Abstract:Compared to traditional swell height measurement methods, spaceborne global navigation satellite system reflectometry (GNSS-R) has many advantages, including remote sensing capabilities, global coverage, real-time monitoring, etc. It can provide wave observation data with high spatiotemporal resolution and is not limited by time, weather, and other conditions. Spaceborne GNSS-R provides a very effective method for estimating swell height, which can monitor and measure wave changes over a large area of the ocean surface in real time. This is of great significance for understanding the marine environment, climate change, and weather forecasting. However, there is relatively little research on the estimation of swell height using this technology, especially in the retrieval model of swell height. For this purpose, the article proposes a global ocean swell height retrieval method based on the convolutional neural network (CNN), convolutional long short-term memory (ConvLSTM) and fully connected network (FCN) hybrid deep learning model (i.e., CNN-ConvLSTM-FCN) for spaceborne GNSS-R. CNN-ConvLSTM-FCN model not only uses CNN to extract spatial features around specular points (SPs) from a two-dimensional (2-D) matrix of a single image (bistatic radar scattering cross-section (BRCS), effective scattering area, or power delay-Doppler map (DDM), but also uses ConvLSTM network to infer feature relationships and FCN to output estimated swell heights. The hybrid model improves its retrieval ability by simultaneously considering feature information related to time and space. The performance of the CNN-ConvLSTM-FCN model in retrieving swell height was tested using ERA5 and WaveWatch III swell height as reference data. The results show that when ERA5 data is used as a reference, compared to the empirical model method based on DDM average (DDMA) observable, the proposed CNN-ConvLSTM-FCN swell height retrieval model improves root mean square error (RMSE), correlation coefficient (CC), and mean absolute percentage error (MAPE) by 50.76%, 26.28%, and 29.63%, respectively. When WaveWatch III data is used as a reference, improvements in RMSE, CC, and MAPE are 51.09%, 25.35%, and 44.21%, respectively. The CNN-ConvLSTM-FCN model can demonstrate the ability of high-precision and high-resolution ocean swell height retrieval on a global scale, providing a new reference method for spaceborne GNSS-R ocean swell height estimation.

Prediction of Mean Sea Level with GNSS-VLM Correction Using a Hybrid Deep Learning Model in Australia

Prediction of Sea Level with Vertical Land Movement Correction Using Deep Learning

A deep-learning model for national scale modelling and mapping of sea level rise in Malaysia: the past, present, and future

Assessment and Prediction of Sea Level and Coastal Wetland Changes in Small Islands Using Remote Sensing and Artificial Intelligence

Advancing Accuracy in Sea Level Estimation with GNSS-R: A Fusion of LSTM-DNN-Based Deep Learning and SNR Residual Sequences

An improved sparrow search algorithm and CNN-BiLSTM neural network for predicting sea level height

Leveraging GNSS tropospheric products for machine learning-based land subsidence prediction

Improved tide level prediction model combined GA-BP neural networks and GNSS SNR data

Sea level prediction using ARIMA, SVR and LSTM neural network: assessing the impact of ensemble Ocean-Atmospheric processes on models’ accuracy

A deep-learning model for rapid spatiotemporal prediction of coastal water levels

New double decomposition deep learning methods for river water level forecasting

Multi-decadal Sea Level Prediction using Neural Networks and Spectral Clustering on Climate Model Large Ensembles and Satellite Altimeter Data

Enhancing GNSS Deformation Monitoring Forecasting with a Combined VMD-CNN-LSTM Deep Learning Model

Estimating sea surface swell height using a hybrid model combining CNN, ConvLSTM, and FCN based on spaceborne GNSS-R data from the CYGNSS mission

Nearshore significant wave height prediction based on MIC-LSTM model

Short-Term Prediction of Groundwater Level Based on Spatiotemporal Correlation

Sea-Level Estimation from GNSS-IR under Loose Constraints Based on Local Mean Decomposition

Sea level Projections with Machine Learning using Altimetry and Climate Model ensembles

Enhancing Sea Level Rise Estimation and Uncertainty Assessment from Satellite Altimetry through Spatiotemporal Noise Modeling

Enhancing Sea Surface Height Retrieval with Triple Features Using Support Vector Regression

Hourly Sea Level Prediction‐Based GNSS‐IR Inversions by Combining the Least Squares Learning Cross‐Checking Method With the Gaussian Kernel Model L2 Constraint and LSTM