Predicted Sea Level Research Articles

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

Pacific Island countries are vulnerable to the impacts of climate change, which include the risks of increased ocean temperatures, sea level rise and coastal wetland loss. The destruction of wetlands leads not only to a loss of carbon sequestration but also triggers the release of already sequestered carbon, in turn exacerbating global warming. These climate change effects are interrelated, and small island nations continuously need to develop adaptive and mitigative strategies to deal with them. However, accurate and reliable research is needed to know the extent of the climate change effects with future predictions. Hence, this study develops a new hybrid Convolutional Neural Network (CNN) Multi-Layer Bidirectional Long Short-Term Memory (BiLSTM) deep learning model with Multivariate Variational Mode Decomposition (MVMD) to predict the sea level for study sites in the Solomon Islands and Federated States of Micronesia (FSM). Three other artificial intelligence (AI) models (Random Forest (FR), multilinear regression (MLR) and multi-layer perceptron (MLP) are used to benchmark the CNN-BiLSTM model. In addition to this, remotely sensed satellite Landsat imagery data are also used to assess and predict coastal wetland changes using a Random Forest (RF) classification model in the two small Pacific Island states. The CNN-BiLSTM model was found to provide the most accurate predictions (with a correlation coefficient of >0.99), and similarly a high level of accuracy (>0.98) was achieved using a Random Forest (RF) model to detect wetlands in both study sites. The mean sea levels were found to have risen 6.0 ± 2.1 mm/year in the Solomon Islands and 7.2 ± 2.2 mm/year in the FSM over the past two decades. Coastal wetlands in general were found to have decreased in total area for both study sites. The Solomon Islands recorded a greater decline in coastal wetland between 2009 and 2022.

Empirical-singular-wavelet based machine learning models for sea level forecasting in the bosphorus strait: A performance analysis

Forecasting sea level is significant for sustainable water supply management, flood mitigation, shoreline maintenance, ecological sustainability, and economic advancement. In this study, a novel approach that incorporates decomposition methods with machine learning algorithms is presented. The objective is to enhance the accuracy of sea level predictions by tackling the challenges associated with sea level modelling, which arise from its complex characteristics and influence of various factors. This study introduced predictive frameworks incorporating Wavelet Decomposition (WD), Singular Spectrum Analysis (SSA), Empirical Mode Decomposition (EMD) with XGBoost (XGB), Decision Trees (DT), k-Nearest Neighbors (KNN), and Support Vector Machines (SVM). The study conducted predictions for time spans of 1, 7, and 30 days, evaluating the accuracy of the model's predictions using performance metrics. For the first day prediction, the XGB model achieved highest NSE value of 0.86, whereas the hybrid model (SSA-XGB) attained NSE value of 0.99. The hybrid model (EMD-XGB) produced remarkable results, attaining NSE value of 0.70 for the 7 day prediction time horizon. Additionally, the hybrid models (EMD-XGB and EMD-KNN) remarkably achieved NSE values exceeding 0.60 for 30-day time horizon. The results demonstrate that the hybrid models accurately predict daily sea levels with a lead time of up to 7 days. Additionally, the hybrid models demonstrated satisfactory prediction performance for time spans of up to 30 days. The hybrid models presented enhancements in both prediction accuracy and extended lead time predictions.

Sea level rise estimation and projection from long-term multi-mission satellite altimetry and tidal data in the Southeast Asia region

