A total of 35 recent surface sediment samples were collected from nine distinct locations across the Indonesian Marine Region, as shown in Fig. 1, spanning a bathymetric transect from 0 to 4327 m water depth (Fig. 2). The sampling locations cover various oceanographic conditions, from the Java Sea, Sunda Strait, Makassar Strait, Sumba Strait, and island of Simeulue to offshore Papua. This spatial coverage represents the environmental heterogeneity across the Indonesian Archipelago, ranging from shallow shelves to deep basins. Complete station metadata including coordinates, sampling dates, and environmental conditions are provided in Table 1. It is important to note that data on the two most important environmental factors within the TROX model – organic carbon flux and dissolved oxygen concentration – were not recorded at the sites.

Based on the classification scheme adapted from Tipsword et al. (1966), Ingle (1980), and van Morkhoven et al. (1986), samples were grouped into six bathymetric zones with regional adaptations for Indonesian oceanographic conditions as shown in Table 2. The transitional zone represents beach face to shallow subtidal environments above a fair-weather wave base, characterized by frequent sediment reworking, high-energy conditions, and a strong influence of wave action and long-shore currents. These stations experience active hydrodynamic winnowing and potential onshore transport of biogenic material, which significantly affect foraminiferal assemblage composition.

The neritic classification follows Tipsword et al. (1966). The inner neritic zone extends from low tide to the fair-weather wave base, characterized by wave influence, normal marine salinity (35 ‰), and well-oxygenated conditions (>5 mL L⁻¹ O₂). The food supply is abundant in the inner neritic zone, driven by terrestrial input and primary productivity in the photic zone, and decreases toward the outer neritic zone. The middle neritic zone spans from the fair-weather wave base to the storm wave base (∼ 100 m), where episodic storm influence occurs with reduced wave energy. The outer neritic zone extends from the storm wave base to the shelf edge, characterized by minimal wave influence and increased planktonic input (Murray, 2006).

The bathyal zone combines classifications from Ingle (1980) and van Morkhoven et al. (1986). The upper bathyal zone represents the upper continental slope with oxygen-minimum zone influence (2–4 mL L⁻¹ O₂) and variable organic matter flux. The lower bathyal zone is characterized by low-oxygen conditions (< 2 mL L⁻¹ O₂), high hydrostatic pressure (>100 bars), and specialized deep-water assemblages adapted to dysoxic and variable organic carbon flux conditions (Murray, 2006). This regional adaptation is scientifically justified for the Indo-Pacific where different oceanographic conditions create bathyal environments extending to approximately 4000 m depth, differing from the traditional 2000 m limit used in Atlantic-based schemes.

Sample collection employed grab samplers for shallow waters (0–168 m), box cores for mid-depth sampling (219–1750 m), and gravity cores for deeper areas (750–4327 m). For all sampling methods, only the uppermost sediment layer (0–10 cm) was analyzed to ensure examination of (sub)recent assemblages. While this depth represents different temporal spans across the bathymetric gradient (shorter time periods in high-energy shallow environments, potentially several thousand years in deeper, low-sedimentation environments), this approach provides the most recent assemblages available and maintains consistency in methodology across all sample sites. As such, our samples will contain a better representative time-averaged “dead assemblage” for comparison with downcore studies.

Statistical analyses were conducted using PAST (Paleontological Statistics) software (Hammer et al., 2001). Samples were standardized by sample totals prior to analysis to account for variations in total foraminiferal counts between samples. Hierarchical cluster analysis was performed to group samples based on foraminiferal compositional similarity using Bray–Curtis dissimilarity with average linkage (UPGMA) clustering. Bray–Curtis dissimilarity is appropriate for ecological abundance data because it excludes joint absences (double zeros) and focuses on shared species presence, making it more ecologically meaningful than Euclidean distance for community data (Legendre and Legendre, 2012). Unlike Euclidean distance, which treats shared absences as evidence of similarity, Bray–Curtis recognizes that the mutual absence of species provides no information about ecological similarity. The dissimilarity ranges from 0 (identical communities) to 1 (completely different communities). Cluster stability was assessed using cophenetic correlation coefficients to evaluate how faithfully the dendrogram preserves the original dissimilarity relationships.

