Temporal dynamics and network drivers of coral reef structural-functional relationships in the Nansha Islands, South China Sea - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Eco Environ Health . 2026 Mar 13;5(2):100232. doi: 10.1016/j.eehl.2026.100232 Search in PMC Search in PubMed View in NLM Catalog Add to search Temporal dynamics and network drivers of coral reef structural-functional relationships in the Nansha Islands, South China Sea Yanyan Zhou Yanyan Zhou a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China Find articles by Yanyan Zhou a , Xianzhi Lin Xianzhi Lin a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China Find articles by Xianzhi Lin a , Haoxuan Cheng Haoxuan Cheng a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China b University of Chinese Academy of Science, Beijing 100049, China Find articles by Haoxuan Cheng a, b , Shuo Zhai Shuo Zhai a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China b University of Chinese Academy of Science, Beijing 100049, China Find articles by Shuo Zhai a, b , Sen Du Sen Du a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China Find articles by Sen Du a , Lizhao Chen Lizhao Chen a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China Find articles by Lizhao Chen a , Li Zhang Li Zhang a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China c Key Laboratory of Tropical Marine Biotechnology of Hainan Province, Sanya Institute of Ocean Eco-Environmental Engineering, Sanya 572000, China Find articles by Li Zhang a, c, ⁎ Author information Article notes Copyright and License information a State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China b University of Chinese Academy of Science, Beijing 100049, China c Key Laboratory of Tropical Marine Biotechnology of Hainan Province, Sanya Institute of Ocean Eco-Environmental Engineering, Sanya 572000, China ⁎ Corresponding author. State Key Laboratory of Tropical Oceanography, Guangdong Provincial Key Laboratory of Applied Marine Biology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China. [email protected] Received 2025 May 20; Revised 2025 Aug 2; Accepted 2026 Mar 11; Collection date 2026 Jun. © 2026 The Authors This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). PMC Copyright notice PMCID: PMC13090306 PMID: 42004007 Abstract For the conservation of coral reefs through resilience-based management, elucidating the evolutionary mechanisms underlying ecosystem functionality and stability emerges as a critical scientific priority. We employed energy flux network analysis to investigate the functional-structural dynamics of the Nansha Islands coral reef ecosystem after the El Niño event. Based on fieldwork (2016−2020), an array of food-web metrics was evaluated at both the entire ecosystem level and the individual taxa level. Simulated and δ 15 N-based trophic levels agreed ( r 2 = 0.83, P < 0.0001), validating model robustness. As the coral coverage increased from 14.1% to 42.1%, there was a corresponding increase in total system throughput (TST), net primary production (NPP), transfer efficiency (TE), and ascendency/capacity (A/C) by factors of 1.10, 1.43, 0.52, and 0.22, respectively, while the overhead/capacity (O/C) decreased by 0.11, indicating an increase in ecosystem maturity. Despite these changes, the relatively low connectance index (CI, 0.15) and omnivory index (OI, 0.18) suggested that the system was still at an immature stage. A low A/C (40.21%) coupled with a high O/C (59.79%) indicated the higher resilience of the ecosystem, with more than half of the energy allocated to buffering El Niño. By integrating informational flexibility with throughput, fish primarily enhance resilience, whereas benthic consumers mainly contribute to increased throughput. Acanthuridae appeared to play a crucial role in actively boosting the resilience of the ecosystem. Overall, the Nansha Islands coral reef ecosystem is evolving towards a more stable and mature state, warranting continued monitoring as a potential model system for ecological restoration. Keywords: South China sea, Coral reefs ecosystem, Resilience, Entropy, Trophic levels Graphical abstract Open in a new tab Highlights • Nansha coral reef ecosystem exhibit significant resistance to El Niño disturbance. • Order-entropy equilibrium confers dual reef resilience: stability and adaptation. • New framework quantifies key taxa driving resilience. 