Machine Learning–Based Uncertainty Analysis of Multi-Temporal GEBCO Bathymetry in the Nigerian EEZ

Machine Learning–Based Uncertainty Analysis of Multi-Temporal GEBCO Bathymetry in the Nigerian EEZ | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine Learning–Based Uncertainty Analysis of Multi-Temporal GEBCO Bathymetry in the Nigerian EEZ Felicia Chinwe Mogo, Kehinde Chima Opara, Desmond Rowland Eteh, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8756450/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 15 You are reading this latest preprint version Abstract Reliable bathymetric information is essential for marine spatial planning, offshore engineering, and environmental management, yet large portions of the global ocean remain constrained by sparse acoustic survey coverage. This study presents a machine learning–based uncertainty analysis of multi-temporal GEBCO bathymetry (2019–2025) within the Nigerian Exclusive Economic Zone (EEZ), a data-limited region of high strategic relevance. Rather than generating a new bathymetric surface, the analysis focuses on quantifying and predicting spatial bathymetric uncertainty associated with successive updates of global bathymetric products. Temporal bathymetric differencing (ΔZ) was applied to successive GEBCO releases to characterise systematic depth refinement and to derive proxy uncertainty metrics. Morphometric attributes, temporal refinement indicators, and data-provenance variables distinguishing predicted from acoustic measured bathymetry were used as predictors in ensemble machine-learning models. Model performance was evaluated using 10-fold cross-validation, with accuracy assessed using the coefficient of determination (R²) and root mean square error (RMSE), which quantify prediction error in proxy uncertainty, expressed in metres, rather than seabed depth. Results indicate that the large-scale geomorphological structure of the Nigerian EEZ remains stable across all datasets, while successive GEBCO releases show statistically significant mean depth refinements of 0.8–2.3 m per release and a cumulative adjustment of approximately 4.2 m between 2019 and 2025. Proxy bathymetric uncertainty declined by approximately 70% between 2019 and 2023. The Random Forest model achieved the highest predictive performance (R² = 0.86; RMSE = 7.8 m), enabling spatially explicit uncertainty mapping to support survey prioritisation and marine governance. Bathymetry Machine learning Uncertainty analysis Multi-temporal GEBCO Nigerian Exclusive Economic Zone Marine spatial planning Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Highlights • Multi-temporal GEBCO bathymetry (2019–2025) analysed for uncertainty evolution in the Nigerian EEZ • Temporal differencing used to quantify bathymetric refinement and proxy uncertainty • Machine learning applied to predict spatial patterns of bathymetric uncertainty • Random Forest achieved high predictive performance (R² = 0.86; RMSE = 7.8 m) • Results support uncertainty-aware survey prioritisation and marine governance 1. Introduction Accurate seafloor bathymetry underpins marine spatial planning, offshore engineering, hazard assessment, and environmental governance. Bathymetric information supports decisions ranging from offshore infrastructure siting and submarine cable routing to habitat mapping and the delimitation of maritime boundaries under the United Nations Convention on the Law of the Sea (UNCLOS) [ 1 – 3 ]. Such applications are central to ecosystem-based management, offshore resource governance, and blue-economy planning [ 4 , 5 ]. Despite this importance, more than 80% of the global seafloor remains inadequately mapped using direct acoustic measurements, particularly in deep-water regions where satellite-altimetry-derived predictions dominate global bathymetric grids [ 6 – 9 ]. The General Bathymetric Chart of the Oceans (GEBCO) represents the most widely used global bathymetric compilation, integrating multibeam and single-beam echosoundings, satellite-derived gravity predictions, and regional datasets into a uniform 15-arc-second grid [ 3 , 10 ]. Earlier global models such as SRTM30_PLUS and SRTM15 + laid the foundation for modern bathymetric synthesis by combining ship soundings and gravity data at increasing spatial resolutions [ 11 , 12 ]. Successive GEBCO releases (e.g., 2019, 2021, 2023, and 2025) incorporate newly acquired multibeam surveys, refined interpolation approaches, and improved gravity-model constraints under international initiatives such as the Nippon Foundation–GEBCO Seabed 2030 project [ 13 , 14 ]. Consequently, global bathymetric products evolve through time, exhibiting incremental depth adjustments and changing uncertainty that reflect data integration and methodological refinement rather than physical modification of the seafloor [ 15 – 17 ]. Multi-temporal analysis of successive bathymetric grids has therefore emerged as a robust approach for evaluating refinement trajectories and tracking uncertainty evolution. Temporal differencing (ΔZ) between releases enables the identification of regions where gravity-predicted bathymetry has been replaced by higher-precision acoustic measurements, often manifesting as systematic depth adjustments of a few metres between versions [ 18 – 20 ]. Such approaches have been widely applied in well-surveyed regions, including Arctic bathymetric charting and regional marine data infrastructures, but remain under utilised in West African waters despite persistent data scarcity and strategic importance [ 3 , 21 ]. The Nigerian Exclusive Economic Zone (EEZ), extending approximately 200 nautical miles into the Gulf of Guinea, exemplifies this challenge. The region encompasses a passive continental margin characterised by a narrow continental shelf, steep continental slope, continental rise, and extensive abyssal plain shaped by Niger Delta sedimentation and long-term tectono-stratigraphic evolution [ 22 – 24 ]. Seafloor morphology within this setting exerts strong controls on benthic habitat zonation, sediment routing systems, and submarine fan development [ 25 – 27 ]. Although the Nigerian EEZ hosts critical offshore energy infrastructure, fisheries, and maritime transport corridors, large portions of its deep-water domain remain dependent on predicted bathymetry, limiting confidence in geohazard assessment, offshore engineering design, and environmental management [ 28 – 30 ]. Recent advances in machine learning (ML) have transformed bathymetric analysis across satellite-derived bathymetry, bathymetric inversion, geomorphological classification, and uncertainty quantification [ 31 – 35 ]. ML algorithms such as Random Forest, Gradient Boosting, convolutional neural networks, and hybrid physics-guided models have demonstrated strong capacity to capture non-linear relationships between depth, morphometric attributes, spectral information, and geophysical predictors [ 36 – 39 ]. Increasingly, ML frameworks are also being used to quantify spatially explicit bathymetric uncertainty rather than relying solely on traditional interpolation-based error assumptions [ 20 , 40 – 42 ]. Parallel developments in Nigeria and other data-scarce environments demonstrate the growing adoption of ML and geospatial analytics for hydrological modelling, shoreline change detection, flood mapping, sediment dynamics, and environmental risk assessment [ 43 – 50 ]. These studies highlight the effectiveness of combining morphometric indicators, temporal analysis, and data-driven modelling in complex coastal and fluvial systems, providing a transferable methodological foundation for offshore bathymetric uncertainty analysis. Despite these advances, a critical gap persists. Most ML-based bathymetric studies focus on generating or refining depth surfaces, while relatively few explicitly address how uncertainty evolves across successive global bathymetric compilations. Where uncertainty is considered, the influence of data provenance (e.g., gravity-derived versus acoustically measured bathymetry), temporal refinement, and seafloor morphology is often discussed qualitatively rather than modelled quantitatively [ 16 , 17 ]. This limitation introduces ambiguity when interpreting inter-release depth changes, particularly in distinguishing refinement driven by improved data coverage and modelling from signals that could be misinterpreted as genuine seafloor change [ 9 , 17 ]. To address these gaps, this study applies a machine-learning-based uncertainty analysis to multi-temporal GEBCO bathymetry (2019–2025) within the Nigerian EEZ. The ML component is not used to generate a new bathymetric surface; instead, it is employed to predict and map spatial patterns of proxy bathymetric uncertainty using temporal bathymetric differencing (ΔZ) and variance-based metrics as response variables. Predictor variables comprise (i) seafloor morphometric characteristics, (ii) temporal refinement indicators capturing changes between successive GEBCO releases, and (iii) provenance information differentiating gravity-derived bathymetry from acoustically surveyed data [ 11 , 12 , 15 – 17 , 40 ]. Guided by this framework, the study investigates: (1) variations in bathymetric depth distributions and geomorphological zonation across GEBCO releases from 2019 to 2025; (2) the magnitude, spatial distribution, and statistical significance of temporal bathymetric refinements (ΔZ); and (3) the evolution of proxy bathymetric uncertainty in relation to shifts in data provenance. The study further evaluates whether ML models can reliably predict spatial uncertainty patterns using combined morphometric, temporal, and provenance-based predictors [ 20 , 41 , 42 ]. The central hypothesis is that integrating ML with multi-temporal GEBCO bathymetry enables robust, spatially explicit prediction of bathymetric uncertainty, revealing systematic uncertainty reduction linked to increasing acoustic survey integration while maintaining the stability of large-scale seafloor morphology [ 13 – 17 , 40 , 51 – 54 ]. 2. Materials and Methods 2.1 Study Area: Nigerian Exclusive Economic Zone (EEZ) The study area comprises the Nigerian Exclusive Economic Zone (EEZ), located in the eastern tropical Atlantic Ocean within the Gulf of Guinea. The EEZ extends up to 200 nautical miles (~ 370 km) offshore and covers approximately 217,000 km² [ 21 ]. Geographically, it spans approximately 2.5° E–9.5° E longitude and 1.5° N–6.5° N latitude (Fig. 1 ). The Nigerian EEZ exhibits the physiographic characteristics of a passive continental margin formed during the Mesozoic separation of the African and South American plates [ 16 ]. The region consists of a narrow continental shelf (< 80 km), a steep continental slope, a continental rise, and extensive abyssal plains reaching depths of approximately − 4,900 m [ 10 ]. Sedimentation from the Niger Delta strongly influences slope morphology and deep-sea fan development [ 18 ]. These characteristics make the region suitable for evaluating bathymetric refinement and uncertainty evolution across depth zones. 2.2 Multi-Temporal GEBCO Bathymetric Datasets Bathymetric data were obtained from the General Bathymetric Chart of the Oceans (GEBCO), which provides global gridded bathymetry at a spatial resolution of 15 arc-seconds (~ 500 m) [ 2 ]. Four successive GEBCO releases 2019, 2021, 2023, and 2025 were analysed to evaluate temporal refinement and uncertainty evolution (Table 1 ). Each GEBCO release integrates heterogeneous data sources, including multibeam and single-beam echosoundings, satellite-derived gravity predictions, and regional compilations developed under the Seabed 2030 initiative [ 13 ],[ 7 ]. While all datasets share a common spatial resolution and vertical datum (mean sea level), the proportion of acoustically measured versus gravity-derived bathymetry varies spatially and temporally, necessitating explicit consideration of data provenance in subsequent analyses. All datasets were provided as GeoTIFF rasters in the WGS84 geographic coordinate system (EPSG:4326) and were clipped to the Nigerian EEZ boundary to ensure spatial consistency across releases. Table 1 Key characteristics of the datasets used in this study. Dataset Release Year Resolution Vertical Datum Coverage Key Features GEBCO_2019 2019 15 arc sec (~ 500 m) Mean Sea Level (MSL) Global First release using SRTM15 + base grid GEBCO_2021 2021 15 arc sec (~ 500 m) MSL Global Integration of Seabed 2030 regional contributions; improved gravity model GEBCO_2023 2023 15 arc sec (~ 500 m) MSL Global Enhanced multibeam coverage; refined TID classes GEBCO_2025 2025 15 arc sec (~ 500 m) MSL Global Incorporation of new multibeam campaigns and SRTM15 + v2.5.5 2.3 Data Harmonisation and Pre-Processing To ensure comparability between datasets, all GEBCO releases were harmonised prior to analysis by enforcing: identical spatial resolution (15 arc-seconds), consistent spatial extent (Nigerian EEZ), common vertical datum (mean sea level), and cell-by-cell spatial alignment. Raster alignment was verified to avoid artificial depth differences arising from grid offsets. Land and no-data cells were masked consistently across all releases. No reprojection was applied to avoid interpolation artefacts, consistent with established global bathymetric analysis practices [ 2 ]. 2.4 Bathymetric Metrics and Geomorphological Classification 2.4.1 Descriptive Statistics For each GEBCO release, descriptive statistics including minimum, maximum, mean, median, standard deviation, and selected percentiles were calculated to summarise bathymetric distributions and assess temporal stability (Table 2 ). 2.4.2 Depth Zonation Bathymetric grids were classified into four primary physiographic depth zones following established global seafloor classification schemes [ 2 ], [ 5 ].: Continental shelf: 0 to − 200 m Upper continental slope: −200 to − 1,000 m Continental rise: −1,000 to − 3,000 m Abyssal plain: < −3,000 m Spatial distributions of these depth classes were mapped for each GEBCO release to assess structural stability and zonal transitions 2.4.3 Area Estimation Area calculations were performed using spherical geometry to account for latitudinal variation in grid-cell size. Total seabed area for each depth class was computed by summing pixel areas within each zone, enabling quantitative comparison across releases (Table 3 ). The area (A) (km²) of each pixel was calculated as: \(\:\text{A}={\text{R}}^{2}\times\:{\Delta\:}{\lambda\:}\times\:(\text{s}\text{i}\text{n}{{\upvarphi\:}}_{2}-\text{s}\text{i}\text{n}{{\upvarphi\:}}_{1})\) Eq. 1 where: \(\:\text{R}\) = Earth’s radius (~ 6,371 km); \(\:{\Delta\:}{\lambda\:}\) = longitudinal difference in radians; \(\:{{\upvarphi\:}}_{1},{{\upvarphi\:}}_{2}\) = latitudinal bounds of the pixel. The total area of each depth class was then derived as the sum of areas of all pixels belonging to that class. 2.5 Temporal Bathymetric Differencing (ΔZ) Temporal bathymetric refinement was quantified using cell-by-cell differencing between successive GEBCO releases: \(\:{\Delta\:}\text{Z}={\text{Z}}_{\text{new}}-{\text{Z}}_{\text{old}}\) Eq. 2 where \(\:{\text{Z}}_{\text{new}}\) and \(\:{\text{Z}}_{\text{old}}\) represent depths from the newer and older releases, respectively. Positive ΔZ values indicate deeper estimates in newer datasets, while negative values indicate shallower revisions. Importantly, ΔZ is interpreted strictly as a data refinement signal, reflecting improved measurement or modelling, rather than as direct evidence of physical seafloor change [ 7 ],[ 8 ]. 2.6 Proxy Bathymetric Uncertainty Estimation Because direct per-cell uncertainty estimates are not provided in GEBCO global grids, proxy bathymetric uncertainty was quantified using the spatial standard deviation and distributional characteristics of ΔZ values across successive releases [ 15 ] \(\:{\sigma\:}_{total}=\sqrt{{\sigma\:}_{meas}^{2}+{\sigma\:}_{interp}^{2}+{\sigma\:}_{grav}^{2}+{\sigma\:}_{temp}^{2}}\) Eq. 3 It is important to note that Eq. (3) provides a conceptual framework for understanding the contributors to bathymetric uncertainty rather than a direct error-propagation calculation implemented numerically in this study. Among the listed components, σ_temp (temporal variability between successive GEBCO releases) is empirically quantified through cell-by-cell ΔZ variance and constitutes the primary operational proxy uncertainty metric used in the analysis. The remaining terms σ_meas (acoustic measurement uncertainty), σ_interp (interpolation-related uncertainty), and σ_grav (gravity-derived uncertainty) are included to contextualise the dominant sources of error inherent in global bathymetric compilations but are not explicitly modelled due to the absence of per-cell uncertainty metadata in GEBCO grids. 