Detrended correspondence analysis (DCA) was employed as the primary ordination method to visualize the relationship between samples along environmental gradients. DCA is particularly suitable for ecological data exhibiting unimodal species responses along environmental gradients, where species show peak abundances at intermediate values of environmental variables (Hill and Gauch, 1980). The method employs detrending by segments to remove the arch effect commonly observed in correspondence analysis and uses reciprocal averaging to simultaneously ordinate both samples and species. DCA axis lengths were calculated to confirm gradient length, and the relationship between DCA axis-1 scores and water depth was quantified using linear regression to validate depth as the primary environmental gradient. Principal component analysis (PCA) was performed separately for each bathymetric zone at both the genus and species levels to identify additional environmental factors controlling foraminiferal assemblages within zones. PCA results for zones with fewer than five samples are presented as descriptive only and are not interpreted as robust ecological signals due to rank limitations. Detailed PCAs are in Document S1 in the Supplement.

Indicator value (IndVal) analysis following Dufrêne and Legendre (1997) was used to identify indicator species for each bathymetric zone. IndVal combines specificity (how concentrated a species occurrence is in particular groups) with fidelity (the faithfulness of the occurrence of a species in particular groups), providing a robust measure of species–environment associations. The IndVal method calculates the product of the relative frequency and relative average abundance of species in each group, with values ranging from 0 % to 100 %. Species with IndVal values ≥ 25 % were classified as primary indicators. The 25 % threshold represents taxa with approximately 50 % occurrence within their preferred environment and has been established as a conservative cutoff for identifying ecologically meaningful indicator species (Dufrêne and Legendre, 1997). Primary indicator taxa were defined as those appearing consistently in both DCA/PCA factor loadings and IndVal analyses, while secondary indicators were identified through either ordination loadings or IndVal analysis alone. To ensure robustness of these classifications, all identified taxa were independently validated through ordination analyses, where primary indicators consistently appeared as strong axis loadings. Complete IndVal results for all taxa at the genus and species levels are provided in Tables S2–S3 in the Supplement.

Complete foraminiferal data, including specimen counts, species richness, and %P for all samples, are presented in Table 3, while sediment characteristics are provided in Table 4. The %P values in Table 3 provide the exact quantitative values for each sample, while Fig. 3 illustrates the overall bathymetric trend of these ratios and the derived paleobathymetric equation.

Sediment analysis in Table 4 revealed a clear trend from coarse sand in transitional environments to predominantly silt in deeper environments. The %P showed a systematic exponential increase with depth, from being nearly absent in transitional environments (0 %–0.6 %) to showing very high values in lower bathyal environments (up to 97 %). Regression analysis of %P versus depth yielded a strong exponential relationship (r2=0.86; Fig. 3), which can be expressed using the exponential-growth model.

Benthic species richness peaked in middle neritic environments, with the maximum of 47 species per sample recorded at 84 m depth (sample no. IKN-03) before declining in deeper waters (Table 3; Fig. 4). In contrast, planktonic species richness increased continuously with depth from 0 to 2 species in transitional environments to a maximum of 17–23 species in lower bathyal settings (Table 3; Fig. 4).

Hierarchical cluster analysis using Bray–Curtis dissimilarity with average linkage revealed a clear two-level hierarchical structure (Fig. 5). Prior to clustering, samples were ordered by water depth to respect the primary bathymetric gradient controlling foraminiferal distribution. This depth-ordered constraint prevents ecologically implausible groupings while allowing the algorithm to identify assemblage similarities within the bathymetric framework.

Cluster stability was assessed using cophenetic correlation, where r=0.64 indicates moderate preservation of the original dissimilarity structure, which is typical for complex ecological datasets spanning extreme environmental gradients (Legendre and Legendre, 2012). The ecological validity of the clustering is strongly supported by independent validation: (1) DCA axis 1 shows strong correlation with water depth (r2=0.82, p<0.001), independently confirming depth as the primary structuring gradient; (2) %P increases systematically within the cluster hierarchy (0.63 % in cluster I to >90 % in subcluster IId); (3) IndVal analysis identifies statistically significant indicator species (p<0.05) for each cluster; and (4) cross-tabulation reveals 89 % classification accuracy when comparing cluster assignments to bathymetric zones (Table 3). This multi-method validation demonstrates that the clusters are ecologically robust and paleoenvironmentally meaningful despite moderate dendrogram distortion.