1. Introduction Coral reef ecosystems are recognized as some of the most productive and valuable ecological systems, offering significant ecological, social, and economic benefits. Alarmingly, 14% of global coral reefs were lost between 2009 and 2018, with over 90% currently at risk due to anthropogenic and global stressors [ [1] , [2] , [3] ]. To ensure the preservation and restoration of coral reef resilience, it is imperative to investigate how ecosystem structure and function adapt following stressors or disturbances [ 4 ]. While previous studies have identified community stability and high species diversity as potential indicators of resilience and stability, these concepts remain subjects of ongoing debate [ [5] , [6] , [7] ]. Moreover, these studies often prioritize taxonomic composition, treating energy flows and ecosystem processes as secondary and derivative. Field-based measurements of biomass and specimen enumeration are methodologically simpler but insufficient for accurately quantifying material flux dynamics within ecosystems [ 8 ]. Energy and material flows represent essential processes driving ecosystem functionality [ 9 ]. Ecological Network Analysis (ENA) provides a robust methodological framework for quantifying energy and material flows within ecosystems [ 10 ]. It has been widely applied to assess the effects of perturbations and evaluate ecosystem integrity [ 8 ]. The Ecopath model, one of the most widely utilized tools for ecosystem network analysis in marine, estuarine, and other aquatic systems [ [11] , [12] , [13] ], offers detailed insights into trophic interactions and the structural and functional organization of ecosystems [ 14 ]. Previous studies on ecological network analysis suggest that ecosystem dynamics are heavily influenced by flow configurations, with resilience and stability primarily arising from redundancy in energy flows [ [15] , [16] , [17] ]. Temporal patterns in ecosystem indices, such as total system throughput (TST), average mutual information (AMI), and ascendency (A), can reflect system development and stability; declining values signal degradation, while increasing values indicate improved resilience and functionality [ [18] , [19] , [20] ]. However, establishing clear links between ecosystem properties and responses to external perturbations remains a significant challenge [ 21 ]. From a thermodynamic perspective, ecosystems balance the competing forces of order and entropy. Systems with greater energy reserves exhibit higher entropy and enhanced resilience, while energy depletion leads to increased order, stability, and maturation [ 22 ]. Metrics such as the A/C (ascendency/development capacity) and O/C (overhead/development capacity) ratios enable quantitative assessments of ecosystem order and entropy, allowing for estimations of resilience through entropy variations expressed in trophic energy flows [ 23 ]. Furthermore, Lewis et al. [ 24 ] introduced the redundancy/ascendency index (R/A), which explores relationships between energy flow in taxon-specific subcomponents and overall ecosystem resilience. They observed that short-term energy flow shifts mediated by specific taxa provided sufficient energy and flexibility within the food web to enhance resilience throughout the entire ecosystem. These energy flow-based insights can advance our understanding of mechanisms underlying ecosystem stability and resilience in response to stressors. This study aims to investigate the energy-driven dynamics of coral reef ecosystems to address the pivotal question: How can internal processes be quantified to evaluate ecosystem function, resilience, and stability? Coral reefs in China are predominantly distributed in tropical waters south of 20.5° N, particularly around the Nansha Islands, which are noted for their exceptional marine biodiversity [ 25 ]. Previous studies have demonstrated that the distribution, taxonomic composition, and biodiversity of coral reef organisms correlate with disturbance regimes [ 26 , 27 ]. Some studies showed that external disturbances (such as heatwaves) caused episodic coral mortality in China, yet there were intervals during which coral reefs exhibited opportunities for recovery [ 28 , 29 ]. Other recent studies found that coral reef fishes tended to miniaturize, with simplified composition and function, which was potentially associated with coral mortality [ 30 , 31 ]. Despite these findings, studies addressing the functional structure of coral reef ecosystems remain limited. Investigations by Zhang et al. [ 32 ] and Hong et al. [ 33 ] focused on macroscopic ecosystem indices such as TST, A/C, Finn’s cycling index, and system omnivory index but overlooked energy flow dynamics within specific ecosystem components. This lack of analysis hampers a deeper understanding of how coral reef resilience, stability, and function evolve in response to stressors and disturbances. Nansha Islands, situated in the tropical South China Sea (western Pacific), comprise over 200 coral atolls, cays, and shoals, forming the region’s largest coral reef system. The ecosystem supports