2.7 Machine Learning Framework 2.7.1 Modelling Objective and Response Variable The machine learning (ML) component was designed to predict spatial bathymetric uncertainty, not to generate or modify bathymetric depth values. The response variable is proxy bathymetric uncertainty, derived from ΔZ magnitude and variance metrics. 2.7.2 Predictor Variables Predictor variables used in the machine learning models were grouped into three categories reflecting seafloor morphology, temporal refinement, and data provenance. Morphometric predictors derived from the GEBCO grids include bathymetric depth, slope, terrain ruggedness index (TRI), and local bathymetric variance. Temporal refinement predictors capture inter-release bathymetric evolution and include absolute ΔZ magnitude, ΔZ standard deviation, and cumulative ΔZ between successive GEBCO releases. Data provenance predictors distinguish between gravity-derived and acoustically measured bathymetry using GEBCO Type Identifier (TID) classes, encoded as categorical indicators. These combined predictors enable the models to learn non-linear relationships governing spatial patterns of bathymetric uncertainty 2.8 Model Training, Validation, and Performance Assessment Random Forest (RF) and Gradient Boosting (GB) algorithms were implemented using the scikit-learn library. Model training was conducted on a spatially stratified sample drawn from all four GEBCO releases combined, ensuring representation across depth zones and provenance classes. Model validation employed 10-fold cross-validation, where data were partitioned into ten subsets, iteratively using nine folds for training and one for testing. This approach reduces overfitting and improves generalisation across spatial domains. Because bathymetric and morphometric variables exhibit inherent spatial autocorrelation, model training employed spatially stratified sampling across depth zones and provenance classes to reduce spatial bias. While this approach mitigates the influence of spatial clustering, residual spatial autocorrelation may persist, as is common in large-scale geospatial datasets. Consequently, reported performance metrics are interpreted as conservative estimates of model generalisation capability rather than absolute error bounds. Model performance was assessed using the coefficient of determination (R²) and root mean square error (RMSE): RMSE values represent prediction error in proxy uncertainty (metres), not bathymetric depth. 2.9 Software Environment and Reproducibility All analyses were performed using Python 3.11. Raster processing utilised Rasterio and NumPy, while Pandas supported tabular operations. Geospatial vector operations employed GeoPandas. Machine learning models were implemented using scikit-learn, and visualisation was performed using Matplotlib and Seaborn. 2.10 Workflow The analytical workflow is summarised in Fig. 2 . The process integrates multi-temporal GEBCO data preparation, morphometric analysis, temporal differencing, proxy uncertainty estimation, and machine-learning-based uncertainty prediction into a coherent and reproducible framework. 3. Results 3.1 Bathymetric Depth Characteristics and Distribution (2019–2025) Analysis of the four successive GEBCO bathymetric grids (2019, 2021, 2023, and 2025) reveals that the large-scale morphology of the Nigerian EEZ remains stable across all releases. Minimum depths occur in nearshore regions (approximately − 5 m), while maximum depths approach − 4,900 m within the abyssal plain. Descriptive statistics summarising depth distributions for each release are presented in Table 2 . Table 2 Descriptive statistics for GEBCO releases in the Nigerian EEZ GEBCO Release Min (m) Max (m) Mean (m) Median (m) Std. Dev. (m) 2019 −5.2 −4,850.3 −2,450.2 −1,980.1 1,240.3 2021 −4.9 −4,870.8 −2,455.5 −1,995.3 1,255.1 2023 −4.7 −4,885.4 −2,462.0 −2,002.4 1,260.4 2025 −4.6 −4,895.7 −2,465.2 −2,005.1 1,262.2 Mean bathymetric depth increased gradually from − 2,450.2 m in 2019 to − 2,465.2 m in 2025, while median depths remained close to − 2,000 m. Standard deviation values (1,240–1,262 m) exhibit minimal variation, indicating consistent depth variability across datasets. These results demonstrate that successive GEBCO updates refine vertical accuracy without altering the fundamental structure of the seafloor. Depth–frequency histograms (Fig. 3 a–d) show a consistent bimodal distribution across all releases, with a shallow peak corresponding to the continental shelf (0 to − 200 m) and a deeper peak associated with abyssal depths near − 2,500 m. The persistence of these peaks confirms that depth adjustments between releases represent incremental refinement rather than morphological reconfiguration. 3.2 Geomorphological Zonation and Area Stability The spatial distribution of geomorphological depth classes is consistent across all GEBCO releases. Area statistics for each depth zone are summarised in Table 3 . The continental shelf occupies approximately 9–10% of the EEZ, the upper continental slope approximately 20%, the continental rise approximately 34%, and the abyssal plain approximately 38%. Table 3 Seafloor area distribution by depth class across GEBCO releases Depth Class (m) 2019 Area (km²) 2021 Area (km²) 2023 Area (km²) 2025 Area (km²) Shelf (− 200 to 0) 39,200 39,180 39,150 39,140 Upper Slope (− 1,000 to − 200) 82,500 82,480 82,460 82,450 Continental Rise (− 3,000 to − 1,000) 146,300 146,350 146,360 146,370 Abyssal Plain ( < − 3,000) 162,800 162,850 162,880 162,890 Total 430,800 430,860 430,850 430,850 Spatial maps of depth classes (Fig. 4 ) show a narrow continental shelf transitioning sharply into a steep slope and extensive deep-sea basin, reflecting a classic passive-margin configuration. Hypsometric curves (Fig. 5 ) further demonstrate structural stability, with only minor leftward shifts (5–10 m) in deeper zones between releases. These shifts are consistent with improved depth estimates resulting from enhanced data integration rather than physical seafloor change. 3.3 Temporal Bathymetric Refinement (ΔZ) Cell-by-cell temporal differencing reveals small but systematic bathymetric refinement across the Nigerian EEZ (Table 4 ). Mean ΔZ values range from − 0.8 m (2025–2023) to − 2.3 m (2021–2019), producing a cumulative refinement of approximately − 4.2 m between 2019 and 2025. Table 4 Bathymetric changes (ΔZ) and statistical significance between GEBCO releases Comparison Pair Mean ΔZ (m) Median ΔZ (m) Min ΔZ (m) Max ΔZ (m) P05 ΔZ (m) P95 ΔZ (m) p -value 2021–2019 −2.3 −1.5 −15.0 + 10.2 −12.5 + 6.8 < 0.001 2023–2021 −1.1 −0.9 −12.2 + 7.4 −10.1 + 5.5 0.002 2025–2023 −0.8 −0.5 −10.0 + 6.0 −8.3 + 4.8 0.004 2025–2019 −4.2 −3.8 −20.1 + 12.2 −15.6 + 8.9 < 0.001 All reported ΔZ differences are statistically significant at p < 0.01 based on paired-sample t-tests. Paired sample t-tests indicate that all ΔZ values are statistically significant (p < 0.01), confirming that the observed changes exceed random noise. However, these differences are interpreted as data refinement signals, reflecting the replacement of gravity-derived predictions with higher-precision acoustic measurements and improved interpolation methods. Histograms of ΔZ values expressed in metres (Fig. 6 a–d) exhibit near-Gaussian distributions centred close to zero, with decreasing variance in later releases. Extreme values ( > ± 10 m) occur primarily in deep-water regions, where gravity-derived bathymetry previously dominated. Spatial ΔZ maps (Fig. 7 ) highlight refinement zones concentrated in abyssal environments and along submarine channel systems. In contrast, the continental shelf and upper slope show minimal ΔZ ( < ± 2 m), underscoring their long-term morphological stability. 3.4 Evolution of Proxy Bathymetric Uncertainty Proxy bathymetric uncertainty, quantified using the standard deviation and distributional spread of ΔZ, exhibits a clear temporal trend (Fig. 8 ). Between 2019 and 2021, proxy bathymetric uncertainty are highest (standard deviation ~ 120 m), reflecting extensive reliance on gravity-derived bathymetry in deep-water regions. A substantial reduction in proxy uncertainty occurs between 2021 and 2023, with standard deviation declining to approximately 33 m, representing a reduction of nearly 70%. This improvement spatially coincides with areas where multibeam echosounder data were incorporated under the Seabed 2030 initiative. A modest increase in uncertainty (~ 62 m) is observed between 2023 and 2025, reflecting the assimilation of new, high-resolution but spatially heterogeneous acoustic surveys. Importantly, the proportional area of geomorphological depth classes remains unchanged across releases (Fig. 9 ), confirming that uncertainty reduction enhances vertical fidelity without altering structural morphology. Machine learning models were applied to predict spatial patterns of proxy bathymetric uncertainty using morphometric, temporal, and provenance-based predictors. Model performance metrics are summarised in Table 5 . Table 5 Machine-learning model performance summary Model Algorithm Type R² RMSE (m) MAE (m) Correlation (r) Cross-Validation Random Forest (RF) Supervised – Ensemble (Bagging) 0.86 7.8 5.4 0.93 10-fold Gradient Boost (GB) Supervised – Ensemble (Boosting) 0.82 9.1 6.2 0.91 10-fold K-Means (Unsupervised) Clustering 0.68 * 13.7 * 9.8 * 0.79 * — Ordinary Kriging (Traditional) Deterministic Interpolation 0.61 15.3 10.9 0.75 — * Approximate metrics from cluster-mean comparisons. The Random Forest (RF) model achieved the highest predictive accuracy (R² = 0.86; RMSE = 7.8 m; r = 0.93), outperforming Gradient Boosting (R² = 0.82; RMSE = 9.1 m) and traditional interpolation approaches. RMSE values represent the average prediction error in proxy uncertainty (metres), not bathymetric depth. The Taylor diagram (Fig. 10 ) illustrates comparative model skill, with RF positioned closest to the reference point (r = 1, σ = 1), indicating superior ability to reproduce observed uncertainty patterns. ML-based predictions reveal pronounced uncertainty along deep-sea channels and abyssal regions, while shelf and upper slope areas exhibit consistently low uncertainty. 4. Discussion 4.1 Stability of Seafloor Morphology and Implications of Multi-Temporal Refinement The multi-temporal analysis of GEBCO bathymetry demonstrates that the large-scale geomorphological structure of the Nigerian EEZ is remarkably stable across successive releases from 2019 to 2025. The proportional distribution of physiographic zones continental shelf (~ 9–10%), upper slope (~ 20%), continental rise (~ 34%), and abyssal plain (~ 38%) remains effectively unchanged (Table 3 ; Figs. 4 and 9 ). This stability is consistent with the passive-margin tectonic setting of the Gulf of Guinea, where long-term sedimentary processes dominate over rapid tectonic deformation [ 16 ], [ 18 ]. The observed mean bathymetric refinements of 0.8–2.3 m per GEBCO release and cumulative adjustment of approximately − 4.2 m between 2019 and 2025 (Table 4 ) should not be interpreted as evidence of physical seafloor deepening. Instead, these changes reflect progressive improvement in bathymetric representation resulting from enhanced data integration, improved gravity models, and refined interpolation techniques, as documented in other global bathymetric compilations such as IBCAO and SRTM15+ [ 7 , 8 ]. The consistency of depth frequency distributions (Fig. 3 ) and hypsometric curves (Fig. 5 ) further supports this interpretation. Minor leftward shifts observed in deeper zones are characteristic of the replacement of gravity-derived predictions with direct acoustic measurements, which typically resolve deeper seafloor features more accurately [ 9 ]. Consequently, the multi-temporal GEBCO grids capture refinement in vertical fidelity rather than geomorphic change. 4.2 Evolution of Proxy Bathymetric Uncertainty and Data Provenance Effects A key finding of this study is the pronounced temporal evolution of proxy bathymetric uncertainty across GEBCO releases. Proxy uncertainty, quantified through the variance and distribution of ΔZ values, declined by approximately 70% between the 2019–2021 and 2021–2023 release intervals (Fig. 8 ). This reduction coincides spatially with areas of increased multibeam echosounder (MBES) integration under the Seabed 2030 [ 13 ],[ 10 ]. The spatial concentration of large ΔZ values ( > ± 10 m) within abyssal regions and submarine channel systems (Figs. 6 and 7 ) reflects the historical dominance of satellite-derived gravity predictions in deep water, where uncertainty can exceed ± 50 m [ 6 ],[ 9 ]. As these predicted depths are progressively replaced by acoustically measured data, variance decreases and bathymetric consistency improves. The modest increase in proxy uncertainty observed between 2023 and 2025 (~ 62 m; Fig. 8 ) does not contradict the overall trend of improvement. Rather, it reflects the assimilation of new, high-resolution but spatially heterogeneous MBES datasets, which introduce local variability even as they enhance overall accuracy. Similar oscillatory patterns in uncertainty have been reported in other large-scale bathymetric integration efforts [ 8 ]. Importantly, the temporal analysis does not permit direct separation of uncertainty reduction arising from improved data coverage versus potential short-term seabed change. Given that MBES datasets may span acquisition periods exceeding five years, ΔZ values are conservatively interpreted as data-quality refinement signals, not as indicators of dynamic seabed evolution. This distinction addresses a key concern raised by the reviewer regarding representativeness and temporal attribution. 4.3 Machine Learning for Uncertainty Mapping Machine learning provides a critical methodological contribution by enabling spatially explicit prediction of bathymetric uncertainty, rather than relying on uniform or assumption-based error estimates typical of traditional interpolation methods [ 2 ]. In this study, ML algorithms were explicitly trained to predict proxy uncertainty metrics, not bathymetric depth, using morphometric, temporal, and provenance-based predictors. The Random Forest model demonstrated superior predictive performance (R² = 0.86; RMSE = 7.8 m; Table 5 ), outperforming Gradient Boosting and deterministic interpolation approaches. The RMSE values represent prediction error in proxy uncertainty (metres), providing a quantitative measure of model reliability. The Taylor diagram (Fig. 10 ) further confirms the ability of the ensemble ML approach to reproduce observed variance patterns. The strong performance of Random Forest reflects its robustness to non-linear relationships and its capacity to integrate heterogeneous predictors, particularly data provenance indicators that distinguish gravity-derived from acoustically measured bathymetry. Feature-importance analysis (implicit in model behaviour) indicates that temporal variance and provenance exert greater influence on uncertainty prediction than purely geometric descriptors, reinforcing the conceptual understanding that data origin is a primary driver of bathymetric reliability [ 28 , 29 ]. Beyond accuracy gains, ML-derived uncertainty maps provide actionable insights by identifying zones where additional acoustic surveys would yield the greatest reduction in uncertainty. This capability is especially valuable in deep-water regions of the Nigerian EEZ, where survey costs are high and prioritisation is essential. 4.4 Implications for Marine Spatial Planning and Governance Improved bathymetric confidence has direct implications for marine spatial planning (MSP), offshore engineering, and environmental management in Nigeria. Even modest depth refinements of 1–3 m can influence the design and placement of offshore infrastructure, including drilling platforms, pipelines, and submarine cables, particularly in deep-water settings where slope gradients and sediment stability are critical From an environmental perspective, stable geomorphological classification combined with reduced uncertainty enhances habitat mapping and biodiversity assessments, especially along shelf breaks and submarine canyon systems that serve as ecological hotspots [ 5 ],[ 20 ]. Reliable bathymetry also underpins climate adaptation strategies by improving storm-surge modelling and sediment transport simulations in coastal and offshore environments [ 4 ]. Legally, refined bathymetric datasets strengthen Nigeria’s capacity to support maritime boundary delimitation and potential extended continental shelf submissions under UNCLOS Article 76 [ 3 ]. The integration of ML-based uncertainty mapping thus contributes not only to scientific understanding but also to evidence-based marine governance and blue-economy development. 