scholarly journals Basin-Scale Prediction of Sea Surface Temperature with Artificial Neural Networks

2018 ◽  
Vol 35 (7) ◽  
pp. 1441-1455 ◽  
Author(s):  
Kalpesh Patil ◽  
M. C. Deo
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Indian Ocean ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Sea Surface ◽  
Prediction Skill ◽  
Basin Scale ◽  
Ann Models ◽  
Artificial Neural

AbstractThe prediction of sea surface temperature (SST) on the basis of artificial neural networks (ANNs) can be viewed as complementary to numerical SST predictions, and it has fairly sustained in the recent past. However, one of its limitations is that such ANNs are site specific and do not provide simultaneous spatial information similar to the numerical schemes. In this work we have addressed this issue by presenting basin-scale SST predictions based on the operation of a very large number of individual ANNs simultaneously. The study area belongs to the basin of the tropical Indian Ocean (TIO) having coordinates of 30°N–30°S, 30°–120°E. The network training and testing are done on the basis of HadISST data of the past 140 yr. Monthly SST anomalies are predicted at 3813 nodes in the basin and over nine time steps into the future with more than 20 million ANN models. The network testing indicated that the prediction skill of ANNs is attractive up to certain lead times depending on the subbasin. The ANN models performed well over both the western Indian Ocean (WIO) and eastern Indian Ocean (EIO) regions up to 5 and 4 months lead time, respectively, as judged by the error statistics of the correlation coefficient and the normalized root-mean-square error. The prediction skill of the ANN models for the TIO region is found to be better than the physics-based coupled atmosphere–ocean models. It is also observed that the ANNs are capable of providing an advanced warning of the Indian Ocean dipole as well as abnormal basin warming.

scholarly journals The Pacific–Indian Ocean associated mode in CMIP5 models

Ocean Science ◽  
2020 ◽  
Vol 16 (2) ◽  
pp. 469-482 ◽  
Author(s):  
Minghao Yang ◽  
Xin Li ◽  
Weilai Shi ◽  
Chao Zhang ◽  
Jianqi Zhang
Keyword(s):  
Indian Ocean ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Climate Models ◽  
Southern Oscillation ◽  
Sea Surface ◽  
Cmip5 Models ◽  
Future Climate Change ◽  
The Pacific ◽  
Basin Scale

Abstract. The Pacific–Indian Ocean associated mode (PIOAM), defined as the first dominant mode (empirical orthogonal function, EOF1) of sea surface temperature anomalies (SSTAs) in the Pacific–Indian Ocean between 20∘ S and 20∘ N, is the product of the tropical air–sea interaction at the cross-basin scale and the main mode of ocean variation in the tropics. Evaluating the capability of current climate models to simulate the PIOAM and finding the possible factors that affect the simulation results are beneficial in the pursuit of more accurate future climate change prediction. Based on the 55-year Hadley Centre Global Sea Ice and Sea Surface Temperature (HadISST) dataset and the output data from 21 Coupled Model Intercomparison Project (CMIP) phase 5 (CMIP5) models, the PIOAM in these CMIP5 models is assessed. Instead of using the time coefficient (PC1) of the PIOAM as its index, we chose to utilize the alternative PIOAM index (PIOAMI), defined with SSTA differences in the boxes, to describe the PIOAM. It is found that the explained variance of the PIOAM in almost all 21 CMIP5 models is underestimated. Although all models reproduce the spatial pattern of the positive sea surface temperature anomaly in the eastern equatorial Pacific well, only one-third of these models successfully simulate the El Niño–Southern Oscillation (ENSO) mode with the east–west inverse phase in the Pacific Ocean. In general, CCSM4, GFDL-ESM2M and CMCC-CMS have a stronger capability to capture the PIOAM than the other models. The strengths of the PIOAM in the positive phase in less than one-fifth of the models are slightly greater, and very close to the HadISST dataset, especially CCSM4. The interannual variation of the PIOAM can be measured by CCSM4, GISS-E2-R and FGOALS-s2.

scholarly journals The Pacific-Indian Ocean Associated Mode in CMIP5 Models

2019 ◽  
Author(s):  
Minghao Yang ◽  
Xin Li ◽  
Weilai Shi ◽  
Chao Zhang ◽  
Jianqi Zhang
Keyword(s):  
Indian Ocean ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Climate Models ◽  
Sea Surface ◽  
Cmip5 Models ◽  
Future Climate Change ◽  
Main Mode ◽  
The Pacific ◽  
Basin Scale

