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

Author(s):  
Kalpesh Patil ◽  
M. C. Deo
2018 ◽  
Vol 35 (7) ◽  
pp. 1441-1455 ◽  
Author(s):  
Kalpesh Patil ◽  
M. C. Deo

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.


2014 ◽  
Vol 142 (5) ◽  
pp. 1771-1791 ◽  
Author(s):  
Mohamed Helmy Elsanabary ◽  
Thian Yew Gan

Abstract Rainfall is the primary driver of basin hydrologic processes. This article examines a recently developed rainfall predictive tool that combines wavelet principal component analysis (WPCA), an artificial neural networks-genetic algorithm (ANN-GA), and statistical disaggregation into an integrated framework useful for the management of water resources around the upper Blue Nile River basin (UBNB) in Ethiopia. From the correlation field between scale-average wavelet powers (SAWPs) of the February–May (FMAM) global sea surface temperature (SST) and the first wavelet principal component (WPC1) of June–September (JJAS) seasonal rainfall over the UBNB, sectors of the Indian, Atlantic, and Pacific Oceans where SSTs show a strong teleconnection with JJAS rainfall in the UBNB (r ≥ 0.4) were identified. An ANN-GA model was developed to forecast the UBNB seasonal rainfall using the selected SST sectors. Results show that ANN-GA forecasted seasonal rainfall amounts that agree well with the observed data for the UBNB [root-mean-square errors (RMSEs) between 0.72 and 0.82, correlation between 0.68 and 0.77, and Hanssen–Kuipers (HK) scores between 0.5 and 0.77], but the results in the foothills region of the Great Rift Valley (GRV) were poor, which is expected since the variability of WPC1 mainly comes from the highlands of Ethiopia. The Valencia and Schaake model was used to disaggregate the forecasted seasonal rainfall to weekly rainfall, which was found to reasonably capture the characteristics of the observed weekly rainfall over the UBNB. The ability to forecast the UBNB rainfall at a season-long lead time will be useful for an optimal allocation of water usage among various competing users in the river basin.


2017 ◽  
Vol 17 (14) ◽  
pp. 8771-8788 ◽  
Author(s):  
Kan Yi ◽  
Junfeng Liu ◽  
George Ban-Weiss ◽  
Jiachen Zhang ◽  
Wei Tao ◽  
...  

Abstract. The response of surface ozone (O3) concentrations to basin-scale warming and cooling of Northern Hemisphere oceans is investigated using the Community Earth System Model (CESM). Idealized, spatially uniform sea surface temperature (SST) anomalies of ±1 °C are individually superimposed onto the North Pacific, North Atlantic, and North Indian oceans. Our simulations suggest large seasonal and regional variability in surface O3 in response to SST anomalies, especially in the boreal summer. The responses of surface O3 associated with basin-scale SST warming and cooling have similar magnitude but are opposite in sign. Increasing the SST by 1 °C in one of the oceans generally decreases the surface O3 concentrations from 1 to 5 ppbv. With fixed emissions, SST increases in a specific ocean basin in the Northern Hemisphere tend to increase the summertime surface O3 concentrations over upwind regions, accompanied by a widespread reduction over downwind continents. We implement the integrated process rate (IPR) analysis in CESM and find that meteorological O3 transport in response to SST changes is the key process causing surface O3 perturbations in most cases. During the boreal summer, basin-scale SST warming facilitates the vertical transport of O3 to the surface over upwind regions while significantly reducing the vertical transport over downwind continents. This process, as confirmed by tagged CO-like tracers, indicates a considerable suppression of intercontinental O3 transport due to increased tropospheric stability at lower midlatitudes induced by SST changes. Conversely, the responses of chemical O3 production to regional SST warming can exert positive effects on surface O3 levels over highly polluted continents, except South Asia, where intensified cloud loading in response to North Indian SST warming depresses both the surface air temperature and solar radiation, and thus photochemical O3 production. Our findings indicate a robust linkage between basin-scale SST variability and continental surface O3 pollution, which should be considered in regional air quality management.


Sign in / Sign up

Export Citation Format

Share Document