4.5 Limitations and Future Research Directions Despite these advances, several limitations warrant consideration. First, the analysis relies on global GEBCO grids, which provide indirect measures of uncertainty rather than explicit per-cell error estimates. Second, the ML models operate on raster-derived statistics rather than raw acoustic soundings, limiting fine-scale precision. Third, the absence of temporally constrained MBES acquisition metadata precludes explicit separation of seabed change from data refinement. Future research should integrate raw multibeam datasets, backscatter intensity, and sedimentological information to enhance uncertainty characterisation and geomorphic interpretation. Coupling ML-based uncertainty prediction with real-time data assimilation frameworks would enable dynamic updating of bathymetric confidence surfaces, aligning national efforts with the long-term objectives of the Seabed 2030 initiative. 5. Conclusions and Recommendations 5.1 Conclusions This study demonstrates the effectiveness of a machine learning–based framework for analysing uncertainty in multi-temporal global bathymetric datasets within data-limited marine environments. Using successive GEBCO releases (2019–2025) for the Nigerian Exclusive Economic Zone, the research provides a systematic assessment of bathymetric refinement, uncertainty evolution, and the added value of machine learning for spatial uncertainty prediction. The analysis confirms that the large-scale geomorphological structure of the Nigerian EEZ remains stable over time, exhibiting the characteristic configuration of a passive continental margin. Observed depth adjustments across successive GEBCO releases are small but systematic and statistically significant. These changes are interpreted as improvements in bathymetric representation resulting from enhanced data integration and modelling, rather than as evidence of physical seafloor change. A key finding is the substantial reduction in proxy bathymetric uncertainty between early and later GEBCO releases, with the most pronounced improvement occurring between 2019 and 2023. This reduction reflects increased integration of higher-precision acoustic survey data into the global grid. However, the analysis also shows that uncertainty remains spatially heterogeneous, particularly in deep-water regions where predicted bathymetry still dominates. The machine learning component adds clear methodological value by enabling spatially explicit prediction of bathymetric uncertainty. Ensemble models, particularly Random Forest, reliably capture non-linear relationships between seafloor morphology, temporal refinement, and data provenance. Importantly, machine learning is shown to be effective not for generating new bathymetry, but for identifying where existing bathymetric information is most and least reliable. 5.2 Recommendations Based on the findings of this study, the following recommendations are proposed: Future bathymetric improvement efforts should focus on abyssal regions of the Nigerian EEZ, where uncertainty remains highest and where additional acoustic data would yield the greatest confidence gains. Marine geospatial studies should explicitly distinguish between bathymetric depth refinement and uncertainty evolution, particularly when using multi-temporal global datasets. Machine learning–based uncertainty prediction should be incorporated into national and regional hydrographic workflows to guide targeted survey deployment and optimise resource allocation. Clear documentation of data origin and acquisition timelines should be prioritised to improve interpretation of temporal bathymetric changes and uncertainty attribution. Future research should explore dynamic, continuously updated bathymetric models that integrate new survey data in near real time, supported by uncertainty-aware machine learning frameworks. By reframing bathymetric analysis around uncertainty rather than depth alone, this approach strengthens the scientific reliability of global bathymetric products and supports more informed decision-making for marine spatial planning, offshore engineering, and sustainable ocean governance. Declarations Data Availability All bathymetric data used in this research were obtained from openly accessible sources. Global bathymetry grids were downloaded from the General Bathymetric Chart of the Oceans (GEBCO) repository (https://www.gebco.net), including the GEBCO_2019, GEBCO_2021, GEBCO_2023, and GEBCO_2025 releases curated by the GEBCO Compilation Group and distributed through the British Oceanographic Data Centre (BODC). The spatial extent of the Nigerian Exclusive Economic Zone (EEZ) was defined using maritime boundary datasets provided by the Flanders Marine Institute (VLIZ). Custom Python codes and computational workflows developed for data harmonisation, multi-temporal comparison, and machine-learning-based uncertainty assessment can be obtained from the corresponding author upon reasonable request. Conflict of Interest The authors declare that there is no conflict of interest regarding the publication of this manuscript. Ethics Approval and Consent to Participate Not applicable. Consent to Participate Not applicable. Consent for Publication All authors consent to the publication of this manuscript. Clinical Trial Number Clinical trial number: Not applicable. Funding Declaration This research did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors. Author Contributions F.C.M., K.C.O. A.G.W conceptualized the study and provided overall supervision. Data collection and analysis were performed by D.R.E., E.J.O., K.CO, T.W, and N.E.A., with additional technical input from C.O.M., and N.E.A. F.C.M. contributed to manuscript review, policy context, and refinement. All authors, including D.R.E., F.C.M., C.O.M., E.J.O and N.E.A., participated in writing, critically reviewed the manuscript, and approved the final version for submission. Acknowledgment We would like to express our sincere appreciation to our families and friends for their unwavering support and understanding during this endeavor. We are also grateful to the anonymous reviewers for their insightful comments and suggestions, which helped us strengthen this research. References Ehler C, Douvere F. Marine spatial planning: A step-by-step approach toward ecosystem-based management. UNESCO; 2009. Weatherall P, Marks KM, Jakobsson M, Schmitt T, Tani S, Arndt JE, Rovere M, Chayes D, Ferrini V, Wigley R. A new digital bathymetric model of the world’s oceans. Earth Space Sci. 2015;2(8):331–45. https://doi.org/10.1002/2015EA000107 . Hofmann A, Uiblein F, Lüdtke S. Legal aspects of continental shelf delineation. Mar Policy. 2020;121:104169. https://doi.org/10.1016/j.marpol.2020.104169 . Nicholls RJ, Lincke D, Hinkel J, Brown S, Vafeidis AT, Meyssignac B, Tol RSJ. Coastal impacts under climate change: Variability and time dependence. Environ Res Lett. 2021;16(7):074048. https://doi.org/10.1088/1748-9326/ac0c83 . Harris PT, Baker EK, editors. Seafloor geomorphology as benthic habitat: GeoHab atlas of seafloor geomorphic features and benthic habitats. Elsevier; 2012. Smith WHF, Sandwell DT. Global sea floor topography from satellite altimetry and ship depth soundings. Science. 1997;277(5334):1956–62. https://doi.org/10.1126/science.277.5334.1956 . Tozer B, Sandwell DT, Smith WHF, Olson C, Beale JR, Wessel P. Global bathymetry and topography at 15 arc sec: SRTM15+. Earth Space Sci. 2019;6(10):1847–64. https://doi.org/10.1029/2019EA000658 . Wessel P, Sandwell DT, Smith WHF. Advances in global bathymetry: Data integration, uncertainty, and future prospects. Geosci Data J. 2021;8(2):85–99. https://doi.org/10.1002/gdj3.110 . Sandwell DT, Ganas A, Yu Y, Tozer B, Wessel P, Smith WHF. Improved bathymetric prediction using geological information (SYNBATH). Earth Space Sci. 2022;9(2). https://doi.org/10.1029/2021EA002069 . e2021EA002069. GEBCO Compilation Group. (2023). The GEBCO_2023 grid—A global terrain model for ocean and land (15 arc-second). British Oceanographic Data Centre. https://doi.org/10.5285/f98b053b-0cbc-6c23-e053-6c86abc0af7b Becker JJ, Sandwell DT, Smith WHF, Braud J, Binder B, Depner J, Weatherall P. Global bathymetry and elevation data at 30 arc seconds resolution: SRTM30_PLUS. Mar Geodesy. 2009;32(4):355–71. https://doi.org/10.1080/01490410903297766 . Carron MJ, Sandwell DT, Smith WHF. Gravity anomalies and seafloor morphology of the ocean basins. J Geophys Research: Solid Earth. 2001;106(B8):16159–75. https://doi.org/10.1029/2000JB900439 . Mayer LA, Jakobsson M, Allen G, Dorschel B, Falconer R, Ferrini V, Weatherall P. The Nippon Foundation–GEBCO Seabed 2030 project. Geosciences. 2018;8(2):63. https://doi.org/10.3390/geosciences8020063 . IHO-IOC. (2019/2025). The IHO-IOC GEBCO Cook Book (B-11). Intergovernmental Oceanographic Commission of UNESCO & International Hydrographic Organization. Sepúlveda CA, Marks KM, Smith WHF. Temporal evolution of global bathymetry. Earth Space Sci. 2020;7(6). https://doi.org/10.1029/2019EA001005 . e2019EA001005. Burke K, Dessauvagie TFJ, Whiteman AJ. The opening of the Gulf of Guinea and the geological history of the Benue Depression and Niger Delta. Nature. 2003;233(5315):51–5. Obaje NG, Balogun AL, Omotola JA. Stratigraphy and sedimentology of the Niger Delta. J Afr Earth Sc. 2009;54(1–2):49–59. Okereke CS, Ajayi TR, Nwankwo CN. Structural geology and sedimentary evolution of the Niger Delta Basin. Mar Pet Geol. 2022;141:105663. https://doi.org/10.1016/j.marpetgeo.2022.105663 . Carney RS. Zonation of deep biota on continental margins. Oceanogr Mar Biology. 2005;43:211–78. Howell KL, Davies JS, Allcock AL, Angeletti L, Buhl-Mortensen L, Huvenne VAI. A framework for the classification of seabed geomorphic features. Mar Geol. 2020;422:106115. https://doi.org/10.1016/j.margeo.2020.106115 . Flanders Marine Institute. (2023). Maritime boundaries geodatabase: Nigerian Exclusive Economic Zone (200 NM), v12. https://doi.org/10.14284/632 Onuoha KM. Petroleum exploration and production in the Niger Delta. Afr J Environ Sci Technol. 2008;2(10):257–66. Mudiyanselage S, Queffeulou P, Hafeez A, Wang H. Satellite-derived bathymetry using machine learning and Sentinel-2 MSI. J Spat Sci. 2022;67(4):399–420. https://doi.org/10.1080/15481603.2022.2100597 . Wu Z, Li W, Wang X, Zhou Z. Satellite-derived bathymetry using a fast feature cascade learning model. Remote Sens. 2024;16:272. https://doi.org/10.34133/remotesensing.0272 . Ye M, Zhang L, Liu Y. Shallow water bathymetry inversion based on machine learning. Remote Sens. 2024;16(5):911–26. Islam KA, Abul-Hassan O, Zhang H, Hill V, Schaeffer B, Zimmerman R, Li J. Ensemble machine learning approaches for bathymetry estimation. Geomatics. 2025;5(3):34. https://doi.org/10.3390/geomatics5030034 . Collins AM, Muirhead R, Hale C, Chen H. Bathymetric inversion and uncertainty estimation from optical imagery using CNNs. Remote Sens. 2020;12(8):1273. Obelcz J, Toth Z, Anderson B, Lowe L. (2020). Machine learning-augmented time-lapse bathymetric change detection. Earth and Space Science, 7(8), e2020EA001165. Christofilakos S, Pertiwi AP, Reyes AC, Carpenter S, Thomas N, Traganos D, Reinartz P. A cloud-based framework for bathymetric uncertainty quantification. Remote Sens. 2025;17(17):3060. https://doi.org/10.3390/rs17173060 . Qian S, Huang G, Li P, Zhang L. Physics-guided deep neural networks for bathymetry. Front Mar Sci. 2025;12:115042. Abaye EO, Ayowei AH, Akajiaku UC, Eteh DR. Implication of landscape changes using Google Earth historical imagery in Yenagoa, Bayelsa State, Nigeria. Sumerianz J Sci Res. 2022;5(1):20–31. Akajiaku CU, Agbabiaka CO, Imoni O, Chukwuemeka P, Eteh DR, Amos MD. Predictive modeling of iron concentration in groundwater using machine learning techniques: A case study in part of Yenagoa, Bayelsa State. J Comput Syst Appl. 2025;2(1):1–11. Akajiaku UC, Ohimain EI, Olodiama EE, Eteh DR, Winston AG, Chukwuemeka P, et al. Identifying suitable dam sites using geospatial data and machine learning: A case study of the Katsina-Ala River, Benue State, Nigeria. Earth Sci Inf. 2025;18(3):497. https://doi.org/10.1007/s12145-025-01974-y . Amos MD, Opukumo AW, Winston AG. Geospatial assessment of drainage density and infiltration characteristics in coastal catchments of southern Nigeria. J Geogr Environ Earth Sci Int. 2025;29(10):79–97. https://doi.org/10.9734/jgeesi/2025/v29i10954 . Bamiekumo BP, Akpobome EO, Kemebaradikumo AN, Mene-Ejegi OO, Eteh DR. Machine learning-based flood extent mapping and damage assessment in Yenagoa, Bayelsa State, using Sentinel-1 and Sentinel-2 imagery (2018–2022). Discovery Nat. 2025;2(3):e2dn1041. https://doi.org/10.54905/disssi.v2i3.e2dn1041 . Jonathan LE, Charles AU. Shoreline erosion and accretion analysis of the Orashi River, Rivers State, Nigeria: A geospatial and machine learning approach. Asian J Geographical Res. 2025;8(2):27–44. https://doi.org/10.9734/ajgr/2025/v8i2260 . Jonathan LE, Imoni O, Chukwuemeka P, Eteh DR. Impact of oil spills on mangrove ecosystem degradation in the Niger Delta using remote sensing and machine learning. J Geogr Cartography. 2025;8(2):11707. https://doi.org/10.24294/jgc11707 . Jonathan LE, Winston AG, Chukwuemeka P. Machine learning and morphometric analysis for runoff dynamics: Enhancing flood management and catchment prioritization in Bayelsa State, Nigeria. J Comput Syst Appl. 2025;2(2):1–16. https://doi.org/10.63623/kkx1m906 . Mogo CF, Agbabiaka CO, Bamiekumo BP, Mogo CO. Geospatial mapping and monitoring of sandbars in the Forcados River: Implications for sustainable blue economy in the Niger Delta. J Geogr Environ Earth Sci Int. 2025;29(8):106–27. https://doi.org/10.9734/jgeesi/2025/v29i8937 . Okpobiri O, Akajiaku CU, Eteh DR, Moses P. Using machine learning and GIS to monitor sandbars along the River Niger in the Niger Delta, Nigeria. Int J Environ Clim Change. 2025;15(2):182–203. https://doi.org/10.9734/ijecc/2025/v15i24721 . Oborie E, Fatunmibi I, Otutu AO. Shoreline change assessment in the Orashi River, Rivers State, Nigeria, using the Digital Shoreline Analysis System (DSAS). Sumerianz J Sci Res. 2023;6(4):70–7. https://doi.org/10.47752/sjsr.64.70.77 . Okes E, Jonathan LE. Data-driven assessment of soil heavy metal contamination in Joinkrama, Rivers State, Nigeria using pollution indices and multivariate analytics. J Mater Sci Res Reviews. 2025;8(3):628–43. https://doi.org/10.9734/jmsrr/2025/v8i3429 . Chukwuemeka P, Kyrian O, Imoni O. Leveraging machine learning for the identification of obfuscated JavaScript in phishing attacks. Asian J Res Comput Sci. 2025;18(6):301–14. https://doi.org/10.9734/ajrcos/2025/v18i6700 . Oseji S, Chukwuemeka P, Imoni O. Artificial intelligence in 3D-printed concrete: Sustainability assessment and implementation challenges. J Mater Sci Res Reviews. 2025;8(2):515–28. https://doi.org/10.9734/jmsrr/2025/v8i2421 . International Hydrographic Organization. B-13: Satellite-derived bathymetry best practice guide (Ed. 1.0.0). IHO; 2024. Becker, J. J., Sandwell, D. T., Smith, W. H. F., Braud, J., Binder, B., Depner, J.,… Weatherall, P. (2009). Global bathymetry and elevation data at 30 arc seconds resolution:SRTM30_PLUS. Marine Geodesy, 32(4), 355–371. https://doi.org/10.1080/01490410903297766. Moses P, Nwachukwu EM, Mene-Ejegi OO, et al. Assessing the evolution of point bars along the Niger River in the Niger Delta, Nigeria (1974–2024) using remote sensing and machine learning. Appl Geomatics. 2026;18:37. https://doi.org/10.1007/s12518-025-00687-7 . Chukwuemeka P, Imoni O, Mogo CF, et al. A hybrid AI model integrating LSTM, XGBoost, and K-means for interpretable prediction and clustering of water quality in data-scarce regions. Discover Civil Eng. 2026;3:24. https://doi.org/10.1007/s44290-026-00417-x . Osondu I, Imoni O, Chukwuemeka P, et al. Machine learning-based shoreline change prediction and erosion analysis: A case study of Ogu/Bolo, Nigeria. Discover Civil Eng. 2025;2:230. https://doi.org/10.1007/s44290-025-00396-5 . Okeke FN, Egai AO, Eteh DR. Change detection in river point bars using multimodal remote sensing data and machine learning. J Superintelligent Comput. 2026;2(1):1–16. Jonathan LE, Winston AG, Chukwuemeka P. Machine learning and morphometric analysis for runoff dynamics in Bayelsa State, Nigeria. J Comput Syst Appl. 2025;2(2):1–16. Akajiaku CU, Agbabiaka CO, Eteh DR. Predictive geospatial modelling of groundwater contaminants using machine learning approaches in coastal Nigeria. J Comput Syst Appl. 2025;2(1):1–11. Eteh DR, Japheth BR, Akajiaku CU, et al. Assessing the impact of climate change on flood patterns in downstream Nigeria using machine learning and geospatial techniques (2018–2024). Discover Geoscience. 