high biodiversity, including 558 reef-associated fish species and 324 hermatypic coral species, and constitutes China’s largest tropical marine fishery resource [ 34 ]. Additionally, coral reef ecosystems in the Nansha Islands act as natural “coastal buffers”. It was reported that stable coral reefs could reduce wave energy by over 70%, mitigating the impact of typhoons, storm surges, and coastal erosion on nearby islands [ 35 ]. Thus, the stability of the Nansha Islands ecosystem is crucial to ecological security, economic development, and national territorial security. The ecosystem has been severely influenced by climate-driven disturbances, particularly during the 2015−2016 El Niño event, which triggered widespread coral mortality throughout the archipelago [ 36 ]. Subsequent monitoring has documented notable recovery in key ecological parameters, including coral cover regeneration, algal community dynamics, and parrotfish population stabilization, even in the absence of significant thermal stress mitigation measures [ 37 ]. Despite these observations, the fundamental structural and functional mechanisms underlying this ecological resilience remain poorly understood. To address this knowledge gap, our study utilized Ecopath modeling to quantitatively assess trophodynamic changes at a representative reef site during the critical post-disturbance recovery period (2016−2020). By analyzing system-wide and taxon-specific carbon flows, we characterized the temporal dynamics of the reef’s structure and function after the El Niño event. Our objectives were: (1) to evaluate the development process and current status of the ecosystem; (2) to assess taxon-specific functional roles and identify key contributors to resilience. This work advanced the understanding of Nansha’s coral reef dynamics and provided a critical theoretical foundation for resilience assessment and restoration strategy development in tropical coral reefs globally, particularly for immature systems undergoing recovery. 2. Materials and methods 2.1. Study area sampling The study area is in the southern part of Meiji Reef in the Nansha Islands in the South China Sea, located at 9°55′ N, 115°32′ E [ 25 ]. Fishing, tourism, and commercial activities are prohibited on Meiji Reef. Sampling surveys were conducted in May 2016, May 2017, July 2019, and May 2020 through underwater visual censuses by SCUBA divers. Four sampling sites were selected to survey species ( Fig. 1 ). Three water depth strata (3 m, 8 m, and 15 m) were sampled per site, each comprising 1 transect and 30 quadrats. Fish species, abundance, and size were estimated through belt-transects (60 m × 5 m, n = 3) at each site. The same transects were utilized to record the species, size, and abundance of coral and macrobenthos, such as starfish, sea cucumbers, Tridacna , and lobsters. For the benthos, including macroalgae, turf algae, and small invertebrates, the number and weight of the benthos were estimated using a sampling frame (0.5 m × 0.5 m, n = 90) for each site. The sampling frame was placed to cover the surface of the seabed, and the area within the frame was considered a sample plot. Within the sample plot, surface animals and macroalgae (excluding coralline algae) were picked, and the reef rocks and sediments were dug out. Turf algae were collected by haphazardly placing a 10-cm diameter PVC ring on the reef surface and scraping material from the substratum using a stainless-steel scraper. Upon scraping the material from the substratum, the hand water pump was activated simultaneously and kept inactive at all other times. After collecting each sample underwater, each PVC container was promptly sealed in an individual sterilized plastic bag. The samples were filtered onto a 20-μm nylon mesh using a diaphragm vacuum pump. The reef rocks were crushed and sorted to reveal the animals living inside them. Zooplankton were collected via vertical tows using a 160-μm mesh plankton net (bolting silk), with six replicate samples obtained per site. Six surface water samples were collected at each station using a water sampler, pre-filtered through a 160-μm mesh, and subsequently filtered onto 0.7-μm glass fiber filters to collect phytoplankton. Fig. 1. Open in a new tab Investigation sites of Nansha coral reef ecosystem during the period 2016–2020. 