Improving Australian Rainfall Prediction Using Sea Surface Salinity

被引:5
作者
Rathore, Saurabh [1 ,2 ]
Bindoff, Nathaniel L. [1 ,3 ,4 ,5 ]
Ummenhofer, Caroline C. [3 ,6 ]
Phillips, Helen E. [1 ,3 ]
Feng, Ming [7 ,8 ]
Mishra, Mayank [9 ]
机构
[1] Univ Tasmania, Inst Marine & Antarctic Studies, Hobart, Tas, Australia
[2] ARC Ctr Excellence Climate Syst Sci, Hobart, Tas, Australia
[3] ARC Ctr Excellence Climate Extremes, Sydney, NSW, Australia
[4] CSIRO Oceans & Atmosphere, Hobart, Tas, Australia
[5] Australian Antarctic Program Partnership, Hobart, Tas, Australia
[6] Woods Hole Oceanog Inst, Woods Hole, MA 02543 USA
[7] CSIRO Oceans & Atmosphere, Indian Ocean Marine Res Ctr, Crawley, WA, Australia
[8] CSIRO, Ctr Southern Hemisphere Oceans Res, Hobart, Tas, Australia
[9] Indian Inst Technol Kharagpur, Ctr Oceans Rivers Atmosphere & Land Sci, Kharagpur, W Bengal, India
基金
美国国家科学基金会;
关键词
ENSO; Flood events; Hydrologic cycle; Machine learning; Rainfall; Salinity; Seasonal forecasting; Soil moisture; CLIMATE VARIABILITY; SUMMER RAINFALL; WINTER RAINFALL; ENSO MODOKI; TEMPERATURE; PREDICTABILITY; PRECIPITATION; REANALYSIS; ALGORITHMS; REGRESSION;
D O I
10.1175/JCLI-D-20-0625.1
中图分类号
P4 [大气科学(气象学)];
学科分类号
0706 ; 070601 ;
摘要
This study uses sea surface salinity (SSS) as an additional precursor for improving the prediction of summer [December-February (DJF)] rainfall over northeastern Australia. From a singular value decomposition between SSS of prior seasons and DJF rainfall, we note that SSS of the Indo-Pacific warm pool region [SSSP (150 degrees E-165 degrees W and 10 degrees S-10 degrees N) and SSSI (50 degrees-95 degrees E and 10 degrees S-10 degrees N)] covaries with Australian rainfall, particularly in the northeast region. Composite analysis that is based on high or low SSS events in the SSSP and SSSI regions is performed to understand the physical links between the SSS and the atmospheric moisture originating from the regions of anomalously high or low, respectively, SSS and precipitation over Australia. The composites show the signature of co-occurring La Nina and negative Indian Ocean dipole with anomalously wet conditions over Australia and conversely show the signature of co-occurring El Nino and positive Indian Ocean dipole with anomalously dry conditions there. During the high SSS events of the SSSP and SSSI regions, the convergence of incoming moisture flux results in anomalously wet conditions over Australia with a positive soil moisture anomaly. Conversely, during the low SSS events of the SSSP and SSSI regions, the divergence of incoming moisture flux results in anomalously dry conditions over Australia with a negative soil moisture anomaly. We show from the random-forest regression analysis that the local soil moisture, El Nino-Southern Oscillation (ENSO), and SSSP are the most important precursors for the northeast Australian rainfall whereas for the Brisbane region ENSO, SSSP, and the Indian Ocean dipole are the most important. The prediction of Australian rainfall using random-forest regression shows an improvement by including SSS from the prior season. This evidence suggests that sustained observations of SSS can improve the monitoring of the Australian regional hydrological cycle.
引用
收藏
页码:2473 / 2490
页数:18
相关论文
共 87 条
  • [1] Permutation importance: a corrected feature importance measure
    Altmann, Andre
    Tolosi, Laura
    Sander, Oliver
    Lengauer, Thomas
    [J]. BIOINFORMATICS, 2010, 26 (10) : 1340 - 1347
  • [2] Influence of the Indian Ocean Dipole on the Australian winter rainfall
    Ashok, K
    Guan, ZY
    Yamagata, T
    [J]. GEOPHYSICAL RESEARCH LETTERS, 2003, 30 (15) : CLM6 - 1
  • [3] On the potential impact of sea surface salinity observations on ENSO predictions
    Ballabrera-Poy, J
    Murtugudde, R
    Busalacchi, AJ
    [J]. JOURNAL OF GEOPHYSICAL RESEARCH-OCEANS, 2002, 107 (C12)
  • [4] Evaluation of the ECMWF ocean reanalysis system ORAS4
    Balmaseda, Magdalena Alonso
    Mogensen, Kristian
    Weaver, Anthony T.
    [J]. QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY, 2013, 139 (674) : 1132 - 1161
  • [5] Random forests
    Breiman, L
    [J]. MACHINE LEARNING, 2001, 45 (01) : 5 - 32
  • [6] Climate change contributes to more frequent consecutive positive Indian Ocean Dipole events
    Cai, W.
    Sullivan, A.
    Cowan, T.
    [J]. GEOPHYSICAL RESEARCH LETTERS, 2009, 36
  • [7] An Asymmetry in the IOD and ENSO Teleconnection Pathway and Its Impact on Australian Climate
    Cai, Wenju
    van Rensch, Peter
    Cowan, Tim
    Hendon, Harry H.
    [J]. JOURNAL OF CLIMATE, 2012, 25 (18) : 6318 - 6329
  • [8] Teleconnection Pathways of ENSO and the IOD and the Mechanisms for Impacts on Australian Rainfall
    Cai, Wenju
    van Rensch, Peter
    Cowan, Tim
    Hendon, Harry H.
    [J]. JOURNAL OF CLIMATE, 2011, 24 (15) : 3910 - 3923
  • [9] Ocean Salinity as a Precursor of Summer Rainfall over the East Asian Monsoon Region
    Chen, Biao
    Qin, Huiling
    Chen, Guixing
    Xue, Huijie
    [J]. JOURNAL OF CLIMATE, 2019, 32 (17) : 5659 - 5676
  • [10] Statistical Uncertainty Estimation Using Random Forests and Its Application to Drought Forecast
    Chen, Junfei
    Li, Ming
    Wang, Weiguang
    [J]. MATHEMATICAL PROBLEMS IN ENGINEERING, 2012, 2012