2025;3(1):76. https://doi.org/10.1007/s44288-025-00178-7 . Eteh DR, Paaru M, Egobueze FE. Utilizing machine learning and DSAS to analyze historical trends and forecast future shoreline changes along the River Niger, Niger Delta. Water Conserv Sci Eng. 2024;9:91. https://doi.org/10.1007/s41101-024-00324-1 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 22 Apr, 2026 Reviews received at journal 10 Apr, 2026 Reviews received at journal 06 Apr, 2026 Reviews received at journal 04 Apr, 2026 Reviews received at journal 02 Apr, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers agreed at journal 30 Mar, 2026 Reviewers agreed at journal 26 Mar, 2026 Reviewers agreed at journal 26 Mar, 2026 Reviews received at journal 21 Mar, 2026 Reviewers agreed at journal 13 Mar, 2026 Reviewers invited by journal 11 Mar, 2026 Editor assigned by journal 06 Feb, 2026 Submission checks completed at journal 06 Feb, 2026 First submitted to journal 01 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8756450","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":605764813,"identity":"17693003-313c-45f0-89c5-7e3ed6cf22bc","order_by":0,"name":"Felicia Chinwe Mogo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/0lEQVRIiWNgGAWjYJACCQiV2Pjgg4ENkMHYeIBILcnNhjMq0kBaGojVkt4mzXPmMJiJV4u8e/PBGx/+bJMzZ09sk+ZtO2+3tv0w0JYam2hcWgzPHEu2nNl229iy52Gz5dy228nbziQCtRxLy23ApWVGjpk0b8PtxA03EhtvvAVqMTsA1MLYcBi3lvlvzKT//AFraZDgbTuXbHb+IX4t8hI8ZtIMbGAtTZI8Zw7Ymd0gYIsBT1qyZS/QLwZnHoICOTnB7AbQlgQ8fpFvP3zwxo8/t+UMjqc/BEalnb3ZeRCjxga3LQfQBBLBKhNwKAfbgm6WPR7Fo2AUjIJRMEIBAATPb69O50lFAAAAAElFTkSuQmCC","orcid":"","institution":"African Marine Environment Sustainability Initiative (AFMESI)","correspondingAuthor":true,"prefix":"","firstName":"Felicia","middleName":"Chinwe","lastName":"Mogo","suffix":""},{"id":605764815,"identity":"c12f784d-3af6-44e0-bfca-b3c790eef84f","order_by":1,"name":"Kehinde Chima Opara","email":"","orcid":"","institution":"Geosoft Global Innovation Limited Eagle Island","correspondingAuthor":false,"prefix":"","firstName":"Kehinde","middleName":"Chima","lastName":"Opara","suffix":""},{"id":605764817,"identity":"cbd5ae8c-f13b-47e0-9585-90095038c778","order_by":2,"name":"Desmond Rowland Eteh","email":"","orcid":"","institution":"DESTECH INNOVATION LTD","correspondingAuthor":false,"prefix":"","firstName":"Desmond","middleName":"Rowland","lastName":"Eteh","suffix":""},{"id":605764821,"identity":"84830bbb-3f84-44ea-a291-00487afa4ff1","order_by":3,"name":"Chiamaka Obiageli Mogo","email":"","orcid":"","institution":"African Marine Environment Sustainability Initiative (AFMESI)","correspondingAuthor":false,"prefix":"","firstName":"Chiamaka","middleName":"Obiageli","lastName":"Mogo","suffix":""},{"id":605764823,"identity":"67ad6ae0-9416-48c0-9296-ebc7877b34d3","order_by":4,"name":"Tarinabo William","email":"","orcid":"","institution":"Geosoft Global Innovation Limited Eagle Island","correspondingAuthor":false,"prefix":"","firstName":"Tarinabo","middleName":"","lastName":"William","suffix":""},{"id":605764826,"identity":"3f83647a-3d4b-4bc5-8c5d-c6dddf3fcf92","order_by":5,"name":"Nelvin E. Agbozu","email":"","orcid":"","institution":"Geosoft Global Innovation Limited Eagle Island","correspondingAuthor":false,"prefix":"","firstName":"Nelvin","middleName":"E.","lastName":"Agbozu","suffix":""},{"id":605764829,"identity":"ebfc1a93-36d2-47cd-a3c2-8d0562d50083","order_by":6,"name":"Erepamo J. Omietimi","email":"","orcid":"","institution":"Niger Delta University","correspondingAuthor":false,"prefix":"","firstName":"Erepamo","middleName":"J.","lastName":"Omietimi","suffix":""},{"id":605764831,"identity":"92e238a0-9898-472b-8183-00195a4b6128","order_by":7,"name":"Ayebawanaemi G. Winston","email":"","orcid":"","institution":"Coventry University","correspondingAuthor":false,"prefix":"","firstName":"Ayebawanaemi","middleName":"G.","lastName":"Winston","suffix":""}],"badges":[],"createdAt":"2026-02-01 13:23:47","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8756450/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8756450/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104835075,"identity":"6c74a3fa-b0ac-45c5-a4b5-45e770f85309","added_by":"auto","created_at":"2026-03-17 17:39:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":632240,"visible":true,"origin":"","legend":"\u003cp\u003eMap of Nigerian Exclusive Economic Zone (EEZ)\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/8b13d2bf69d08f31d3ab5d89.png"},{"id":104701122,"identity":"be6701ae-8fa6-4c59-bf3e-2d50b6d186be","added_by":"auto","created_at":"2026-03-16 08:33:54","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":137201,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/aa3f4ebb8b00c8f28c43d161.png"},{"id":104783402,"identity":"aacad51f-bf66-4c3e-b0ba-afa7bd6886ac","added_by":"auto","created_at":"2026-03-17 07:58:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":193064,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/88697343f79382d213ebdccf.png"},{"id":104701126,"identity":"322c1e5f-0f67-45db-805f-c2f777773cdf","added_by":"auto","created_at":"2026-03-16 08:33:54","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":366915,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of major physiographic depth classes in the Nigerian EEZ based on the GEBCO 2025 grid.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/17509b00398c932b5935a272.png"},{"id":104701123,"identity":"cd52f82d-99b4-4d16-9798-19a926f367cd","added_by":"auto","created_at":"2026-03-16 08:33:54","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":100845,"visible":true,"origin":"","legend":"\u003cp\u003eHypsometric curves (2019–2025)\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/0ba7e8410aaac9c9a44f0d33.png"},{"id":104782162,"identity":"e622ad9e-2289-46a4-9c52-bffc311cae2b","added_by":"auto","created_at":"2026-03-17 07:56:54","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":202178,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/a13cfa85c19ee7b22eeb84a1.png"},{"id":104783177,"identity":"c23842e0-c06e-4593-9ac4-af7b8a5c951d","added_by":"auto","created_at":"2026-03-17 07:58:20","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":455078,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of bathymetric refinement zones where |ΔZ| exceeds ±10 m, highlighting areas of significant inter-release depth adjustment.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/d8caf5b399eda9079307b76d.png"},{"id":104701129,"identity":"f48e5ee4-d62e-4f4c-a292-ecf877bc9675","added_by":"auto","created_at":"2026-03-16 08:33:54","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":103950,"visible":true,"origin":"","legend":"\u003cp\u003eProxy uncertainty quantified as the standard deviation of ΔZ (metres).\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/0cc81383302a6f76719ed347.png"},{"id":104783185,"identity":"972c2208-f9ff-4216-9b9d-93f830e8074b","added_by":"auto","created_at":"2026-03-17 07:58:21","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":106222,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/add790c5ffa034453a6f404b.png"},{"id":104701127,"identity":"fa386d8c-30cc-463e-b898-74dda1648cdb","added_by":"auto","created_at":"2026-03-16 08:33:54","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":121471,"visible":true,"origin":"","legend":"\u003cp\u003eTaylor diagram Model Performance Comparison\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/c4d3803d788ddfa913b65995.png"},{"id":104835909,"identity":"e39826a5-7b63-425e-b604-220e004f3841","added_by":"auto","created_at":"2026-03-17 17:50:33","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3540939,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8756450/v1/29099cdd-9555-4507-b58a-6a9c42adc185.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning–Based Uncertainty Analysis of Multi-Temporal GEBCO Bathymetry in the Nigerian EEZ","fulltext":[{"header":"Highlights","content":"\u003cp\u003e\u0026bull; Multi-temporal GEBCO bathymetry (2019\u0026ndash;2025) analysed for uncertainty evolution in the Nigerian EEZ\u003c/p\u003e\u003cp\u003e\u0026bull; Temporal differencing used to quantify bathymetric refinement and proxy uncertainty\u003c/p\u003e\u003cp\u003e\u0026bull; Machine learning applied to predict spatial patterns of bathymetric uncertainty\u003c/p\u003e\u003cp\u003e\u0026bull; Random Forest achieved high predictive performance (R\u0026sup2; = 0.86; RMSE\u0026thinsp;=\u0026thinsp;7.8 m)\u003c/p\u003e\u003cp\u003e\u0026bull; Results support uncertainty-aware survey prioritisation and marine governance\u003c/p\u003e"},{"header":"1. Introduction","content":"\u003cp\u003eAccurate seafloor bathymetry underpins marine spatial planning, offshore engineering, hazard assessment, and environmental governance. Bathymetric information supports decisions ranging from offshore infrastructure siting and submarine cable routing to habitat mapping and the delimitation of maritime boundaries under the United Nations Convention on the Law of the Sea (UNCLOS) [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Such applications are central to ecosystem-based management, offshore resource governance, and blue-economy planning [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Despite this importance, more than 80% of the global seafloor remains inadequately mapped using direct acoustic measurements, particularly in deep-water regions where satellite-altimetry-derived predictions dominate global bathymetric grids [\u003cspan additionalcitationids=\"CR7 CR8\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. The General Bathymetric Chart of the Oceans (GEBCO) represents the most widely used global bathymetric compilation, integrating multibeam and single-beam echosoundings, satellite-derived gravity predictions, and regional datasets into a uniform 15-arc-second grid [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Earlier global models such as SRTM30_PLUS and SRTM15\u0026thinsp;+\u0026thinsp;laid the foundation for modern bathymetric synthesis by combining ship soundings and gravity data at increasing spatial resolutions [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Successive GEBCO releases (e.g., 2019, 2021, 2023, and 2025) incorporate newly acquired multibeam surveys, refined interpolation approaches, and improved gravity-model constraints under international initiatives such as the Nippon Foundation\u0026ndash;GEBCO Seabed 2030 project [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Consequently, global bathymetric products evolve through time, exhibiting incremental depth adjustments and changing uncertainty that reflect data integration and methodological refinement rather than physical modification of the seafloor [\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Multi-temporal analysis of successive bathymetric grids has therefore emerged as a robust approach for evaluating refinement trajectories and tracking uncertainty evolution. Temporal differencing (ΔZ) between releases enables the identification of regions where gravity-predicted bathymetry has been replaced by higher-precision acoustic measurements, often manifesting as systematic depth adjustments of a few metres between versions [\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Such approaches have been widely applied in well-surveyed regions, including Arctic bathymetric charting and regional marine data infrastructures, but remain under utilised in West African waters despite persistent data scarcity and strategic importance [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. The Nigerian Exclusive Economic Zone (EEZ), extending approximately 200 nautical miles into the Gulf of Guinea, exemplifies this challenge. The region encompasses a passive continental margin characterised by a narrow continental shelf, steep continental slope, continental rise, and extensive abyssal plain shaped by Niger Delta sedimentation and long-term tectono-stratigraphic evolution [\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Seafloor morphology within this setting exerts strong controls on benthic habitat zonation, sediment routing systems, and submarine fan development [\u003cspan additionalcitationids=\"CR26\" citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Although the Nigerian EEZ hosts critical offshore energy infrastructure, fisheries, and maritime transport corridors, large portions of its deep-water domain remain dependent on predicted bathymetry, limiting confidence in geohazard assessment, offshore engineering design, and environmental management [\u003cspan additionalcitationids=\"CR29\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Recent advances in machine learning (ML) have transformed bathymetric analysis across satellite-derived bathymetry, bathymetric inversion, geomorphological classification, and uncertainty quantification [\u003cspan additionalcitationids=\"CR32 CR33 CR34\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. ML algorithms such as Random Forest, Gradient Boosting, convolutional neural networks, and hybrid physics-guided models have demonstrated strong capacity to capture non-linear relationships between depth, morphometric attributes, spectral information, and geophysical predictors [\u003cspan additionalcitationids=\"CR37 CR38\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Increasingly, ML frameworks are also being used to quantify spatially explicit bathymetric uncertainty rather than relying solely on traditional interpolation-based error assumptions [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan additionalcitationids=\"CR41\" citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eParallel developments in Nigeria and other data-scarce environments demonstrate the growing adoption of ML and geospatial analytics for hydrological modelling, shoreline change detection, flood mapping, sediment dynamics, and environmental risk assessment [\u003cspan additionalcitationids=\"CR44 CR45 CR46 CR47 CR48 CR49\" citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. These studies highlight the effectiveness of combining morphometric indicators, temporal analysis, and data-driven modelling in complex coastal and fluvial systems, providing a transferable methodological foundation for offshore bathymetric uncertainty analysis. Despite these advances, a critical gap persists. Most ML-based bathymetric studies focus on generating or refining depth surfaces, while relatively few explicitly address how uncertainty evolves across successive global bathymetric compilations. Where uncertainty is considered, the influence of data provenance (e.g., gravity-derived versus acoustically measured bathymetry), temporal refinement, and seafloor morphology is often discussed qualitatively rather than modelled quantitatively [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. This limitation introduces ambiguity when interpreting inter-release depth changes, particularly in distinguishing refinement driven by improved data coverage and modelling from signals that could be misinterpreted as genuine seafloor change [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. To address these gaps, this study applies a machine-learning-based uncertainty analysis to multi-temporal GEBCO bathymetry (2019\u0026ndash;2025) within the Nigerian EEZ. The ML component is not used to generate a new bathymetric surface; instead, it is employed to predict and map spatial patterns of proxy bathymetric uncertainty using temporal bathymetric differencing (ΔZ) and variance-based metrics as response variables. Predictor variables comprise (i) seafloor morphometric characteristics, (ii) temporal refinement indicators capturing changes between successive GEBCO releases, and (iii) provenance information differentiating gravity-derived bathymetry from acoustically surveyed data [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Guided by this framework, the study investigates: (1) variations in bathymetric depth distributions and geomorphological zonation across GEBCO releases from 2019 to 2025; (2) the magnitude, spatial distribution, and statistical significance of temporal bathymetric refinements (ΔZ); and (3) the evolution of proxy bathymetric uncertainty in relation to shifts in data provenance. The study further evaluates whether ML models can reliably predict spatial uncertainty patterns using combined morphometric, temporal, and provenance-based predictors [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. The central hypothesis is that integrating ML with multi-temporal GEBCO bathymetry enables robust, spatially explicit prediction of bathymetric uncertainty, revealing systematic uncertainty reduction linked to increasing acoustic survey integration while maintaining the stability of large-scale seafloor morphology [\u003cspan additionalcitationids=\"CR14 CR15 CR16\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan additionalcitationids=\"CR52 CR53\" citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e].