The dendrogram separated samples into two main clusters (Fig. 5). Cluster I comprises exclusively transitional environment samples (P. Bembang and P. Parangireng at 0 m depth), characterized by extremely low %P (0 %–0.6 %) and coarse sand substrates. Cluster II encompasses all deeper samples (1.5–4327 m) and is further subdivided into four subclusters reflecting increasing depth and changing environmental controls.

Subcluster IIa represents the largest and most heterogeneous grouping, containing 18 samples spanning inner neritic to outer neritic (1.5–168 m depth). This subcluster includes all inner neritic samples on both fine-sand (SSTH1, SSTH2) and silt substrates (sem-A9, sem-A21, sem-A11, sem-A6, sem-A5), all middle neritic samples (SSTH8, SS10, SS-09, IKN-01, IKN-03, CLP-03), and all outer neritic samples (GC-05, CLP-05, CLP-06, CLP-07, IKN-04). Despite spanning three bathymetric zones, these samples share compositional similarities reflecting the dominance of shelf environmental conditions with %P ranging from 2.5 % to 79 % (Table 3). The grouping of different neritic zones into a single subcluster reflects the gradual nature of environmental change across the continental shelf, where substrate characteristics and hydrodynamic conditions exert strong control on assemblage composition alongside depth.

Subclusters IIb and IIc each contain single isolated upper bathyal samples (IKN-05 at 219 m and Sim-ST14 at 376 m, respectively). The separation of these samples into distinct subclusters, despite belonging to the same bathymetric zone as some samples in subcluster IId, likely reflects regional oceanographic differences or local environmental heterogeneity such as substrate composition, organic matter flux, or oxygen availability. These isolated positions in the dendrogram highlight the importance of local environmental factors in shaping foraminiferal assemblages beyond simple depth control.

Subcluster IId contains 11 samples spanning upper bathyal to lower bathyal environments (750–4327 m depth), including the remaining upper bathyal samples (Sum ST-06, Sum ST-01, Sum ST-12, Sum ST-09) and all lower bathyal samples (Sum ST-13, Sum ST-11, OS-10, Sim-ST02, Sim-ST08, Sum ST-08, OS-07). This grouping reflects the transition to deep-water assemblages characterized by high %P (81 %–97 %); increased influence of biogeochemical factors (oxygen-minimum zones, organic matter flux); and dominance of characteristic deep-water taxa such as Uvigerina, Bulimina, and Globobulimina.

DCA results (Fig. 6) confirmed depth as the primary control on foraminiferal distribution. Axis 1 (eigenvalue = 0.675) explained the majority of variance and showed a strong correlation with water depth. Samples are distributed from right to left along axis 1, with transitional and shallow environments positioned on the right and bathyal environments on the left. Axis 2 (eigenvalue = 0.180) represents secondary environmental factors, likely reflecting substrate variability, organic carbon supply, and oxygen conditions.

The bathymetric distribution framework established in the study (Table 5) demonstrates a clear ecological stratification of foraminiferal assemblages across depth zones. This zonation pattern reflects adaptations to changing environmental parameters with increasing depth and has important implications for paleoenvironmental interpretations in Indonesian seas specifically.