ABSTRACT The Southeast Asian (SEA) region represents a complex coastal geographical location. Most countries in SEA are encircled by the ocean, with lowlands coastal areas. The advantages of long-term satellite altimetry and tide gauge measurements have enabled climate-related research, particularly sea level rise, to be examined extensively in spatial and temporal scales. In this study, sea level anomaly (SLA) derived from 26 years of multi-mission satellite altimeter data plus approximately 28 years of tidal data have been utilized. Then, the projection for every 10 years starting from 2030 until 2100 has been analysed using machine learning regression. The projections are then evaluated based on the global model of the Intergovernmental Panel on Climate Change (IPCC). The findings presented that sea level rise trends around SEA with overall means of 4.70 ± 0.37 mm yr−1 for the satellite altimeter and 3.41 ± 0.24 mm yr−1 for the coastal tide gauge. Sea level projection from coastal tide gauge and satellite altimeter are expected to rise up to 34.87 and 33.71 cm, respectively, in 2100. These outcomes prove that the altimetry data from Radar Altimeter Database System (RADS) and the redundant tidal data are capable in measuring the sea level in SEA. These results are also expected to be valuable for multidisciplinary environmental studies in SEA, such as coastal flooding, coastal erosion, and other effect of global warming. Thus, it would make a meaningful contribution towards Sustainable Development Goal (SDG) 13, climate action, by providing the sea level rate and prediction information.

Abrupt climate fluctuations in Tibet as imprints of multiple meltwater events during the early to mid-Holocene

Understanding the impact of meltwater discharge during the final stage of the Laurentide Ice Sheet (LIS) has important implications for predicting sea level rise and climate change. Here we present a high-resolution ice-core isotopic record from the central Tibetan Plateau (TP), where the climate is sensitive to the meltwater forcing, and explore possible signals of the climate response to potential LIS meltwater discharges in the early to mid-Holocene. The record shows four abrupt large fluctuations during the 7–9 ka BP (kiloannum before present), reflecting large shifts of the mid-latitude westerlies and the Indian summer monsoon (ISM) over this period, and they corresponded to possible LIS freshwater events documented in other paleoclimate records. Our study suggests that multiple rapid meltwater discharge events might have occurred during the final stage of LIS. The finding implies the possibility of rapid sea level rise and unstable climate in the transition zone between the mid-latitude westerlies and the ISM due to fast polar ice retreat under the anthropogenic global warming.

Skillful multiyear to decadal predictions of sea level in the North Atlantic Ocean and U.S. East Coast

Long-term sea-level rise and multiyear to decadal sea level variations pose substantial risks for flooding and erosion in coastal communities. Here we use observations and climate model predictions to show that sea level variations along the U.S. East Coast are skillfully predictable 3 to 10 years in advance. The most predictable component of sea level is a basin scale upward trend, predictable a decade in advance and primarily a response to increasing greenhouse gases. Significant additional predictability comes from multidecadal variations of the Atlantic Meridional Overturning Circulation (AMOC). While perfect model simulations show AMOC-related sea level predictability of 5-7 years, model biases and initialization uncertainties reduce the realized predictive skill to 3-5 years, depending on location. Overall, greenhouse gas warming and predictable AMOC variations lead to multiyear to decadal prediction skill for sea level along the U.S. East Coast. Such skill could have significant societal benefit for planning and adaptation.

Clarifying the Distinction between Steric and Baroclinic Sea Surface Height

Abstract One of the most fundamental uses of ocean models is for the prediction of sea level. Vertical integration of the hydrostatic equation leads to the partitioning of sea level in terms of atmospheric pressure, steric height, and bottom pressure. In an effort to validate the baroclinic wave dynamics of numerical ocean models, some researchers have compared the steric height from models with the sea level anomaly derived from satellite altimetry. The use of steric height in these comparisons captures the qualitative aspects of the baroclinic waves, but it neglects a nonnegligible contribution from bottom pressure. A more accurate evaluation of baroclinic wave dynamics using sea level would involve projecting the pressure field onto orthogonal barotropic and baroclinic components to isolate the baroclinic sea level anomaly. This note illustrates the quantitative difference between steric height and baroclinic sea level, which amounts to approximately a 20% bias of steric height over baroclinic sea level, depending on location. Significance Statement The bottom pressure variability associated with superinertial internal waves is not negligible, but it has been neglected in a recent series of ocean model validation studies. Baroclinic surface pressure, and its associated sea level anomaly, should be used instead of steric height for model–data comparisons of baroclinic sea level variability.