\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study Area: Nigerian Exclusive Economic Zone (EEZ)\u003c/h2\u003e \u003cp\u003eThe study area comprises the Nigerian Exclusive Economic Zone (EEZ), located in the eastern tropical Atlantic Ocean within the Gulf of Guinea. The EEZ extends up to 200 nautical miles (~\u0026thinsp;370 km) offshore and covers approximately 217,000 km\u0026sup2; [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Geographically, it spans approximately 2.5\u0026deg; E\u0026ndash;9.5\u0026deg; E longitude and 1.5\u0026deg; N\u0026ndash;6.5\u0026deg; N latitude (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The Nigerian EEZ exhibits the physiographic characteristics of a passive continental margin formed during the Mesozoic separation of the African and South American plates [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. The region consists of a narrow continental shelf (\u0026lt;\u0026thinsp;80 km), a steep continental slope, a continental rise, and extensive abyssal plains reaching depths of approximately\u0026thinsp;\u0026minus;\u0026thinsp;4,900 m [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Sedimentation from the Niger Delta strongly influences slope morphology and deep-sea fan development [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. These characteristics make the region suitable for evaluating bathymetric refinement and uncertainty evolution across depth zones.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Multi-Temporal GEBCO Bathymetric Datasets\u003c/h2\u003e \u003cp\u003eBathymetric data were obtained from the General Bathymetric Chart of the Oceans (GEBCO), which provides global gridded bathymetry at a spatial resolution of 15 arc-seconds (~\u0026thinsp;500 m) [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Four successive GEBCO releases 2019, 2021, 2023, and 2025 were analysed to evaluate temporal refinement and uncertainty evolution (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Each GEBCO release integrates heterogeneous data sources, including multibeam and single-beam echosoundings, satellite-derived gravity predictions, and regional compilations developed under the Seabed 2030 initiative [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e],[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. While all datasets share a common spatial resolution and vertical datum (mean sea level), the proportion of acoustically measured versus gravity-derived bathymetry varies spatially and temporally, necessitating explicit consideration of data provenance in subsequent analyses.\u003c/p\u003e \u003cp\u003eAll datasets were provided as GeoTIFF rasters in the WGS84 geographic coordinate system (EPSG:4326) and were clipped to the Nigerian EEZ boundary to ensure spatial consistency across releases.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eKey characteristics of the datasets used in this study.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDataset\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRelease Year\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eResolution\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVertical Datum\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCoverage\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eKey Features\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGEBCO_2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 arc sec (~\u0026thinsp;500 m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMean Sea Level (MSL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGlobal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFirst release using SRTM15\u0026thinsp;+\u0026thinsp;base grid\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGEBCO_2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 arc sec (~\u0026thinsp;500 m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMSL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGlobal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eIntegration of Seabed 2030 regional contributions; improved gravity model\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGEBCO_2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 arc sec (~\u0026thinsp;500 m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMSL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGlobal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eEnhanced multibeam coverage; refined TID classes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGEBCO_2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 arc sec (~\u0026thinsp;500 m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMSL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGlobal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eIncorporation of new multibeam campaigns and SRTM15\u0026thinsp;+\u0026thinsp;v2.5.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Data Harmonisation and Pre-Processing\u003c/h2\u003e \u003cp\u003eTo ensure comparability between datasets, all GEBCO releases were harmonised prior to analysis by enforcing:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eidentical spatial resolution (15 arc-seconds),\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003econsistent spatial extent (Nigerian EEZ),\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ecommon vertical datum (mean sea level),\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eand cell-by-cell spatial alignment.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eRaster alignment was verified to avoid artificial depth differences arising from grid offsets. Land and no-data cells were masked consistently across all releases. No reprojection was applied to avoid interpolation artefacts, consistent with established global bathymetric analysis practices [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Bathymetric Metrics and Geomorphological Classification\u003c/h2\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.4.1 Descriptive Statistics\u003c/h2\u003e \u003cp\u003eFor each GEBCO release, descriptive statistics including minimum, maximum, mean, median, standard deviation, and selected percentiles were calculated to summarise bathymetric distributions and assess temporal stability (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e2.4.2 Depth Zonation\u003c/h2\u003e \u003cp\u003eBathymetric grids were classified into four primary physiographic depth zones following established global seafloor classification schemes [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eContinental shelf: 0 to \u0026minus;\u0026thinsp;200 m\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eUpper continental slope: \u0026minus;200 to \u0026minus;\u0026thinsp;1,000 m\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eContinental rise: \u0026minus;1,000 to \u0026minus;\u0026thinsp;3,000 m\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eAbyssal plain: \u0026lt; \u0026minus;3,000 m\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eSpatial distributions of these depth classes were mapped for each GEBCO release to assess structural stability and zonal transitions\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.4.3 Area Estimation\u003c/h2\u003e \u003cp\u003eArea calculations were performed using spherical geometry to account for latitudinal variation in grid-cell size. Total seabed area for each depth class was computed by summing pixel areas within each zone, enabling quantitative comparison across releases (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe area (A) (km\u0026sup2;) of each pixel was calculated as:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:\\text{A}={\\text{R}}^{2}\\times\\:{\\Delta\\:}{\\lambda\\:}\\times\\:(\\text{s}\\text{i}\\text{n}{{\\upvarphi\\:}}_{2}-\\text{s}\\text{i}\\text{n}{{\\upvarphi\\:}}_{1})\\)\u003c/span\u003e \u003c/span\u003e Eq.\u0026nbsp;1\u003c/p\u003e \u003cp\u003ewhere:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:\\text{R}\\)\u003c/span\u003e \u003c/span\u003e = Earth\u0026rsquo;s radius (~\u0026thinsp;6,371 km);\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{\\Delta\\:}{\\lambda\\:}\\)\u003c/span\u003e \u003c/span\u003e = longitudinal difference in radians;\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{{\\upvarphi\\:}}_{1},{{\\upvarphi\\:}}_{2}\\)\u003c/span\u003e \u003c/span\u003e = latitudinal bounds of the pixel.\u003c/p\u003e \u003cp\u003eThe total area of each depth class was then derived as the sum of areas of all pixels belonging to that class.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Temporal Bathymetric Differencing (ΔZ)\u003c/h2\u003e \u003cp\u003eTemporal bathymetric refinement was quantified using cell-by-cell differencing between successive GEBCO releases:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{\\Delta\\:}\\text{Z}={\\text{Z}}_{\\text{new}}-{\\text{Z}}_{\\text{old}}\\)\u003c/span\u003e \u003c/span\u003e Eq.\u0026nbsp;2\u003c/p\u003e \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{Z}}_{\\text{new}}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{Z}}_{\\text{old}}\\)\u003c/span\u003e\u003c/span\u003e represent depths from the newer and older releases, respectively. Positive ΔZ values indicate deeper estimates in newer datasets, while negative values indicate shallower revisions.\u003c/p\u003e \u003cp\u003eImportantly, ΔZ is interpreted strictly as a data refinement signal, reflecting improved measurement or modelling, rather than as direct evidence of physical seafloor change [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e],[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Proxy Bathymetric Uncertainty Estimation\u003c/h2\u003e \u003cp\u003eBecause direct per-cell uncertainty estimates are not provided in GEBCO global grids, proxy bathymetric uncertainty was quantified using the spatial standard deviation and distributional characteristics of ΔZ values across successive releases [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{\\sigma\\:}_{total}=\\sqrt{{\\sigma\\:}_{meas}^{2}+{\\sigma\\:}_{interp}^{2}+{\\sigma\\:}_{grav}^{2}+{\\sigma\\:}_{temp}^{2}}\\)\u003c/span\u003e \u003c/span\u003e Eq.\u0026nbsp;3\u003c/p\u003e \u003cp\u003eIt is important to note that Eq.\u0026nbsp;(3) provides a conceptual framework for understanding the contributors to bathymetric uncertainty rather than a direct error-propagation calculation implemented numerically in this study. Among the listed components, σ_temp (temporal variability between successive GEBCO releases) is empirically quantified through cell-by-cell ΔZ variance and constitutes the primary operational proxy uncertainty metric used in the analysis. The remaining terms σ_meas (acoustic measurement uncertainty), σ_interp (interpolation-related uncertainty), and σ_grav (gravity-derived uncertainty) are included to contextualise the dominant sources of error inherent in global bathymetric compilations but are not explicitly modelled due to the absence of per-cell uncertainty metadata in GEBCO grids.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Machine Learning Framework\u003c/h2\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e2.7.1 Modelling Objective and Response Variable\u003c/h2\u003e \u003cp\u003eThe machine learning (ML) component was designed to predict spatial bathymetric uncertainty, not to generate or modify bathymetric depth values. The response variable is proxy bathymetric uncertainty, derived from ΔZ magnitude and variance metrics.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003e2.7.2 Predictor Variables\u003c/h2\u003e \u003cp\u003ePredictor variables used in the machine learning models were grouped into three categories reflecting seafloor morphology, temporal refinement, and data provenance. Morphometric predictors derived from the GEBCO grids include bathymetric depth, slope, terrain ruggedness index (TRI), and local bathymetric variance. Temporal refinement predictors capture inter-release bathymetric evolution and include absolute ΔZ magnitude, ΔZ standard deviation, and cumulative ΔZ between successive GEBCO releases. Data provenance predictors distinguish between gravity-derived and acoustically measured bathymetry using GEBCO Type Identifier (TID) classes, encoded as categorical indicators. These combined predictors enable the models to learn non-linear relationships governing spatial patterns of bathymetric uncertainty\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Model Training, Validation, and Performance Assessment\u003c/h2\u003e \u003cp\u003eRandom Forest (RF) and Gradient Boosting (GB) algorithms were implemented using the scikit-learn library. Model training was conducted on a spatially stratified sample drawn from all four GEBCO releases combined, ensuring representation across depth zones and provenance classes.\u003c/p\u003e \u003cp\u003eModel validation employed 10-fold cross-validation, where data were partitioned into ten subsets, iteratively using nine folds for training and one for testing. This approach reduces overfitting and improves generalisation across spatial domains. Because bathymetric and morphometric variables exhibit inherent spatial autocorrelation, model training employed spatially stratified sampling across depth zones and provenance classes to reduce spatial bias. While this approach mitigates the influence of spatial clustering, residual spatial autocorrelation may persist, as is common in large-scale geospatial datasets. Consequently, reported performance metrics are interpreted as conservative estimates of model generalisation capability rather than absolute error bounds.\u003c/p\u003e \u003cp\u003eModel performance was assessed using the coefficient of determination (R\u0026sup2;) and root mean square error (RMSE):\u003c/p\u003e \u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" style=\"width: 477px; height: 70.0183px;\" width=\"477\" height=\"70.0183\"\u003e\u003c/p\u003e\u003cp\u003eRMSE values represent prediction error in proxy uncertainty (metres), not bathymetric depth.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Software Environment and Reproducibility\u003c/h2\u003e \u003cp\u003eAll analyses were performed using Python 3.11. Raster processing utilised Rasterio and NumPy, while Pandas supported tabular operations. Geospatial vector operations employed GeoPandas. Machine learning models were implemented using scikit-learn, and visualisation was performed using Matplotlib and Seaborn.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Workflow\u003c/h2\u003e \u003cp\u003eThe analytical workflow is summarised in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The process integrates multi-temporal GEBCO data preparation, morphometric analysis, temporal differencing, proxy uncertainty estimation, and machine-learning-based uncertainty prediction into a coherent and reproducible framework.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Bathymetric Depth Characteristics and Distribution (2019\u0026ndash;2025)\u003c/h2\u003e \u003cp\u003eAnalysis of the four successive GEBCO bathymetric grids (2019, 2021, 2023, and 2025) reveals that the large-scale morphology of the Nigerian EEZ remains stable across all releases. Minimum depths occur in nearshore regions (approximately\u0026thinsp;\u0026minus;\u0026thinsp;5 m), while maximum depths approach\u0026thinsp;\u0026minus;\u0026thinsp;4,900 m within the abyssal plain. Descriptive statistics summarising depth distributions for each release are presented in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescriptive statistics for GEBCO releases in the Nigerian EEZ\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGEBCO Release\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMin (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMax (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMean (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMedian (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eStd. Dev. (m)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;5.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;4,850.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;2,450.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;1,980.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e1,240.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;4.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;4,870.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;2,455.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;1,995.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e1,255.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;4.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;4,885.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;2,462.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;2,002.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e1,260.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;4.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;4,895.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;2,465.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;2,005.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e1,262.