In the transitional environment, the occurrence of Amphistegina and Sphaerogypsina (Fig. 7) carries notable ecological implications as these organisms are characteristically displaced from their original neritic habitats through hydrodynamic transport mechanisms including wave action and current systems (Hohenegger, 2004). Their presence in coarse-sand substrates indicates high-energy environments influenced by strong currents. Renema and Troelstra (2001) similarly observed the dominance of Amphistegina in high-energy environments exhibiting effective water circulation patterns throughout various Indonesian coastal systems. The extremely low %P (0 %–0.6 %) confirms the shallow-water nature of this environment, while the limited species richness (21 benthic species, 1 planktonic species) reflects the physically demanding conditions that limit the number of adapted species. The presence of robust, stress-tolerant foraminifera adaptable to physical disturbance is characteristic of such environments (Murray, 2006). The assemblage represents transported fauna from adjacent inner neritic habitats rather than an in situ community, a critical consideration for paleoenvironmental interpretation. Active hydrodynamic winnowing in this zone preferentially removes fine sediment and less robust tests, concentrating larger, more resistant forms like Amphistegina and Calcarina (Murray, 2006).

The inner neritic environment exhibits pronounced community differentiation based on substrate composition, featuring markedly discrete assemblages between fine-sand- and silt-dominated bottom environments. Fine-sand habitats support predominant populations of Ammonia sp., Asterorotalia trispinosa, and Elphidium hispidulum (Fig. 8). These assemblages reflect moderate energy conditions with good water circulation, consistent with the ecological preferences of these taxa as described by Murray (2006), who noted that Asterorotalia typically inhabits well-oxygenated environments with moderate currents, while Elphidium and Ammonia are commonly found in nearshore settings with active water exchange and moderate wave energy (Hayward et al., 2004). In contrast, silt substrates harbor significant populations of agglutinated forms like Haplophragmoides sp. and also Elphidium (Fig. 9), which indicates lower energy settings with higher organic content. According to Murray (2006), agglutinated foraminifera such as Haplophragmoides are typically associated with fine-grained sediments and environments with reduced water circulation, where organic matter accumulation is enhanced.

Beyond the primary-characteristic assemblages shown in Figs. 8 and 9, secondary indicators identified through single-method analysis provide additional insights into the environmental complexity of the inner neritic zone. For fine-sand substrates (Fig. 8), secondary indicators include Peneroplis, Adelosina, Operculina, Baggina, Cellanthus, Spiroloculina, Triloculina, Haplophragmoides, and Ammobaculites at the genus level, with species-level secondary indicators encompassing various Quinqueloculina species (Q. parkeri, Q. strigillata, Q. seminulum, Q. lamarckiana, Q. philippinensis, Q. crassicarinata, Q. angulariformis), Adelosina laevigata, Peneroplis species (P. pertusus, P. carinatus, P. planatus), Ammonia gaimardii, Elphidium neosimplex, Operculina species (Operculina sp., O. philippinensis), Spiroloculina species (S. depressa, S. scrobiculata), Baggina indica, Cellanthus craticulatus, Triloculina trigonula, and Elphidium crispum, adapted to moderate-energy sandy environments. For silt substrates (Fig. 10), secondary indicators include Nonion, Edentostomina, Peneroplis, Ammonia, Asterorotalia, and Ammobaculites at the genus level, with corresponding species-level indicators such as Ammonia beccarii, Elphidium (Elphidium sp. E. neosimplex, E. crispum), Haplophragmoides sp., Nonion boueanum, Quinqueloculina (Quinqueloculina sp. (Q. seminulum and Q. lamarckiana), Spiroloculina scrobiculata, Edentostomina cultrata, Peneroplis sp. (P. carinatus and P. planatus), and Ammobaculites agglutinans, reflecting adaptation to lower-energy, fine-grained conditions.

This specific substrate distribution within the same depth zone highlights the importance of substrate as a primary controlling factor in shallow marine environments, as demonstrated by Jorissen et al. (2007), who emphasized the fundamental role of sediment texture in determining foraminiferal microhabitat preferences. The increasing but still relatively low %P (3 %–19 %) reflects the limited influence of open marine conditions, which is typical for inner neritic environments according to Van der Zwaan et al. (1990). The increase in benthic species richness compared to transitional environments indicates more stable conditions (Murray, 2006), while the limited planktonic species richness reflects the nearshore position of this zone, where reduced oceanic influence restricts planktonic assemblage development (Bé and Tolderlund, 1971).