Prediction of Sea Level in the Arabian Gulf Using Artificial Neural Networks

Creating an efficient model for predicting sea level fluctuations is essential for climate change research. This study examined the effectiveness of utilizing Artificial Neural Networks (ANNs), particularly the recurrent network approach. ANNs were chosen for their capacity to learn from extensive and intricate data and their ability to handle nonlinear correlations. The Long Short-Term Memory (LSTM) algorithm was employed to fill data gaps and predict future sea level records in the Arabian Gulf, especially in Mina Salman. The results were promising, with LSTM successfully filling a 6-year data gap while maintaining low Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) values. The first phase of the model yielded a RMSE value of 63.4 mm and a MAPE value of 3.14%. The same approach was used to retrain the model with a mix of real and predicted values, preserving historical patterns and yearly rates with an RMSE of 66.5 mm and a MAPE of 3.07%. These findings highlight LSTM’s advantages when considering only historical information for predicting the future sea level changes. The research provides valuable insights into predicting sea level changes in regions with limited field data, such as the Arabian Gulf, and emphasizes the potential for further research to enhance sea level prediction models through improved optimization techniques.

Generalized Additive Models for Predicting Sea Level Rise in Coastal Florida

Within the last century, the global sea level has risen between 16 and 21 cm and will likely accelerate into the future. Projections from the Intergovernmental Panel on Climate Change (IPCC) show the global mean sea level (GMSL) rise may increase to up to 1 m (1000 mm) by 2100. The primary cause of the sea level rise can be attributed to climate change through the thermal expansion of seawater and the recession of glaciers from melting. Because of the complexity of the climate and environmental systems, it is very difficult to accurately predict the increase in sea level. The latest estimate of GMSL rise is about 3 mm/year, but as GMSL is a global measure, it may not represent local sea level changes. It is essential to obtain tailored estimates of sea level rise in coastline Florida, as the state is strongly impacted by the global sea level rise. The goal of this study is to model the sea level in coastal Florida using climate factors. Hence, water temperature, water salinity, sea surface height anomalies (SSHA), and El Niño southern oscillation (ENSO) 3.4 index were considered to predict coastal Florida sea level. The sea level changes across coastal Florida were modeled using both multiple regression as a broadly used parametric model and the generalized additive model (GAM), which is a nonparametric method. The local rates and variances of sea surface height anomalies (SSHA) were analyzed and compared to regional and global measurements. The identified optimal model to explain and predict sea level was a GAM with the year, global and regional (adjacent basins) SSHA, local water temperature and salinity, and ENSO as predictors. All predictors including global SSHA, regional SSHA, water temperature, water salinity, ENSO, and the year were identified to have a positive impact on the sea level and can help to explain the variations in the sea level in coastal Florida. Particularly, the global and regional SSHA and the year are important factors to predict sea level changes.

Protective structures as adaptive management strategy in Nature-based Solutions to mitigate sea level rise effects

Nature-based Solutions (NbS) that integrate salt marshes with traditional coastal protection structures are gaining interest in the face of ongoing climate change. However, major uncertainties remain about the combined impact of sea level rise and marsh-supporting management measures (e.g. nourishments and brushwood dams) on vegetation establishment. This study presents inundation percentages, bed level changes and vegetation characteristics of the dominant pioneer species of the constructed Marconi salt marsh for the growing season of 2021. This field data identified hydrodynamic and morphodynamic thresholds for vegetation establishment: <12% monthly inundation, and −0.8 cm < monthly bed level change < 1.4 cm. The existing morphodynamic DET-ESTMORF model was extended with these thresholds for vegetation establishment. A scenario analysis with this model revealed that the implementation of a brushwood dam accommodates the vegetation to establish further seaward. On the long term, until 2100, predicted sea level rise was found to cause the vegetation to retreat, but this retreat was substantially delayed by the presence and positioning of a protective structure. These results emphasize the need for adaptive management of NbS.

