainfall accumulation exceeding a threshold amount over three days, for the purpose of producing a practical, user-friendly burst forecast product. The ensemble hindcasts, made up of 11 members for the period 1990–2012, display good predictive skill out to lead week 2 in the far northern regions, despite overestimating the total number of summer burst days and the proportion of total summer rainfall from bursts. Coinciding with a predicted strong Madden-Julian Oscillation (MJO), the skill in burst event prediction can be extended out to four weeks over the far northern coast in December, however this improvement is not apparent in other months or over the far northeast, which shows generally better forecast skill with a predicted weak MJO. The ability of ACCESS-S1 to skillfully forecast bursts out to 2-3 weeks suggests the Bureau's recent prototype development of a Burst Potential forecast product would be of great interest to northern Australia’s livestock and crop producers, who rely on accurate multi-week rainfall forecasts for managing business decisions.

Sign in / Sign up

Export Citation Format

Share Document