scholarly journals DINCAE 1.0: a convolutional neural network with error estimates to reconstruct sea surface temperature satellite observations

2020 ◽  
Author(s):  
Zhaohui Han
Keyword(s):  
Neural Network ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Convolutional Neural Network ◽  
Error Estimates ◽  
Satellite Observations ◽  
Sea Surface

A convolutional neural network with error estimates to reconstruct sea surface temperature satellite observations (DINCAE)

2020 ◽  
Author(s):  
Alexander Barth ◽  
Aida Alvera Azcárate ◽  
Matjaz Licer ◽  
Jean-Marie Beckers
Keyword(s):  
Neural Network ◽  
Missing Data ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Error Estimates ◽  
Satellite Data ◽  
Cross Validation ◽  
Error Variance ◽  
Satellite Observations ◽  
Sea Surface

<p>A method to reconstruct missing data in satellite data using a neural network is presented. Satellite observations working in the optical and infrared bands are affected by clouds, which obscure part of the ocean underneath. In this paper, a neural network with the structure of a convolutional auto-encoder is developed to reconstruct the missing data based on the available cloud-free pixels in satellite images. However, it is unclear how to handle missing data (or data with variable accuracy) in a neural network when using incomplete satellite data in the training phase. The present work shows a consistent approach which uses essentially the satellite data and its expected error variance as input and provides the reconstructed field along with its expected error variance as output. The approach is motivated by the way models and observations are combined in the frame of data assimilation. The neural network is trained by maximizing the likelihood of the observed value. The corresponding error variances are estimated during training and do not need to be known a priori. The approach, called DINCAE (Data-Interpolating Convolutional Auto-Encoder) is applied to a relatively long time-series of Advanced Very High Resolution Radiometer (AVHRR) sea surface temperature data and compared to DINEOF (Data Interpolating Empirical Orthogonal Functions), a method to reconstruct missing data based on an EOF decomposition. The reconstruction error of both approaches is computed using cross-validation and in situ observations from the World Ocean Database. DINCAE results have lower error, while showing higher variability than the DINEOF reconstruction. The resulting error estimates are also validated using the cross-validation data and they follow closely the expected Gaussian distribution.</p>

scholarly journals DINCAE 1.0: a convolutional neural network with error estimates to reconstruct sea surface temperature satellite observations

2020 ◽  
Vol 13 (3) ◽  
pp. 1609-1622 ◽  
Author(s):  
Alexander Barth ◽  
Aida Alvera-Azcárate ◽  
Matjaz Licer ◽  
Jean-Marie Beckers
Keyword(s):  
Neural Network ◽  
Missing Data ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Error Variance ◽  
Temperature Data ◽  
Satellite Observations ◽  
Sea Surface ◽  
Function Decomposition ◽  
Surface Temperature Data

Abstract. A method to reconstruct missing data in sea surface temperature data using a neural network is presented. Satellite observations working in the optical and infrared bands are affected by clouds, which obscure part of the ocean underneath. In this paper, a neural network with the structure of a convolutional auto-encoder is developed to reconstruct the missing data based on the available cloud-free pixels in satellite images. Contrary to standard image reconstruction with neural networks, this application requires a method to handle missing data (or data with variable accuracy) in the training phase. The present work shows a consistent approach which uses the satellite data and its expected error variance as input and provides the reconstructed field along with its expected error variance as output. The neural network is trained by maximizing the likelihood of the observed value. The approach, called DINCAE (Data INterpolating Convolutional Auto-Encoder), is applied to a 25-year time series of Advanced Very High Resolution Radiometer (AVHRR) sea surface temperature data and compared to DINEOF (Data INterpolating Empirical Orthogonal Functions), a commonly used method to reconstruct missing data based on an EOF (empirical orthogonal function) decomposition. The reconstruction error of both approaches is computed using cross-validation and in situ observations from the World Ocean Database. DINCAE results have lower error while showing higher variability than the DINEOF reconstruction.

scholarly journals DINCAE 1.0: a convolutional neural network with error estimates to reconstruct sea surface temperature satellite observations

2019 ◽  
Author(s):  
Alexander Barth ◽  
Aida Alvera-Azcárate ◽  
Matjaz Licer ◽  
Jean-Marie Beckers
Keyword(s):  
Neural Network ◽  
Missing Data ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Satellite Data ◽  
World Ocean ◽  
Error Variance ◽  
Empirical Orthogonal Functions ◽  
Satellite Observations ◽  
Sea Surface