The sands of time: Predicting sea level rise impacts to barrier island habitats

Coastal beach ecosystems support critical habitat for numerous species and are vulnerable to sea level rise. Sand beaches are spatially and temporally dynamic, making it difficult to accurately predict future habitat loss – estimates that are crucial as species are being assessed for protection. We mapped sand beach habitat on 12 focal barrier islands and low-lying beaches off the Gulf Coast of Florida, USA using two methods for comparison - a remotely sensed land cover and hand-digitized aerial imagery that was collected concurrently with digital elevation data. We then compared estimates of beach habitat lost to sea level rise between the two methods and compared this sensitivity to control mangrove islands. Predictions suggest that most of the beach habitat in our study areas will be underwater within the next century. These beaches represent critical nesting habitat for vulnerable vertebrate species such as the snowy plover (Charadrius nivosus) and the loggerhead turtle (Caretta caretta). Importantly, we found that for some islands, using remotely sensed land cover data led us to under- or over-estimate the amount of beach habitat lost to sea level rise relative to the digitized data because of the temporal mismatch between land cover and elevation data. In contrast, we found relatively little difference between methods in the amount of mangrove forest lost to sea level rise on nearby control islands. We suggest that using land cover data collected at the same time as elevation data can produce more accurate predictions of beach habitat inundation from sea level rise.

Prediction of the Sea Level from the PUMMA System Using SARIMA

The rising sea levels can threaten millions of people residing along the coast or lowlands. The risk can be mitigated by the sea-level prediction done by collecting information on the likelihood of rising sea levels. The Ministry of Marine Affairs and Fisheries of Indonesia has developed Perangkat Ukur Murah untuk Muka Air Laut (Inexpensive Device for Sea Level Measurement, PUMMA) to measure sea levels. PUMMA is located in remote monitoring stations based on Indonesian maritime area. The PUMMA system currently lacks a prediction feature. This objective of this study is to model the sea-level prediction using the dataset for one year, from July 2021 until July 2022. The seasonal autoregressive integrated moving average (SARIMA) method was used because SARIMA proved to be a flexible and versatile method for a dataset having noncomplex nature and seasonal patterns. This study has developed several models of the SARIMA. The model performance was evaluated using the mean absolute percentage error (MAPE), R-squared, mean square error (MSE), and root mean square error (RMSE) metrics. The SARIMA(1, 1, 0)(1, 1, 1)12 model achieved the lowest prediction error with an R-squared of 0.508, MSE of 0.0479, and RMSE of 0.069. Based on the performance, SARIMA(1, 1, 0)(1, 1, 1)12 model is feasible for predicting sea levels using the PUMMA dataset.

Trend Analysis of Tides Level and Projection Sea Level Rise On the West Coast of Peninsular Malaysia

Globally, sea levels are rising due to climate change caused by humans. Current and future sea level rise will have a variety of effects, particularly on coastal infrastructure. This study analysed the tidal level trend and projected sea level rise along the western coast of peninsular Malaysia using tide level records collected between 1986 and 2012. The seasonal Mann-Kendall test and linear trend were used to determine the tide level trend over time and to predict sea level rise for 2050 and 2100. Based on a confidence interval of 95%, the results of the analysis indicated a rising trend at all stations. Tau values for the Mann-Kendall test range between 0.16 and 0.33, while seasonal Mann-Kendall values range between 0.18 and 0.41. Based on the sea level projection analysis, the mean sea level will rise between 8.77 cm and 14.29 cm in 2050, and between 20.44 cm and 33.20 cm in 2100. In conclusion, the sea level trend at all stations on the west coasts of peninsular Malaysia exhibits an upward trend.