The middle neritic environment hosts assemblages dominated by Quinqueloculina seminulum and Pseudorotalia sp. (Fig. 10), indicating more stable marine conditions. These particular taxonomic groups flourish in settings characterized by normal marine salinity and lower energy settings (Murray, 2006). A noticeable sedimentary gradient is observed, transitioning from fine sand in shallower areas to predominantly silt in deeper areas, reflecting decreasing energy conditions with depth (Reading, 1996). This zone represents an ecological optimum for benthic foraminifera, with the highest species richness (120 species) observed across all environments. The moderate %P (12 %–54 %) reflects the increasing influence of open marine conditions (Van der Zwaan et al., 1990), while the increasing planktonic species richness with predominantly globular forms indicates a transitional position between coastal and fully oceanic settings.

Beyond the primary indicators, secondary-characteristic taxa (Fig. 10) for the middle neritic environment include Bombulina, Discorbis, Nonionella, Operculina, Cibicides, Ammonia, Brizalina, and Elphidium at the genus level. Species-level secondary indicators comprise Discorbis sp., Bombulina echinata, Nonionella simplex, Cibicides sp., Operculina philippinensis, Ammonia beccarii, Operculina sp., Quinqueloculina granulocostata, Triloculina tricarinata, and Nonion scaphum. These secondary indicators reinforce the interpretation of stable marine conditions, with normal salinity and good water circulation characteristics.

As depth increases to the outer neritic environment, the assemblage is characterized by Brizalina aenariensis, Gyroidina broeckhiana, and Cibicides sp. (Fig. 11). This transition marks an important ecological threshold where dissolved oxygen concentration begins to emerge as a dominant controlling factor, consistently with the TROX model proposed by Jorissen et al. (1995). The concurrent abundance of Brizalina species, recognized for their adaptation to low-oxygen conditions (Jorissen et al., 1995), alongside Cibicides species, which require higher oxygen levels (Murray, 2006), suggests the presence of diverse microhabitats within this zone. The marked increase in the %P (56 %–79 %) suggests a strong influence of open marine conditions, consistently with Van der Zwaan et al. (1990). In contrast, the slight decline in benthic species richness relative to middle neritic settings points to growing environmental limitations with increasing depth (Murray, 2006). Meanwhile, the continued rise in planktonic species richness, especially the greater presence of carinate forms, further reflects the increasingly oceanic nature of this zone (Schiebel and Hemleben, 2017).

Beyond the primary indicators, secondary-characteristic taxa shown in Fig. 11 for the outer neritic environment include Nonion, Neouvigerina, Virgulina, and Amphistegina at the genus level. Species-level secondary indicators comprise Nonion scaphum, Lagena sulcata, Virgulina pauciloculata, Quinqueloculina crassicarinata, Brizalina earlandi, Brizalina striatula, Textularia agglutinans, Amphistegina papillosa, and Brizalina robusta. The presence of these secondary indicators supports the interpretations of variable oxygen conditions and increased turbidity that characterize the transition from neritic to deeper environments, reflecting the complex microhabitat species richness within this zone.

The upper bathyal fauna is a diversified association (Fig. 12) suited to silty environments and varying oxygen conditions with increased depth. Infaunal species such as Bulimina marginata and Uvigerina asperula have often been recorded in low-oxygen environments (Jorissen et al., 2007), while the presence of Hyalinea balthica serves as a reliable bathymetric indicator for this depth zone (Van der Zwaan and Jorissen, 1991). The high %P (73 %–89 %) suggests the dominance of open marine conditions (Van der Zwaan et al., 1990). The relatively stable benthic species richness indicates a well-adapted community suited to upper continental slope conditions (Gooday, 2003), while the shift in planktonic assemblages toward carinate forms reflects adaptation to deeper, more oceanic waters (Schiebel and Hemleben, 2017). The increasing importance of infaunal taxa in this zone reflects adaptation to exploit organic matter within sediments under reduced oxygen conditions at the sediment–water interface.