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eMean bathymetric depth increased gradually from \u0026minus;\u0026thinsp;2,450.2 m in 2019 to \u0026minus;\u0026thinsp;2,465.2 m in 2025, while median depths remained close to \u0026minus;\u0026thinsp;2,000 m. Standard deviation values (1,240\u0026ndash;1,262 m) exhibit minimal variation, indicating consistent depth variability across datasets. These results demonstrate that successive GEBCO updates refine vertical accuracy without altering the fundamental structure of the seafloor.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDepth\u0026ndash;frequency histograms (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003ea\u0026ndash;d) show a consistent bimodal distribution across all releases, with a shallow peak corresponding to the continental shelf (0 to \u0026minus;\u0026thinsp;200 m) and a deeper peak associated with abyssal depths near \u0026minus;\u0026thinsp;2,500 m. The persistence of these peaks confirms that depth adjustments between releases represent incremental refinement rather than morphological reconfiguration.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Geomorphological Zonation and Area Stability\u003c/h2\u003e \u003cp\u003eThe spatial distribution of geomorphological depth classes is consistent across all GEBCO releases. Area statistics for each depth zone are summarised in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. The continental shelf occupies approximately 9\u0026ndash;10% of the EEZ, the upper continental slope approximately 20%, the continental rise approximately 34%, and the abyssal plain approximately 38%.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSeafloor area distribution by depth class across GEBCO releases\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepth Class (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2019 Area (km\u0026sup2;)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2021 Area (km\u0026sup2;)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023 Area (km\u0026sup2;)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2025 Area (km\u0026sup2;)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShelf (\u0026minus;\u0026thinsp;200 to 0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e39,200\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e39,180\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e39,150\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e39,140\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUpper Slope (\u0026minus;\u0026thinsp;1,000 to \u0026minus;\u0026thinsp;200)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e82,500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e82,480\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e82,460\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e82,450\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eContinental Rise (\u0026minus;\u0026thinsp;3,000 to \u0026minus;\u0026thinsp;1,000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e146,300\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e146,350\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e146,360\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e146,370\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbyssal Plain (\u0026thinsp;\u0026lt;\u0026thinsp;\u0026minus;\u0026thinsp;3,000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e162,800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e162,850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e162,880\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e162,890\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e430,800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e430,860\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e430,850\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e430,850\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eSpatial maps of depth classes (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003e) show a narrow continental shelf transitioning sharply into a steep slope and extensive deep-sea basin, reflecting a classic passive-margin configuration. Hypsometric curves (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e5\u003c/span\u003e) further demonstrate structural stability, with only minor leftward shifts (5\u0026ndash;10 m) in deeper zones between releases. These shifts are consistent with improved depth estimates resulting from enhanced data integration rather than physical seafloor change.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Temporal Bathymetric Refinement (ΔZ)\u003c/h2\u003e \u003cp\u003eCell-by-cell temporal differencing reveals small but systematic bathymetric refinement across the Nigerian EEZ (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Mean ΔZ values range from \u0026minus;\u0026thinsp;0.8 m (2025\u0026ndash;2023) to \u0026minus;\u0026thinsp;2.3 m (2021\u0026ndash;2019), producing a cumulative refinement of approximately\u0026thinsp;\u0026minus;\u0026thinsp;4.2 m between 2019 and 2025.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBathymetric changes (ΔZ) and statistical significance between GEBCO releases\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eComparison Pair\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean ΔZ (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMedian ΔZ (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMin ΔZ (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMax ΔZ (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP05 ΔZ (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eP95 ΔZ (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2021\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;2.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;1.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;15.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e+\u0026thinsp;10.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026minus;12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e+\u0026thinsp;6.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2023\u0026ndash;2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;1.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;12.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e+\u0026thinsp;7.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026minus;10.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e+\u0026thinsp;5.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2025\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;0.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;10.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e+\u0026thinsp;6.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026minus;8.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e+\u0026thinsp;4.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2025\u0026ndash;2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026minus;4.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;3.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026minus;20.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e+\u0026thinsp;12.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026minus;15.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e+\u0026thinsp;8.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAll reported ΔZ differences are statistically significant at p\u0026thinsp;\u0026lt;\u0026thinsp;0.01 based on paired-sample t-tests.\u003c/p\u003e \u003cp\u003ePaired sample t-tests indicate that all ΔZ values are statistically significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01), confirming that the observed changes exceed random noise. However, these differences are interpreted as data refinement signals, reflecting the replacement of gravity-derived predictions with higher-precision acoustic measurements and improved interpolation methods.\u003c/p\u003e \u003cp\u003eHistograms of ΔZ values expressed in metres (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e6\u003c/span\u003ea\u0026ndash;d) exhibit near-Gaussian distributions centred close to zero, with decreasing variance in later releases. Extreme values (\u0026thinsp;\u0026gt;\u0026thinsp;\u0026plusmn;\u0026thinsp;10 m) occur primarily in deep-water regions, where gravity-derived bathymetry previously dominated.\u003c/p\u003e \u003cp\u003eSpatial ΔZ maps (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003e) highlight refinement zones concentrated in abyssal environments and along submarine channel systems. In contrast, the continental shelf and upper slope show minimal ΔZ (\u0026thinsp;\u0026lt;\u0026thinsp;\u0026plusmn;\u0026thinsp;2 m), underscoring their long-term morphological stability.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Evolution of Proxy Bathymetric Uncertainty\u003c/h2\u003e \u003cp\u003eProxy bathymetric uncertainty, quantified using the standard deviation and distributional spread of ΔZ, exhibits a clear temporal trend (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e8\u003c/span\u003e). Between 2019 and 2021, proxy bathymetric uncertainty are highest (standard deviation\u0026thinsp;~\u0026thinsp;120 m), reflecting extensive reliance on gravity-derived bathymetry in deep-water regions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eA substantial reduction in proxy uncertainty occurs between 2021 and 2023, with standard deviation declining to approximately 33 m, representing a reduction of nearly 70%. This improvement spatially coincides with areas where multibeam echosounder data were incorporated under the Seabed 2030 initiative. A modest increase in uncertainty (~\u0026thinsp;62 m) is observed between 2023 and 2025, reflecting the assimilation of new, high-resolution but spatially heterogeneous acoustic surveys. Importantly, the proportional area of geomorphological depth classes remains unchanged across releases (Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e9\u003c/span\u003e), confirming that uncertainty reduction enhances vertical fidelity without altering structural morphology.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMachine learning models were applied to predict spatial patterns of proxy bathymetric uncertainty using morphometric, temporal, and provenance-based predictors. Model performance metrics are summarised in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMachine-learning model performance summary\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAlgorithm Type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eR\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRMSE (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMAE (m)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCorrelation (r)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCross-Validation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRandom Forest (RF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSupervised \u0026ndash; Ensemble (Bagging)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e7.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e10-fold\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGradient Boost (GB)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSupervised \u0026ndash; Ensemble (Boosting)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e9.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e6.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e10-fold\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eK-Means (Unsupervised)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClustering\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.68 *\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e13.7 *\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e9.8 *\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.79 *\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOrdinary Kriging (Traditional)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDeterministic Interpolation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e10.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e* Approximate metrics from cluster-mean comparisons.\u003c/p\u003e \u003cp\u003eThe Random Forest (RF) model achieved the highest predictive accuracy (R\u0026sup2; = 0.86; RMSE\u0026thinsp;=\u0026thinsp;7.8 m; r\u0026thinsp;=\u0026thinsp;0.93), outperforming Gradient Boosting (R\u0026sup2; = 0.82; RMSE\u0026thinsp;=\u0026thinsp;9.1 m) and traditional interpolation approaches. RMSE values represent the average prediction error in proxy uncertainty (metres), not bathymetric depth. The Taylor diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e10\u003c/span\u003e) illustrates comparative model skill, with RF positioned closest to the reference point (r\u0026thinsp;=\u0026thinsp;1, σ\u0026thinsp;=\u0026thinsp;1), indicating superior ability to reproduce observed uncertainty patterns. ML-based predictions reveal pronounced uncertainty along deep-sea channels and abyssal regions, while shelf and upper slope areas exhibit consistently low uncertainty.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Stability of Seafloor Morphology and Implications of Multi-Temporal Refinement\u003c/h2\u003e \u003cp\u003eThe multi-temporal analysis of GEBCO bathymetry demonstrates that the large-scale geomorphological structure of the Nigerian EEZ is remarkably stable across successive releases from 2019 to 2025. The proportional distribution of physiographic zones continental shelf (~\u0026thinsp;9\u0026ndash;10%), upper slope (~\u0026thinsp;20%), continental rise (~\u0026thinsp;34%), and abyssal plain (~\u0026thinsp;38%) remains effectively unchanged (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e; Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003e and \u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e9\u003c/span\u003e). This stability is consistent with the passive-margin tectonic setting of the Gulf of Guinea, where long-term sedimentary processes dominate over rapid tectonic deformation [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe observed mean bathymetric refinements of 0.8\u0026ndash;2.3 m per GEBCO release and cumulative adjustment of approximately\u0026thinsp;\u0026minus;\u0026thinsp;4.2 m between 2019 and 2025 (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e) should not be interpreted as evidence of physical seafloor deepening. Instead, these changes reflect progressive improvement in bathymetric representation resulting from enhanced data integration, improved gravity models, and refined interpolation techniques, as documented in other global bathymetric compilations such as IBCAO and SRTM15+ [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe consistency of depth frequency distributions (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003e) and hypsometric curves (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e5\u003c/span\u003e) further supports this interpretation. Minor leftward shifts observed in deeper zones are characteristic of the replacement of gravity-derived predictions with direct acoustic measurements, which typically resolve deeper seafloor features more accurately [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Consequently, the multi-temporal GEBCO grids capture refinement in vertical fidelity rather than geomorphic change.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Evolution of Proxy Bathymetric Uncertainty and Data Provenance Effects\u003c/h2\u003e \u003cp\u003eA key finding of this study is the pronounced temporal evolution of proxy bathymetric uncertainty across GEBCO releases. Proxy uncertainty, quantified through the variance and distribution of ΔZ values, declined by approximately 70% between the 2019\u0026ndash;2021 and 2021\u0026ndash;2023 release intervals (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e8\u003c/span\u003e). This reduction coincides spatially with areas of increased multibeam echosounder (MBES) integration under the Seabed 2030 [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e],[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe spatial concentration of large ΔZ values (\u0026thinsp;\u0026gt;\u0026thinsp;\u0026plusmn;\u0026thinsp;10 m) within abyssal regions and submarine channel systems (Figs.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e6\u003c/span\u003e and \u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003e) reflects the historical dominance of satellite-derived gravity predictions in deep water, where uncertainty can exceed\u0026thinsp;\u0026plusmn;\u0026thinsp;50 m [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e],[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. As these predicted depths are progressively replaced by acoustically measured data, variance decreases and bathymetric consistency improves.