Predicting sea levels using ML algorithms in selected locations along coastal Malaysia

In consideration of the distinct behavior of machine learning (ML) algorithms, six well-defined ML used were carried out in this study for predicting sea level on a day-to-day basis. Data compiled from 1985 to 2018 was utilized for training and testing the developed models. An assessment of the multiple statistics-driven regression algorithms resulted such that each tested location was associated with a particular preferred model. The following were the developed best models for their respective study areas: In Peninsular Malaysia, the interactions linear regression model was the best at Pulau Langkawi (RMSE = 19.066), the Matern 5/2 gaussian process regression model at Geting (RMSE = 49.891), and the trilayered artificial neural network at Pulau Pinang (RMSE = 20.026), while the linear regression model was the best at Sandakan in Sabah, East Malaysia (RMSE = 14.054). Other metrics, such as MAE and R-square, were also at their best values, each providing its best values, further substantiating the RMSE respectively, at each of the study areas. These empirical statistics (or metrics) also revealed that despite employing sea level as the sole parameter, results obtained were exceptional better when utilizing a 7-day lag, regardless of the model used. Notably, lag variables with less than a 7-day lag could degrade the model's accuracy in representing ground reality. The study emphasizes the importance of thorough training and testing of ML to aid decision-makers in developing mitigation actions for the climate change phenomena of sea level rise through reliable ML.

Multiscale Analysis and Prediction of Sea Level in the Northern South China Sea Based on Tide Gauge and Satellite Data

Under the influence of global warming, the problem of sea-level rise is becoming increasingly prominent. The northern part of the South China Sea (SCS) is low lying, with intense economic development, and densely populated. These characteristics make the region extremely sensitive to the consequences of rising sea levels. This study aims to reveal the trends of sea-level changes in the northern SCS and provide scientific insights into the potential flooding risks in low-lying areas. To achieve this, the Ensemble Empirical Mode Decomposition (EEMD) method is used to analyze the water level time series data from three tide gauges along the coast of Hong Kong. This analysis reveals the multidimensional change characteristics and response mechanisms of the sea level in the SCS. The findings reveal distinct seasonal, interannual, decadal, and interdecadal variations in sea-level changes. Furthermore, we explore the impact of the El Niño-Southern Oscillation (ENSO) on sea-level changes in the study area, finding a 6-month lagged correlation between the sea level and ENSO. Spatially, the rate of sea-level change is faster in nearshore areas than in the open ocean and higher in the northern regions than in the southern regions. The Multifractal Detrended Fluctuation Analysis (MF-DFA) method is employed to analyze the sea-level change time series, revealing long-range correlations and multifractal characteristics. In addition, we propose a sea-level prediction method that combines EEMD with Long Short-Term Memory (LSTM) neural networks and conducts empirical research on sea-level changes in the northern South China Sea. The results indicate that the EEMD-LSTM model outperforms the standalone LSTM model in terms of predictive accuracy, effectively eliminating noise from signals and providing a valuable reference. In summary, this research delves into the multiscale characteristics and influencing factors of sea-level changes in the northern SCS, proposing an improved sea-level prediction method that integrates EEMD and LSTM. The findings lay the groundwork for evaluating the risks of sea-level rise in low-lying regions of the northern SCS and inform future response strategies.

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

The prediction of sea level rise is extremely important for improved future climate change mitigation and adaptation strategies. This study uses a hybrid convolutional neural Network (CNN) and a bidirectional long short-term (BiLSTM) model with successive variational mode decomposition (SVMD) to predict the absolute sea level for two study sites in Australia (Port Kembla and Milner Bay). More importantly, the sea level measurements using a tide gauge were corrected using Global Navigation Satellite System (GNSS) measurements of the vertical land movement (VLM). The SVMD-CNN-BiLSTM model was benchmarked by a multi-layer perceptron (MLP), support vector regression (SVR) and gradient boosting (GB). The SVMD-CNN-BiLSTM model outperformed all the comparative models with high correlation values of more than 0.95 for Port Kembla and Milner Bay. Similarly, the SVMD-CNN-BiLSTM model achieved the highest values for the Willmott index, the Nash–Sutcliffe index and the Legates and McCabe index for both study sites. The projected linear trend showed the expected annual mean sea rise for 2030. Using the current trend, Port Kembla was projected to have an MSL value of 1.03 m with a rate rise of approx. 4.5 mm/year. The rate of the MSL for Milner Bay was comparatively lower with a value of approx. 2.75 mm/year and an expected MSL value of 1.27 m for the year 2030.