The upper bathyal environment shows extensive secondary-characteristic assemblages (Fig. 12), including Oolina, Hyalinea, Cassidulina, Anomalina, Sigmoilopsis, Saracenaria, Pullenia, Brizalina, Baggina, Hoeglundina, Neouvigerina, and Lagena at the genus level. Species-level secondary indicators encompass Oolina globosa, Pyrgo sp., Gyroidina neosoldanii, Saracenaria italica, Textularia sp., Lenticulina sp., Sigmoilopsis schlumbergeri, Ammonia supera, Pullenia subcarinata, Nodosaria sp., Baggina indica, Neouvigerina sp., Bulimina sp., Brizalina aenariensis. Brizalina sp., Neouvigerina ampullacea, Hoeglundina elegans, Uvigerina peregrina, Bolivinita subangularis, Quinqueloculina seminulum, Siphogenerina columellaris, and Ammobaculites agglutinans. This diverse array of secondary indicators reflects the complex environmental gradients present in upper-continental-slope settings, where multiple factors including oxygen availability, organic matter flux, and water mass characteristics interact to shape foraminiferal communities.

In the lower bathyal environment, the dominance of Lenticulina orbicularis, Uvigerina peregrina, and Globobulimina pacifica (Fig. 13) suggests a community highly adapted to deep-sea conditions. Notably, Uvigerina peregrina indicates increased organic matter flux at bathyal depths (Fontanier et al., 2002), while Globobulimina pacifica is adapted to low-oxygen conditions (Jorissen et al., 1995). According to Kaiho (1994), Globobulimina pacifica is particularly tolerant of dysoxic conditions and can penetrate deep into sediments as an infaunal species. The very high %P (90 %–97 %) represents fully oceanic conditions (Van der Zwaan et al., 1990), while the decrease in benthic species richness reflects the increasing environmental constraints of deep-sea habitats (Gooday, 2003). The maximum planktonic species richness, featuring carinate forms and species with cortexes, indicates the fully oceanic character of this zone (Schiebel and Hemleben, 2017).

Secondary indicators of the lower bathyal environment (Fig. 13) include an extensive suite of deep-water adapted genera: Uvigerina, Lenticulina, Cibicides, Amphicoryna, Brizalina, Anomalina, Bulimina, Bolivinita, Dentalina, Heterolepa, Hoeglundina, Cancris, Vulvulina, Nodosaria, Astacolus, Planularia, Martinotiella, and Oridorsalis. At the species level, secondary indicators include Lagena hispidula, Amphicoryna scalaris, Bolivinita quadrilatera, Planulina ariminensis, Brizalina robusta, Bulimina sp., and Brizalina sp. The species richness of secondary indicators in this deepest environment reflects the specialized adaptations required for life in the abyssal–bathyal transition zone, where extreme depth, low-oxygen conditions, and variable organic matter flux create highly selective environmental pressures that determine the composition of deep-sea foraminiferal communities.

A clear transition in controlling environmental factors emerges across the bathymetric gradient. In shallower environments (transitional to middle neritic), physical factors dominate the distribution patterns.

Current and wave energies strongly influence the transitional environment, evidenced by coarse substrates; active winnowing; and the prevalence of robust, transported foraminifera. Hydrodynamic processes including wave action, long-shore currents, and storm events control assemblage composition through selective transport and destruction of tests.

The recognition of distinct indicator species for each bathymetric zone provides a robust framework for paleoenvironmental interpretations, specifically in Indonesian seas and potentially being applicable to the broader Indo-Pacific region. The well-defined depth-related zonation documented in this study can strengthen the use of fossil assemblages for paleobathymetric inferences, consistently with similar approaches developed by Van der Zwaan et al. (1990) and Murray (2006) for other oceanic regions but calibrated for tropical Indo-Pacific conditions.

The clear separation of cluster I (transitional environment) from all deeper samples demonstrates that transported shallow-water assemblages dominated by Amphistegina and Sphaerogypsina can be reliably distinguished from in situ deeper assemblages. This is particularly important for interpreting fossil deposits where hydrodynamic transport may complicate paleobathymetric reconstruction. The extremely low %P (0 %–0.6 %) and coarse substrates provide additional independent evidence for transitional settings.

The grouping of inner, middle, and outer neritic samples into the single large subcluster IIa, despite covering 1.5–168 m depth, highlights both the utility and limitations of foraminiferal assemblages for fine-scale depth discrimination on continental shelves. This clustering pattern indicates the following.