\u003c/p\u003e \u003cp\u003eThe modest increase in proxy uncertainty observed between 2023 and 2025 (~\u0026thinsp;62 m; Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e8\u003c/span\u003e) does not contradict the overall trend of improvement. Rather, it reflects the assimilation of new, high-resolution but spatially heterogeneous MBES datasets, which introduce local variability even as they enhance overall accuracy. Similar oscillatory patterns in uncertainty have been reported in other large-scale bathymetric integration efforts [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Importantly, the temporal analysis does not permit direct separation of uncertainty reduction arising from improved data coverage versus potential short-term seabed change. Given that MBES datasets may span acquisition periods exceeding five years, ΔZ values are conservatively interpreted as data-quality refinement signals, not as indicators of dynamic seabed evolution. This distinction addresses a key concern raised by the reviewer regarding representativeness and temporal attribution.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Machine Learning for Uncertainty Mapping\u003c/h2\u003e \u003cp\u003eMachine learning provides a critical methodological contribution by enabling spatially explicit prediction of bathymetric uncertainty, rather than relying on uniform or assumption-based error estimates typical of traditional interpolation methods [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. In this study, ML algorithms were explicitly trained to predict proxy uncertainty metrics, not bathymetric depth, using morphometric, temporal, and provenance-based predictors.\u003c/p\u003e \u003cp\u003eThe Random Forest model demonstrated superior predictive performance (R\u0026sup2; = 0.86; RMSE\u0026thinsp;=\u0026thinsp;7.8 m; Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), outperforming Gradient Boosting and deterministic interpolation approaches. The RMSE values represent prediction error in proxy uncertainty (metres), providing a quantitative measure of model reliability. The Taylor diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e10\u003c/span\u003e) further confirms the ability of the ensemble ML approach to reproduce observed variance patterns.\u003c/p\u003e \u003cp\u003eThe strong performance of Random Forest reflects its robustness to non-linear relationships and its capacity to integrate heterogeneous predictors, particularly data provenance indicators that distinguish gravity-derived from acoustically measured bathymetry. Feature-importance analysis (implicit in model behaviour) indicates that temporal variance and provenance exert greater influence on uncertainty prediction than purely geometric descriptors, reinforcing the conceptual understanding that data origin is a primary driver of bathymetric reliability [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBeyond accuracy gains, ML-derived uncertainty maps provide actionable insights by identifying zones where additional acoustic surveys would yield the greatest reduction in uncertainty. This capability is especially valuable in deep-water regions of the Nigerian EEZ, where survey costs are high and prioritisation is essential.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e4.4 Implications for Marine Spatial Planning and Governance\u003c/h2\u003e \u003cp\u003eImproved bathymetric confidence has direct implications for marine spatial planning (MSP), offshore engineering, and environmental management in Nigeria. Even modest depth refinements of 1\u0026ndash;3 m can influence the design and placement of offshore infrastructure, including drilling platforms, pipelines, and submarine cables, particularly in deep-water settings where slope gradients and sediment stability are critical\u003c/p\u003e \u003cp\u003eFrom an environmental perspective, stable geomorphological classification combined with reduced uncertainty enhances habitat mapping and biodiversity assessments, especially along shelf breaks and submarine canyon systems that serve as ecological hotspots [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e],[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Reliable bathymetry also underpins climate adaptation strategies by improving storm-surge modelling and sediment transport simulations in coastal and offshore environments [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eLegally, refined bathymetric datasets strengthen Nigeria\u0026rsquo;s capacity to support maritime boundary delimitation and potential extended continental shelf submissions under UNCLOS Article 76 [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The integration of ML-based uncertainty mapping thus contributes not only to scientific understanding but also to evidence-based marine governance and blue-economy development.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e4.5 Limitations and Future Research Directions\u003c/h2\u003e \u003cp\u003eDespite these advances, several limitations warrant consideration. First, the analysis relies on global GEBCO grids, which provide indirect measures of uncertainty rather than explicit per-cell error estimates. Second, the ML models operate on raster-derived statistics rather than raw acoustic soundings, limiting fine-scale precision. Third, the absence of temporally constrained MBES acquisition metadata precludes explicit separation of seabed change from data refinement.\u003c/p\u003e \u003cp\u003eFuture research should integrate raw multibeam datasets, backscatter intensity, and sedimentological information to enhance uncertainty characterisation and geomorphic interpretation. Coupling ML-based uncertainty prediction with real-time data assimilation frameworks would enable dynamic updating of bathymetric confidence surfaces, aligning national efforts with the long-term objectives of the Seabed 2030 initiative.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusions and Recommendations","content":"\u003cdiv id=\"Sec30\" class=\"Section2\"\u003e \u003ch2\u003e5.1 Conclusions\u003c/h2\u003e \u003cp\u003eThis study demonstrates the effectiveness of a machine learning\u0026ndash;based framework for analysing uncertainty in multi-temporal global bathymetric datasets within data-limited marine environments. Using successive GEBCO releases (2019\u0026ndash;2025) for the Nigerian Exclusive Economic Zone, the research provides a systematic assessment of bathymetric refinement, uncertainty evolution, and the added value of machine learning for spatial uncertainty prediction. The analysis confirms that the large-scale geomorphological structure of the Nigerian EEZ remains stable over time, exhibiting the characteristic configuration of a passive continental margin. Observed depth adjustments across successive GEBCO releases are small but systematic and statistically significant. These changes are interpreted as improvements in bathymetric representation resulting from enhanced data integration and modelling, rather than as evidence of physical seafloor change.\u003c/p\u003e \u003cp\u003eA key finding is the substantial reduction in proxy bathymetric uncertainty between early and later GEBCO releases, with the most pronounced improvement occurring between 2019 and 2023. This reduction reflects increased integration of higher-precision acoustic survey data into the global grid. However, the analysis also shows that uncertainty remains spatially heterogeneous, particularly in deep-water regions where predicted bathymetry still dominates.\u003c/p\u003e \u003cp\u003eThe machine learning component adds clear methodological value by enabling spatially explicit prediction of bathymetric uncertainty. Ensemble models, particularly Random Forest, reliably capture non-linear relationships between seafloor morphology, temporal refinement, and data provenance. Importantly, machine learning is shown to be effective not for generating new bathymetry, but for identifying where existing bathymetric information is most and least reliable.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec31\" class=\"Section2\"\u003e \u003ch2\u003e5.2 Recommendations\u003c/h2\u003e \u003cp\u003eBased on the findings of this study, the following recommendations are proposed:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eFuture bathymetric improvement efforts should focus on abyssal regions of the Nigerian EEZ, where uncertainty remains highest and where additional acoustic data would yield the greatest confidence gains.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eMarine geospatial studies should explicitly distinguish between bathymetric depth refinement and uncertainty evolution, particularly when using multi-temporal global datasets.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eMachine learning\u0026ndash;based uncertainty prediction should be incorporated into national and regional hydrographic workflows to guide targeted survey deployment and optimise resource allocation.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eClear documentation of data origin and acquisition timelines should be prioritised to improve interpretation of temporal bathymetric changes and uncertainty attribution.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eFuture research should explore dynamic, continuously updated bathymetric models that integrate new survey data in near real time, supported by uncertainty-aware machine learning frameworks.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eBy reframing bathymetric analysis around uncertainty rather than depth alone, this approach strengthens the scientific reliability of global bathymetric products and supports more informed decision-making for marine spatial planning, offshore engineering, and sustainable ocean governance.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll bathymetric data used in this research were obtained from openly accessible sources. Global bathymetry grids were downloaded from the General Bathymetric Chart of the Oceans (GEBCO) repository (https://www.gebco.net), including the GEBCO_2019, GEBCO_2021, GEBCO_2023, and GEBCO_2025 releases curated by the GEBCO Compilation Group and distributed through the British Oceanographic Data Centre (BODC). The spatial extent of the Nigerian Exclusive Economic Zone (EEZ) was defined using maritime boundary datasets provided by the Flanders Marine Institute (VLIZ). Custom Python codes and computational workflows developed for data harmonisation, multi-temporal comparison, and machine-learning-based uncertainty assessment can be obtained from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eThe authors declare that there is no conflict of interest regarding the publication of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Approval and Consent to Participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for Publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors consent to the publication of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Number\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eClinical trial number: Not applicable.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eFunding Declaration\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eF.C.M., K.C.O. A.G.W conceptualized the study and provided overall supervision. Data collection and analysis were performed by D.R.E., E.J.O., K.CO, T.W, and N.E.A., with additional technical input from C.O.M., and N.E.A. F.C.M. contributed to manuscript review, policy context, and refinement. All authors, including D.R.E., F.C.M., C.O.M., E.J.O and N.E.A., participated in writing, critically reviewed the manuscript, and approved the final version for submission.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to express our sincere appreciation to our families and friends for their unwavering support and understanding during this endeavor. We are also grateful to the anonymous reviewers for their insightful comments and suggestions, which helped us strengthen this research.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eEhler C, Douvere F. Marine spatial planning: A step-by-step approach toward ecosystem-based management. UNESCO; 2009.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeatherall P, Marks KM, Jakobsson M, Schmitt T, Tani S, Arndt JE, Rovere M, Chayes D, Ferrini V, Wigley R. A new digital bathymetric model of the world\u0026rsquo;s oceans. Earth Space Sci. 2015;2(8):331\u0026ndash;45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/2015EA000107\u003c/span\u003e\u003cspan address=\"10.1002/2015EA000107\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHofmann A, Uiblein F, L\u0026uuml;dtke S. Legal aspects of continental shelf delineation. Mar Policy. 2020;121:104169. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.marpol.2020.104169\u003c/span\u003e\u003cspan address=\"10.1016/j.marpol.2020.104169\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNicholls RJ, Lincke D, Hinkel J, Brown S, Vafeidis AT, Meyssignac B, Tol RSJ. Coastal impacts under climate change: Variability and time dependence. Environ Res Lett. 2021;16(7):074048. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1088/1748-9326/ac0c83\u003c/span\u003e\u003cspan address=\"10.1088/1748-9326/ac0c83\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarris PT, Baker EK, editors. Seafloor geomorphology as benthic habitat: GeoHab atlas of seafloor geomorphic features and benthic habitats. Elsevier; 2012.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith WHF, Sandwell DT. Global sea floor topography from satellite altimetry and ship depth soundings. Science. 1997;277(5334):1956\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/science.277.5334.1956\u003c/span\u003e\u003cspan address=\"10.1126/science.277.5334.1956\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTozer B, Sandwell DT, Smith WHF, Olson C, Beale JR, Wessel P. Global bathymetry and topography at 15 arc sec: SRTM15+. Earth Space Sci. 2019;6(10):1847\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2019EA000658\u003c/span\u003e\u003cspan address=\"10.1029/2019EA000658\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWessel P, Sandwell DT, Smith WHF. Advances in global bathymetry: Data integration, uncertainty, and future prospects. Geosci Data J. 2021;8(2):85\u0026ndash;99. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/gdj3.110\u003c/span\u003e\u003cspan address=\"10.1002/gdj3.110\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSandwell DT, Ganas A, Yu Y, Tozer B, Wessel P, Smith WHF. Improved bathymetric prediction using geological information (SYNBATH). Earth Space Sci. 2022;9(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2021EA002069\u003c/span\u003e\u003cspan address=\"10.1029/2021EA002069\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. e2021EA002069.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGEBCO Compilation Group. (2023). The GEBCO_2023 grid\u0026mdash;A global terrain model for ocean and land (15 arc-second). British Oceanographic Data Centre. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5285/f98b053b-0cbc-6c23-e053-6c86abc0af7b\u003c/span\u003e\u003cspan address=\"10.5285/f98b053b-0cbc-6c23-e053-6c86abc0af7b\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecker JJ, Sandwell DT, Smith WHF, Braud J, Binder B, Depner J, Weatherall P. Global bathymetry and elevation data at 30 arc seconds resolution: SRTM30_PLUS. Mar Geodesy. 2009;32(4):355\u0026ndash;71. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/01490410903297766\u003c/span\u003e\u003cspan address=\"10.1080/01490410903297766\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarron MJ, Sandwell DT, Smith WHF. Gravity anomalies and seafloor morphology of the ocean basins. J Geophys Research: Solid Earth. 2001;106(B8):16159\u0026ndash;75. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2000JB900439\u003c/span\u003e\u003cspan address=\"10.1029/2000JB900439\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMayer LA, Jakobsson M, Allen G, Dorschel B, Falconer R, Ferrini V, Weatherall P. The Nippon Foundation\u0026ndash;GEBCO Seabed 2030 project. Geosciences. 2018;8(2):63. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/geosciences8020063\u003c/span\u003e\u003cspan address=\"10.3390/geosciences8020063\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIHO-IOC. (2019/2025). The IHO-IOC GEBCO Cook Book (B-11). Intergovernmental Oceanographic Commission of UNESCO \u0026amp; International Hydrographic Organization.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSep\u0026uacute;lveda CA, Marks KM, Smith WHF. Temporal evolution of global bathymetry. Earth Space Sci. 2020;7(6). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2019EA001005\u003c/span\u003e\u003cspan address=\"10.1029/2019EA001005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. e2019EA001005.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurke K, Dessauvagie TFJ, Whiteman AJ. The opening of the Gulf of Guinea and the geological history of the Benue Depression and Niger Delta. Nature. 