Using EEMD mode decomposition in combination with machine learning models to improve the accuracy of monthly sea level predictions in the coastal area of China

In the context of climate change and human activities, the global sea level is facing a rising trend, which poses serious challenges to the ecological environment of coastal areas. In this study, we selected the monthly mean sea level (MSL) time series of 9 stations in the coastal areas of China as the research object. First, we analyzed the spatiotemporal distribution characteristics of the monthly MSL in the coastal areas of China. Secondly, we analyzed the ability of ensemble empirical mode decomposition (EEMD) to decompose the monthly MSL series. Finally, we choose three machine learning models, namely Back Propagation (BP), K-Nearest Neighbor (KNN), and Long Short-Term Memory (LSTM) neural network models to compare model prediction effect between single machine learning models with machine learning models combined with EEMD. The results show that except for the YANTAI (YT) station, which showed an insignificant downward trend, the monthly MSL of other stations showed an upward trend, indicating that the coastal areas of China are facing the risk of sea level rise. EEMD can effectively reduce the complexity of the original monthly MSL time series, and different intrinsic mode functions (IMFs) reflect changes in monthly MSL at different frequencies. Comparing the single machine learning model and the machine learning model combined with EEMD, it is found that the simulation effect of the machine learning model combined with EEMD is better than that of the single model. The model with the best prediction effect on monthly MSL in the coastal areas of China is LSTM-EEMD, followed by KNN-EEMD. This study provides an important reference for systematically understanding sea level changes and selecting an appropriate monthly MSL prediction model in the coastal areas of China.

Spatial sedimentation and plant captured sediment within seagrass patches

Habitat degradation in coastal ecosystems has resulted in the fragmentation of coastal aquatic vegetation and compromised their role in supplying essential ecological services such as trapping sediment or sequestering carbon. Fragmentation has changed seagrass architecture by decreasing the density of the canopy or engendering small patches of vegetated areas. This study aims to quantify the role different patch sizes of vegetation with different canopy densities have in the spatial distribution of sediment within a patch. To this aim, two canopy densities, four different patch lengths, and two wave frequencies were considered. The amounts of sediment deposited onto the bed, captured by plant leaves, remaining in suspension within the canopy, and remaining in suspension above the canopy were used to understand the impact hydrodynamics has on sediment distribution patterns within seagrass patches. In all the cases studied, patches reduced the suspended sediment concentrations, increased the capture of particles in the leaves, and increased the sedimentation rates to the bed. For the lowest wave frequency studied (0.5 Hz), the sediment deposited to the bottom was enhanced at canopy edges, resulting in spatial heterogeneous sedimentation patterns. Therefore, restoration and preservation of coastal aquatic vegetation landscapes can help face future climate change scenarios where an increase in sedimentation can help mitigate predicted sea level rise in coastal areas.

Multistability and Transient Response of the Greenland Ice Sheet to Anthropogenic CO2 Emissions

AbstractUnderstanding the future fate of the Greenland Ice Sheet (GIS) in the context of anthropogenic CO2 emissions is crucial to predict sea level rise. With the fully coupled Earth system model of intermediate complexity CLIMBER‐X, we study the stability of the GIS and its transient response to CO2 emissions over the next 10 Kyr. Bifurcation points exist at global temperature anomalies of 0.6 and 1.6 K relative to pre‐industrial. For system states in the vicinity of the equilibrium ice volumes corresponding to these temperature anomalies, mass loss rate and sensitivity of mass loss to cumulative CO2 emission peak. These critical ice volumes are crossed for cumulative emissions of 1,000 and 2,500 GtC, which would cause long‐term sea level rise by 1.8 and 6.9 m respectively. In summary, we find tipping of the GIS within the range of the temperature limits of the Paris agreement.