Substrate variability exerts a strong control on shelf assemblages, sometimes comparable to or exceeding the influence of depth alone. The inclusion of both fine-sand and silty inner neritic samples within the same subcluster demonstrates that different microhabitats at similar depths can host compositionally similar assemblages when controlled for substrate type. It must be noted that our samples were taken from a wide range of locations, and information on local primary productivity, dissolved oxygen, pH, and temperature was not gathered for this study.

Gradual environmental transitions across the shelf create overlapping assemblage characteristics. The progressive increase in %P from inner neritic (2.5 %–19 %) through middle neritic (12 %–54 %) to outer neritic (56 %–79 %) (Table 3) provides a more continuous depth proxy than discrete assemblage breaks.

Multiple lines of evidence should be employed for paleobathymetric reconstruction in shelf settings. While foraminiferal assemblages provide valuable constraints, integration with sedimentological data (grain size, sedimentary structures), taphonomic indicators (fragmentation, abrasion), and quantitative %P yields more robust interpretations.

The indicator species identified through IndVal analysis (Tables S2–S3) remain valuable despite the broad clustering of neritic samples. Taxa such as Asterorotalia and Elphidium in inner neritic environments, Quinqueloculina and Pseudorotalia in middle neritic environments, and Brizalina and Gyroidina in outer neritic environments show statistically significant depth preferences that can distinguish between these environments when considered alongside complementary evidence.

The heterogeneous clustering of upper bathyal samples, with some forming isolated subclusters (IIb, IIc) and others grouping with lower bathyal samples (subcluster IId), underscores the importance of regional oceanographic variability in deep-water environments. This pattern suggests that local factors including oxygen-minimum zone position and intensity (which varies spatially across Indonesian basins, e.g., Sumba Strait vs. Simeulue region), organic matter flux patterns (influenced by surface productivity and current systems), and water mass characteristics (particularly where Indonesian Throughflow creates complex intermediate and deep water circulation) exert a significant control on upper bathyal assemblages, potentially overriding simple depth-related patterns. Paleoenvironmental reconstructions in bathyal settings must therefore consider regional oceanographic context and not rely solely on depth indicator species.

The clear grouping of lower bathyal samples (subcluster IId) with high %P (>90 %) and characteristic taxa (Globobulimina, Planulina, Uvigerina peregrina) demonstrates that deep-water assemblages (>1000 m) are more readily distinguished than shelf assemblages. This likely reflects the stronger environmental gradients (oxygen, pressure, temperature, organic matter availability) that characterize the transition from upper continental slopes to deep basins.

From a methodological perspective, the classification accuracy achieved through hierarchical cluster analysis validates the bathymetric zonation scheme adapted for Indonesian seas. However, primarily involving samples near bathymetric boundaries or in transitional oceanographic settings emphasizes the probabilistic rather than deterministic nature of foraminiferal paleobathymetry. Fossil assemblages should be assigned to depth zones with appropriate uncertainty estimates, and, where possible, multiple samples from stratigraphic sequences should be analyzed to establish robust depth trends rather than relying on single-sample interpretations.

The integration of this regionally calibrated indicator species framework with the quantitative %P–depth relationship (Fig. 3; r2=0.86) provides complementary approaches to paleobathymetric reconstruction: the %P offers continuous depth estimates with defined uncertainty, while indicator species provide qualitative environmental context (substrate, oxygen, energy) that enriches paleoenvironmental interpretation beyond simple water depth.

Finally, the shift from physical environmental controls (wave energy, substrate, salinity) in shallow waters to biogeochemical controls (oxygen, organic matter flux) in deep waters, as evidenced by both cluster analysis patterns and DCA ordination (Fig. 6), is consistent with the applicability of the TROX model (Jorissen et al., 1995) to tropical Indo-Pacific settings. This has important implications for interpreting ancient depositional systems as it suggests that similar depth-related transitions in environmental controls likely operated in past tropical seas. Fossil assemblages can therefore be interpreted not only for water depth but also for the dominant environmental processes (physical vs. biogeochemical) controlling benthic ecosystems at the time of deposition.