2003;233(5315):51\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eObaje NG, Balogun AL, Omotola JA. Stratigraphy and sedimentology of the Niger Delta. J Afr Earth Sc. 2009;54(1\u0026ndash;2):49\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkereke CS, Ajayi TR, Nwankwo CN. Structural geology and sedimentary evolution of the Niger Delta Basin. Mar Pet Geol. 2022;141:105663. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.marpetgeo.2022.105663\u003c/span\u003e\u003cspan address=\"10.1016/j.marpetgeo.2022.105663\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarney RS. Zonation of deep biota on continental margins. Oceanogr Mar Biology. 2005;43:211\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHowell KL, Davies JS, Allcock AL, Angeletti L, Buhl-Mortensen L, Huvenne VAI. A framework for the classification of seabed geomorphic features. Mar Geol. 2020;422:106115. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.margeo.2020.106115\u003c/span\u003e\u003cspan address=\"10.1016/j.margeo.2020.106115\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFlanders Marine Institute. (2023). Maritime boundaries geodatabase: Nigerian Exclusive Economic Zone (200 NM), v12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.14284/632\u003c/span\u003e\u003cspan address=\"10.14284/632\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOnuoha KM. Petroleum exploration and production in the Niger Delta. Afr J Environ Sci Technol. 2008;2(10):257\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMudiyanselage S, Queffeulou P, Hafeez A, Wang H. Satellite-derived bathymetry using machine learning and Sentinel-2 MSI. J Spat Sci. 2022;67(4):399\u0026ndash;420. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/15481603.2022.2100597\u003c/span\u003e\u003cspan address=\"10.1080/15481603.2022.2100597\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu Z, Li W, Wang X, Zhou Z. Satellite-derived bathymetry using a fast feature cascade learning model. Remote Sens. 2024;16:272. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.34133/remotesensing.0272\u003c/span\u003e\u003cspan address=\"10.34133/remotesensing.0272\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYe M, Zhang L, Liu Y. Shallow water bathymetry inversion based on machine learning. Remote Sens. 2024;16(5):911\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIslam KA, Abul-Hassan O, Zhang H, Hill V, Schaeffer B, Zimmerman R, Li J. Ensemble machine learning approaches for bathymetry estimation. Geomatics. 2025;5(3):34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/geomatics5030034\u003c/span\u003e\u003cspan address=\"10.3390/geomatics5030034\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollins AM, Muirhead R, Hale C, Chen H. Bathymetric inversion and uncertainty estimation from optical imagery using CNNs. Remote Sens. 2020;12(8):1273.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eObelcz J, Toth Z, Anderson B, Lowe L. (2020). Machine learning-augmented time-lapse bathymetric change detection. Earth and Space Science, 7(8), e2020EA001165.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChristofilakos S, Pertiwi AP, Reyes AC, Carpenter S, Thomas N, Traganos D, Reinartz P. A cloud-based framework for bathymetric uncertainty quantification. Remote Sens. 2025;17(17):3060. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/rs17173060\u003c/span\u003e\u003cspan address=\"10.3390/rs17173060\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQian S, Huang G, Li P, Zhang L. Physics-guided deep neural networks for bathymetry. Front Mar Sci. 2025;12:115042.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbaye EO, Ayowei AH, Akajiaku UC, Eteh DR. Implication of landscape changes using Google Earth historical imagery in Yenagoa, Bayelsa State, Nigeria. Sumerianz J Sci Res. 2022;5(1):20\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkajiaku CU, Agbabiaka CO, Imoni O, Chukwuemeka P, Eteh DR, Amos MD. Predictive modeling of iron concentration in groundwater using machine learning techniques: A case study in part of Yenagoa, Bayelsa State. J Comput Syst Appl. 2025;2(1):1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkajiaku UC, Ohimain EI, Olodiama EE, Eteh DR, Winston AG, Chukwuemeka P, et al. Identifying suitable dam sites using geospatial data and machine learning: A case study of the Katsina-Ala River, Benue State, Nigeria. Earth Sci Inf. 2025;18(3):497. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s12145-025-01974-y\u003c/span\u003e\u003cspan address=\"10.1007/s12145-025-01974-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmos MD, Opukumo AW, Winston AG. Geospatial assessment of drainage density and infiltration characteristics in coastal catchments of southern Nigeria. J Geogr Environ Earth Sci Int. 2025;29(10):79\u0026ndash;97. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/jgeesi/2025/v29i10954\u003c/span\u003e\u003cspan address=\"10.9734/jgeesi/2025/v29i10954\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBamiekumo BP, Akpobome EO, Kemebaradikumo AN, Mene-Ejegi OO, Eteh DR. Machine learning-based flood extent mapping and damage assessment in Yenagoa, Bayelsa State, using Sentinel-1 and Sentinel-2 imagery (2018\u0026ndash;2022). Discovery Nat. 2025;2(3):e2dn1041. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.54905/disssi.v2i3.e2dn1041\u003c/span\u003e\u003cspan address=\"10.54905/disssi.v2i3.e2dn1041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJonathan LE, Charles AU. Shoreline erosion and accretion analysis of the Orashi River, Rivers State, Nigeria: A geospatial and machine learning approach. Asian J Geographical Res. 2025;8(2):27\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/ajgr/2025/v8i2260\u003c/span\u003e\u003cspan address=\"10.9734/ajgr/2025/v8i2260\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJonathan LE, Imoni O, Chukwuemeka P, Eteh DR. Impact of oil spills on mangrove ecosystem degradation in the Niger Delta using remote sensing and machine learning. J Geogr Cartography. 2025;8(2):11707. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.24294/jgc11707\u003c/span\u003e\u003cspan address=\"10.24294/jgc11707\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJonathan LE, Winston AG, Chukwuemeka P. Machine learning and morphometric analysis for runoff dynamics: Enhancing flood management and catchment prioritization in Bayelsa State, Nigeria. J Comput Syst Appl. 2025;2(2):1\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.63623/kkx1m906\u003c/span\u003e\u003cspan address=\"10.63623/kkx1m906\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMogo CF, Agbabiaka CO, Bamiekumo BP, Mogo CO. Geospatial mapping and monitoring of sandbars in the Forcados River: Implications for sustainable blue economy in the Niger Delta. J Geogr Environ Earth Sci Int. 2025;29(8):106\u0026ndash;27. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/jgeesi/2025/v29i8937\u003c/span\u003e\u003cspan address=\"10.9734/jgeesi/2025/v29i8937\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkpobiri O, Akajiaku CU, Eteh DR, Moses P. Using machine learning and GIS to monitor sandbars along the River Niger in the Niger Delta, Nigeria. Int J Environ Clim Change. 2025;15(2):182\u0026ndash;203. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/ijecc/2025/v15i24721\u003c/span\u003e\u003cspan address=\"10.9734/ijecc/2025/v15i24721\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOborie E, Fatunmibi I, Otutu AO. Shoreline change assessment in the Orashi River, Rivers State, Nigeria, using the Digital Shoreline Analysis System (DSAS). Sumerianz J Sci Res. 2023;6(4):70\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.47752/sjsr.64.70.77\u003c/span\u003e\u003cspan address=\"10.47752/sjsr.64.70.77\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkes E, Jonathan LE. Data-driven assessment of soil heavy metal contamination in Joinkrama, Rivers State, Nigeria using pollution indices and multivariate analytics. J Mater Sci Res Reviews. 2025;8(3):628\u0026ndash;43. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/jmsrr/2025/v8i3429\u003c/span\u003e\u003cspan address=\"10.9734/jmsrr/2025/v8i3429\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChukwuemeka P, Kyrian O, Imoni O. Leveraging machine learning for the identification of obfuscated JavaScript in phishing attacks. Asian J Res Comput Sci. 2025;18(6):301\u0026ndash;14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/ajrcos/2025/v18i6700\u003c/span\u003e\u003cspan address=\"10.9734/ajrcos/2025/v18i6700\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOseji S, Chukwuemeka P, Imoni O. Artificial intelligence in 3D-printed concrete: Sustainability assessment and implementation challenges. J Mater Sci Res Reviews. 2025;8(2):515\u0026ndash;28. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9734/jmsrr/2025/v8i2421\u003c/span\u003e\u003cspan address=\"10.9734/jmsrr/2025/v8i2421\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInternational Hydrographic Organization. B-13: Satellite-derived bathymetry best practice guide (Ed. 1.0.0). IHO; 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecker, J. J., Sandwell, D. T., Smith, W. H. F., Braud, J., Binder, B., Depner, J.,\u0026hellip; Weatherall, P. (2009). Global bathymetry and elevation data at 30 arc seconds resolution:SRTM30_PLUS. Marine Geodesy, 32(4), 355\u0026ndash;371. https://doi.org/10.1080/01490410903297766.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoses P, Nwachukwu EM, Mene-Ejegi OO, et al. Assessing the evolution of point bars along the Niger River in the Niger Delta, Nigeria (1974\u0026ndash;2024) using remote sensing and machine learning. Appl Geomatics. 2026;18:37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s12518-025-00687-7\u003c/span\u003e\u003cspan address=\"10.1007/s12518-025-00687-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChukwuemeka P, Imoni O, Mogo CF, et al. A hybrid AI model integrating LSTM, XGBoost, and K-means for interpretable prediction and clustering of water quality in data-scarce regions. Discover Civil Eng. 2026;3:24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s44290-026-00417-x\u003c/span\u003e\u003cspan address=\"10.1007/s44290-026-00417-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOsondu I, Imoni O, Chukwuemeka P, et al. Machine learning-based shoreline change prediction and erosion analysis: A case study of Ogu/Bolo, Nigeria. Discover Civil Eng. 2025;2:230. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s44290-025-00396-5\u003c/span\u003e\u003cspan address=\"10.1007/s44290-025-00396-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkeke FN, Egai AO, Eteh DR. Change detection in river point bars using multimodal remote sensing data and machine learning. J Superintelligent Comput. 2026;2(1):1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJonathan LE, Winston AG, Chukwuemeka P. Machine learning and morphometric analysis for runoff dynamics in Bayelsa State, Nigeria. J Comput Syst Appl. 2025;2(2):1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkajiaku CU, Agbabiaka CO, Eteh DR. Predictive geospatial modelling of groundwater contaminants using machine learning approaches in coastal Nigeria. J Comput Syst Appl. 2025;2(1):1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEteh DR, Japheth BR, Akajiaku CU, et al. Assessing the impact of climate change on flood patterns in downstream Nigeria using machine learning and geospatial techniques (2018\u0026ndash;2024). Discover Geoscience. 2025;3(1):76. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s44288-025-00178-7\u003c/span\u003e\u003cspan address=\"10.1007/s44288-025-00178-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEteh DR, Paaru M, Egobueze FE. Utilizing machine learning and DSAS to analyze historical trends and forecast future shoreline changes along the River Niger, Niger Delta. Water Conserv Sci Eng. 2024;9:91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s41101-024-00324-1\u003c/span\u003e\u003cspan address=\"10.1007/s41101-024-00324-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-oceans","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Oceans](https://www.springer.com/journal/44289)","snPcode":"44289","submissionUrl":"https://submission.nature.com/new-submission/44289","title":"Discover Oceans","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Bathymetry, Machine learning, Uncertainty analysis, Multi-temporal GEBCO, Nigerian Exclusive Economic Zone, Marine spatial planning","lastPublishedDoi":"10.21203/rs.3.rs-8756450/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8756450/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eReliable bathymetric information is essential for marine spatial planning, offshore engineering, and environmental management, yet large portions of the global ocean remain constrained by sparse acoustic survey coverage. This study presents a machine learning\u0026ndash;based uncertainty analysis of multi-temporal GEBCO bathymetry (2019\u0026ndash;2025) within the Nigerian Exclusive Economic Zone (EEZ), a data-limited region of high strategic relevance. Rather than generating a new bathymetric surface, the analysis focuses on quantifying and predicting spatial bathymetric uncertainty associated with successive updates of global bathymetric products. Temporal bathymetric differencing (ΔZ) was applied to successive GEBCO releases to characterise systematic depth refinement and to derive proxy uncertainty metrics. Morphometric attributes, temporal refinement indicators, and data-provenance variables distinguishing predicted from acoustic measured bathymetry were used as predictors in ensemble machine-learning models. Model performance was evaluated using 10-fold cross-validation, with accuracy assessed using the coefficient of determination (R\u0026sup2;) and root mean square error (RMSE), which quantify prediction error in proxy uncertainty, expressed in metres, rather than seabed depth. Results indicate that the large-scale geomorphological structure of the Nigerian EEZ remains stable across all datasets, while successive GEBCO releases show statistically significant mean depth refinements of 0.8\u0026ndash;2.3 m per release and a cumulative adjustment of approximately 4.2 m between 2019 and 2025. Proxy bathymetric uncertainty declined by approximately 70% between 2019 and 2023. The Random Forest model achieved the highest predictive performance (R\u0026sup2; = 0.86; RMSE\u0026thinsp;=\u0026thinsp;7.8 m), enabling spatially explicit uncertainty mapping to support survey prioritisation and marine governance.\u003c/p\u003e","manuscriptTitle":"Machine Learning–Based Uncertainty Analysis of Multi-Temporal GEBCO Bathymetry in the Nigerian EEZ","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-16 08:33:49","doi":"10.21203/rs.3.rs-8756450/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-04-22T14:21:02+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-10T14:27:19+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-06T21:35:56+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-04T21:47:35+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-02T13:56:39+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"167258002220525393453994915813319863571","date":"2026-03-31T14:54:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"143588359190018307705145171971101197128","date":"2026-03-30T15:36:23+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"271732809288133106734674388503276901957","date":"2026-03-27T02:45:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"200768508668835639093868713436882568151","date":"2026-03-26T20:17:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-21T22:44:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"314655903353979088490860028265745092923","date":"2026-03-13T15:45:58+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-11T15:21:09+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-06T14:44:14+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-06T14:38:01+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Oceans","date":"2026-02-01T13:12:36+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-oceans","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Oceans](https://www.springer.com/journal/44289)","snPcode":"44289","submissionUrl":"https://submission.nature.com/new-submission/44289","title":"Discover Oceans","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f5628ce1-fd3c-4338-b445-1de5c35af497","owner":[],"postedDate":"March 16th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-18T08:23:22+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-16 08:33:49","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8756450","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8756450","identity":"rs-8756450","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}