Extreme sea level changes over the tropical western Pacific in 1.5 °C and 2.0 °C warmer climates

The tropical western Pacific and the adjacent South China Sea are home to many low-lying islands and coastal zones that are vulnerable to flood hazards resulting from extreme sea level (ESL) changes. Based on the hourly sea level recorded by 15 tide gauges during the period 1980-2018, this study evaluates the historical trend of ESLs over this region. On this basis, a regression model for hourly future sea-level prediction is established by combining the atmospheric reanalysis products, the tidal harmonics, and the outputs of three climate models archived by the Coupled Model Intercomparison Project Phase 6 (CMIP6) to evaluate the future ESL changes in 1.5 °C and 2.0 °C warmer climates. The historical trend of ESLs show that the ESLs along the coasts and islands of the tropical western Pacific have significantly risen during the past decades, which is mainly contributed by the mean sea level rise. And results from the historical observations and the prediction model show that in a warming climate from 1980 to 2050, both the mean sea levels and ESLs rise with fluctuations. The mean sea level change plays an important role in the secular trend of ESLs, while the interannual-to-decadal variability of ESLs is significantly affected by tides and extreme weather events. Under the warming scenario of 1.5°C, the changes in the return levels of ESL relative to the historical period are generally small at most tide gauge sites. Compared with the situations under 1.5°C warming, the return levels of ESL at most selected tide gauges will rise more significantly under the 2.0°C warming scenario, so the frequency of the current 100-year return level will reduce to less than 10 years at most stations. The above results suggest that this additional 0.5°C warming will cause a huge difference in the ESLs along the coasts and islands of the tropical western Pacific. As proposed in the Paris climate agreement, it is very necessary to limit anthropogenic warming to 1.5°C instead of 2.0°C, which will substantially reduce the potential risk of flood disasters along the coasts and islands of the tropical western Pacific.

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

AbstractMultisatellite systems and multi‐signal‐to‐noise ratio types provide a more prosperous data basis for the inversion of sea level by GNSS‐IR technology. However, there are few studies on data reconstruction of the retrieved sea level height. This study takes the SPBY station as an example and introduces the least squares learning cross‐checking method with the Gaussian kernel model (GKM) L2 constraint. Furthermore, research on hourly sea level reconstruction based on GNSS‐IR is carried out. Compared with the measured sea level, the sea level height in 2021 obtained after the fusion of multisource data has an R of 0.905 and an RMSE of 0.144 m. And the result obtained after data reconstruction has an R of 0.958 and an RMSE of 0.090 m. Compared with the multisource data fusion results before reconstruction, R is increased by 5.9%, and RMSE is decreased by 37.5%. Finally, using the reconstructed sea level data based on Long Short‐Term Memory (LSTM) artificial neural network to carry out the research on sea level prediction, which verifies the conclusion that more reliable forecast values can be obtained based on 5 months of training data. Among them, the R from 1 to 24 hr is 0.905, and the RMSE is 0.145 m. Compared with the inversion accuracy of GNSS‐IR, the R is increased by 2.0%, and the RMSE is decreased by 21.4%. This study demonstrates the feasibility of GNSS‐IR technology and L2‐constrained least squares learning cross‐checking method based on the GKM to reconstruct sea level data with high temporal resolution and high accuracy. The reliability of the sea level prediction based on GKM reconstruction and LSTM is verified in the sea level forecast of the next 24 epochs, which has essential applications in sea level data recovery and forecasting.