Abstract. The Pacific-Indian Ocean associated mode (PIOAM) is the product of the tropical air-sea interaction at the cross-basin scale and the main mode of ocean variation in the tropics. Evaluating the capability of current climate models to simulate the PIOAM and finding the possible factors that affect the simulation results are beneficial to obtain more accurate future climate change prediction. Based on 55-yr the Hadley Centre Global Sea Ice and Sea Surface Temperature (HadISST) reanalysis and the output data from twenty-one Coupled Model Intercomparison Project (CMIP) phase 5 (CMIP5) models, the PIOAM in these CMIP5 models is assessed. It is found that the explained variance of PIOAM in almost all twenty-one CMIP5 models are underestimated. Although all models reproduce the spatial pattern of the positive sea surface temperature anomaly in the eastern equatorial Pacific well, only one-third of these models successfully simulate the ENSO mode with the east-west inverse phase in the Pacific Ocean. In general, CCSM4, GFDL-ESM2M and CMCC-CMS have a stronger capability to capture the PIOAM than that of the other models. The strengths of the PIOAM in the positive phase in less than one-fifth of the models are slightly stronger, and very close to HadISST reanalysis, especially in CCSM4. The interannual variation of PIOAM can be measured by CCSM4, GISS-E2-R and FGOALS-s2. Further analysis indicates that considering the carbon cycle, resolving stratosphere, chemical process or increasing the horizontal resolution of the atmospheric model may effectively improve the performance of the model to simulate the PIOAM.

scholarly journals Interannual coherent variability of SSTA and SSHA in the Tropical Indian Ocean

2012 ◽  
Vol 9 (1) ◽  
pp. 1-24
Author(s):  
J. Q. Feng
Keyword(s):  
Indian Ocean ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Rossby Wave ◽  
Tropical Indian Ocean ◽  
Sea Surface Height ◽  
Ocean Dynamics ◽  
Sea Surface ◽  
Height Anomaly ◽  
Basin Scale

Abstract. Sea surface height derived from the multiple ocean satellite altimeter missions (TOPEX/Poseidon, Jason-1, ERS, Envisat et al.) and sea surface temperature from National Centers for Environmental Prediction (NCEP) over 1993–2008 are analyzed to investigate the coherent patterns between the interannual variability of the sea surface and subsurface in the Tropical Indian Ocean, by jointly adopting Singular Value Decomposition (SVD) and Extended Associate Pattern Analysis (EAPA) methods. Results show that there are two dominant coherent modes with the nearly same main period of about 3–5 yr, accounting for 86 % of the total covariance in all, but 90° phase difference between them. The primary pattern is characterized by a east-west dipole mode associated with the mature phase of ENSO, and the second presents a sandwich mode having one sign anomalies along Sumatra-Java coast and northeast of Madagascar, whilst an opposite sign between the two regions. The robust correlations of the sea surface height anomaly (SSHA) with sea surface temperature anomaly (SSTA) in the leading modes indicate a strong interaction between them, though the highest correlation coefficient appears with a time lag. And there may be some physical significance with respect to ocean dynamics implied in SSHA variability. Analyzing results show that the features of oceanic waves with basin scale, of which the Rossby wave is prominent, are apparent in the dominant modes. It is further demonstrated from the EAPA that the equatorial eastward Kelvin wave and off-equatorial westward Rossby wave as well as their reflection in the east and west boundary, respectively, are important dynamic mechanisms in the evolution of the two leading coherent patterns. Results of the present study suggest that the upper ocean thermal variations on the timescale of interannual coherent with the ocean dynamics in spatial structure and temporal evolution are mainly attributed to the ocean waves.

scholarly journals Interannual Variability of Sea Surface Temperature off Java and Sumatra in a Global GCM*

2008 ◽  
Vol 21 (11) ◽  
pp. 2451-2465 ◽  
Author(s):  
Yan Du ◽  
Tangdong Qu ◽  
Gary Meyers
Keyword(s):  
Indian Ocean ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Mixed Layer ◽  
Heat Budget ◽  
Sea Surface ◽  
Surface Mixed Layer ◽  
Horizontal Advection ◽  
Cold Anomaly ◽  
Sst Anomalies

Abstract Using results from the Simple Ocean Data Assimilation (SODA), this study assesses the mixed layer heat budget to identify the mechanisms that control the interannual variation of sea surface temperature (SST) off Java and Sumatra. The analysis indicates that during the positive Indian Ocean Dipole (IOD) years, cold SST anomalies are phase locked with the season cycle. They may exceed −3°C near the coast of Sumatra and extend as far westward as 80°E along the equator. The depth of the thermocline has a prominent influence on the generation and maintenance of SST anomalies. In the normal years, cooling by upwelling–entrainment is largely counterbalanced by warming due to horizontal advection. In the cooling episode of IOD events, coastal upwelling–entrainment is enhanced, and as a result of mixed layer shoaling, the barrier layer no longer exists, so that the effect of upwelling–entrainment can easily reach the surface mixed layer. Horizontal advection spreads the cold anomaly to the interior tropical Indian Ocean. Near the coast of Java, the northern branch of an anomalous anticyclonic circulation spreads the cold anomaly to the west near the equator. Both the anomalous advection and the enhanced, wind-driven upwelling generate the cold SST anomaly of the positive IOD. At the end of the cooling episode, the enhanced surface thermal forcing overbalances the cooling effect by upwelling/entrainment, and leads to a warming in SST off Java and Sumatra.

Sign in / Sign up

Export Citation Format

Share Document