Abstract. A method to reconstruct missing data in satellite data using a neural network is presented. Satellite observations working in the optical and infrared bands are affected by clouds, which obscure part of the ocean underneath. In this paper, a neural network with the structure of a convolutional auto-encoder is developed to reconstruct the missing data based on the available cloud-free pixels in satellite images. However, it is unclear how to handle missing data (or data with variable accuracy) in a neural network when using incomplete satellite data in the training phase. The present work shows a consistent approach which uses essentially the satellite data and its expected error variance as input and provides the reconstructed field along with its expected error variance as output. The neural network is trained by maximizing likelihood of the observed value. The approach, called DINCAE (Data-Interpolating Convolutional Auto-Encoder) is applied to a relatively long time-series of Advanced Very High Resolution Radiometer (AVHRR) sea surface temperature data and compared to DINEOF (Data Interpolating Empirical Orthogonal Functions), a method to reconstruct missing data based on an EOF decomposition. The reconstruction error of both approaches is computed using cross-validation and in situ observations from the World Ocean Database. DINCAE results have lower error, while showing higher variability than the DINEOF reconstruction.

scholarly journals Impacts of Sea Surface Temperature Uncertainty on the Western North Pacific Subtropical High (WNPSH) and Rainfall

2011 ◽  
Vol 26 (3) ◽  
pp. 371-387 ◽  
Author(s):  
Xiaodong Hong ◽  
Craig H. Bishop ◽  
Teddy Holt ◽  
Larry O’Neill
Keyword(s):  
Surface Temperature ◽  
Sea Surface Temperature ◽  
North Pacific ◽  
Western North Pacific ◽  
Forecast Error ◽  
Error Variance ◽  
Satellite Observations ◽  
Sea Surface ◽  
Subtropical High ◽  
Forecast Error Variance

Abstract This paper examines the sensitivity of short-term forecasts of the western North Pacific subtropical high (WNPSH) and rainfall to sea surface temperature (SST) uncertainty using the Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS). A comparison of analyzed SSTs with satellite observations of SST indicates that SST analysis errors are particularly pronounced on horizontal scales from 100 to 200 km, similar to the mesoscale eddy scales in the Kuroshio region. Since significant oceanic variations occur on these scales, it is of interest to examine the effects of representing this small-scale uncertainty with random, scale-dependent perturbations. An SST ensemble perturbation generation technique is used here that enables temporal and spatial correlations to be controlled and produces initial SST fields comparable to satellite observations. The atmospheric model develops large uncertainty in the Korea and Japan area due to the fluctuation in the horizontal pressure gradient caused by the location of the WNPSH. This, in turn, increases the variance of the low-level jet (LLJ) over southeast China, resulting in large differences in the moist transport flux from the tropical ocean and subsequent rainfall. Validation using bin-mean statistics shows that the ensemble forecast with the perturbed SST better distinguishes large forecast error variance from small forecast error variance. The results suggest that using the SST perturbation as a proxy for the ocean ensemble in a coupled atmosphere and ocean ensemble system is feasible and computationally efficient.

scholarly journals Monthly and Quarterly Sea Surface Temperature Prediction Based on Gated Recurrent Unit Neural Network

2020 ◽  
Vol 8 (4) ◽  
pp. 249 ◽  
Author(s):  
Zhen Zhang ◽  
Xinliang Pan ◽  
Tao Jiang ◽  
Baikai Sui ◽  
Chenxi Liu ◽  
...  
Keyword(s):  
Neural Network ◽  
Prediction Model ◽  
Surface Temperature ◽  
Sea Surface Temperature ◽  
Short Term Memory ◽  
Absolute Error ◽  
Sea Surface ◽  
Neural Network Algorithm ◽  
Stability And Accuracy ◽  
Gated Recurrent Unit

The sea surface temperature (SST) is an important parameter of the energy balance on the Earth’s surface. SST prediction is crucial to marine production, marine protection, and climate prediction. However, the current SST prediction model still has low precision and poor stability. In this study, a medium and long-term SST prediction model is designed on the basis of the gated recurrent unit (GRU) neural network algorithm. This model captures the SST time regularity by using the GRU layer and outputs the predicted results through the fully connected layer. The Bohai Sea, which is characterized by a large annual temperature difference, is selected as the study area, and the SSTs on different time scales (monthly and quarterly) are used to verify the practicability and stability of the model. The results show that the designed SST prediction model can efficiently fit the results of the real sea surface temperature, and the correlation coefficient is above 0.98. Regardless of whether monthly or quarterly data are used, the proposed network model performs better than long short-term memory in terms of stability and accuracy when the length of the prediction increases. The root mean square error and mean absolute error of the predicted SST are mostly within 0–2.5 °C.

Sign in / Sign up

Export Citation Format

Share Document