2.2. Model design Trophic models of the study area were constructed using the Ecopath with Ecosim modeling software 6.5 (EwE) ( http://www.ecopath.org ) [ 14 ]. Ecopath models were organized into 37−40 functional groups, based on biological traits (growth/mortality rates) and feeding patterns, which comprehensively represented energy flow dynamics in the ecosystem [ 38 ]. The models contained 22−25 fish functional groups (representing 103 species), 11 benthic invertebrate groups (representing 86 species), 3 primary producer groups (phytoplankton, turf algae, and 21 macroalgal species), and 1 zooplankton and detritus group ( Appendix B ). Ecopath models were based on a mass-balance system between biomass gains through production and losses because of predation, metabolism, and exports. The general equation can be expressed as follows: B i × ( P B ) i × EE i = ∑ j = 1 j B j × ( Q B ) i × D C i j + Y i + B A i + E i (1) Where, B i and B j were the biomass of prey i and predator j in t/km 2 ; P i was the production rate for the group i in t/(km 2 ·year); Q j was the consumption for the group j in t/(km 2 ·year); ( P/B ) i was the production/biomass ratio of i in year −1 , which was equivalent to the coefficient of total mortality M under steady-state conditions [ 39 ]; ( Q/B ) j was the ratio of consumption to biomass for the given predator j in year −1 ; EE i was the trophic efficiency defined as the proportion of the production of i that was utilized in the system; DC ij was the contribution of i to the diet of the predator j ; Y i was the total fishery catch of i in t/(km 2 ·year); E i was the net migration rate of i (emigration minus immigration) in t/(km 2 ·year); and BA i was the biomass accumulation rate for i in t/(km 2 ·year) [ 14 ]. To summarize some of the relevant results of ecological network analysis, a theoretical framework was used to quantify energy flows in trophic networks. The ratio A/C provided a measure of ecosystem order, and ratio O/C described the ability of the system to resist disturbances [ 21 , 22 ]. A represented the organized power of the system in bit t/(km 2 ·year) and was described as follows: A = ∑ i , j T i j × log ( T i j T s , s T i , s T s , j ) (2) Where, i and j denoted the prey and predator, respectively; s was the sum of flows of prey or predators; T i,s represented the flows from one prey to all their predators in bit t/(km 2 ·year); T s,j the consumption of a predator over all its prey in bit t/(km 2 ·year); and T s,s the total sum of flows over prey and predators in bit t/(km 2 ·year). O represented the energy in reserve of the network in bit t/(km 2 ·year) and was used to estimate its ability to withstand perturbations. It could be expressed as follows. O = - ∑ i , j T i j × log T i j 2 T i , s T s , j (3) C quantified the upper limit of A in bit t/(km 2 ·year) and was estimated from the difference between A and O . It could be described as follows. C = A + O = - ∑ i , j T i j × log T i j T s , s (4) A / C was designated as a measure of ecosystem order, whereas O / C quantified ecosystem entropy. A negative deviation of the entropy from its initial state ( O / C t < O / C 0 ) indicated a gain in order and, consequently, a trend toward ecosystem improvement. δ ( O / C ) = O / C t − O / C 0 (5) Where, δ ( O / C ) denoted the value of relative entropy; and O / C t and O / C 0 were the values of entropy at given time t and initial time 0 , respectively. The value of entropy in 2016 was considered as the initial value ( O / C 0 ). R / A provided the flexibility of internal interactions per unit of flow relative to the amount of constraint on complexity [ 18 ]. Redundancy ( R ) represented the portion of the overhead associated with the compartmental interactions and could be expressed as follows. R = O - O r e s (6) O r e s = R e s × log 2 R e s T (7) Where, O res represented the overhead attributed to respiration in t/(km 2 ·year); R es represented the respiration of each taxon or function group in t/(km 2 ·year); and T represented throughput as the sum of flows for each taxon or function group in t/(km 2 ·year). 2.3. Data sources For fish, the abundance and size were estimated through belt-transects (60 m × 5 m, n = 3) at each site. The length measurements were then used to calculate biomass based on a length-weight relationship [ 40 ]. Coral surface area was determined assuming that massive coral heads had a cylindrical shape. Coral tissue was extracted from the skeleton using a Water Pik (jet-tooth cleaner) with distilled water. Subsequently, the samples were freeze-dried, and the dry weight (DW) per unit area was measured. Coral biomass was calculated as follows. DW = Surface area × c × WW Unit area (8) Where, c is the conversion factor from DW to wet weight (WW). For corals, the DW:WW ratio is 1:5.2 [ 41 ]. For the benthos, including macroalgae, turf algae, and invertebrates, samples were collected from the sampling frame for each site. Then, all benthos were rinsed and weighed [ 42 ]. For zooplankton, samples were collected by vertical hauls from the bottom to the surface for each site and then freeze-dried to a constant weight. The conversion ratio from DW to WW was 10:1 [ 43 ]. The biomass was estimated by integrating site-specific depth and excluding the detritus. For phytoplankton, the biomass was estimated from chlorophyll-α (Chlα) concentrations measured by fluorimetry. The biomass was estimated using a Carbon / Chlα ratio of 84 [ 44 ]. For the detritus, the biomass was estimated using the following formula. log D = - 2.41 + 0.95 log P P + 0.863 log Z e u (9) Where, D was detritus biomass; PP was primary production in g C/(m 2 ·year); and Z eu was the euphotic zone in metres [ 45 ]. All the biomass estimations were standardized to wet weight (t/km 2 ). P / B and Q / B (year −1 ) for fish were estimated using the empirical formula [ 46 , 47 ]. P B = k 0.65 × L ∝ - 0.279 × T 0.43 (10) Where, k corresponded to the growth coefficient (year); L ∝ was an asymptotic length (cm); and T temp corresponded to water temperature [1000/(T + 273.15)] (°C). log Q B = 7.946 - 0.204 log W ∝ - 1.965 T t e m p + 0.083 A c a u + 0.532 h + 0.398 d (11) Where, W ∝ represented the asymptotic weight (g); A cau represented the caudal fin shape parameter (the ratio of the square of caudal fin height to the caudal fin area); h and d were dummy variables indicating the feeding category of the species: carnivore ( h = 0, d = 0), detritivore ( h = 0, d = 1), or herbivore ( h = 1, d = 0). P / B and Q / B for phytoplankton and zooplankton were estimated in situ by the methods described by Wu et al. [ 48 ] and Tan et al. [ 49 ]. P / B and Q / B for the benthos were obtained from published studies conducted in regions close to the study area, such as Xisha Islands, Wuzhizhou Island, Taiwan Island, and Beibu Bay, China [ [50] , [51] , [52] ]. The diet matrix ( Appendix B ) was reconstructed from stomach content analysis, N isotope analysis, published literature [ 32 , 53 ], and Fishbase ( https://www.fishbase.org ). 2.4. Model balancing and verification Before model balancing, PREBAL diagnostics were calculated ( Fig. S1 ), which were used to assess the coherence of the input parameters. In general, the values of B , P / B, and Q / B tended to decrease as the trophic level increased. For model balancing, the following criteria were verified: (1) all of the Ecotrophic Efficiencies ( EE ) of the compartments were <1.0 [ 54 ]; (2) P / Q ratio was lower than 1 for all compartments (optimal values between 0.1 and 0.3) [ 55 ]; (3) Respiration/Assimilation ( R es / A s ) was <1.0; (4) Production/Respiration ( P / R es ) was <1.0 ( Table S2 ). To verify the model, trophic levels estimated from Ecopath were compared with those calculated from N stable isotopes in situ ( Fig. S2 ). 2.5. Stable isotope analysis All samples (white muscle) were collected in study areas and then freeze-dried and ground into powder. δ 15 N was analyzed using an elemental analyzer (Flash2000, Thermo Fisher Scientific, Inc., Italy) connected to an isotope ratio mass spectrometer (Delta V advantage, Thermo Fisher Scientific, Inc., Germany) at the South China Sea Institute of Oceanography, Chinese Academy of Sciences, Guangzhou, China. The δ 15 N was calculated by the following formula. δ 15 N = N 15 / N 14 sample N 15 / N 14 standard - 1 × 10 3 (12) Where, δ 15 N isotope values were expressed as δ values per mil deviation (‰) from the standard reference materials Vienna Pee Dee Belemnite and atmospheric nitrogen, respectively. Protein (Casein) Standard (CatNo.B2155, Elemental Microanalysis Ltd., UK) was used as a certified reference material with an analytical precision of ±0.08‰ for δ 15 N values. The trophic level was estimated according to Post [ 56 ]. The calculation formula was as follows. T L = [ ( δ 15 N species − δ 15 N base ) / 3.4 ] + T L base (13) Where, δ 15 N species represented the value of the species; δ 15 N base represented the value of the baseline organisms; TL base was the TL of baseline organisms, used to calculate the TL s of other consumers in the ecosystem. In this study, Tridacna was identified as the baseline organism [ 56 , 57 ]. 2.6. Statistical analysis Correlation analysis was carried out using SPSS 20.0 software. Pearson’s simple correlation coefficient was used to measure the linear relationship between TLs simulated from Ecopath and determined from δ 15 N values. 3. Results 3.1. Physicochemical parameters The physicochemical parameters of seawater from 2016 to 2020 were summarized in Table S2 . Dissolved oxygen (DO) was 5.06−6.13 mg/L, pH was 8.01–8.37, turbidity was 0.39–0.99 FTU, salinity was 31.00‰–33.91‰, temperature was 29.69–33.86 °C, current velocity was 0.04−0.38 m/s, and suspended solids (SS) were 2.99–7.82 mg/L. Additionally, total inorganic nitrogen, reactive phosphate, and silicate were within the ranges of 0.13–0.42, 0.01–0.03, and 0.58–2.84 μmol/L, respectively. Apart from elevated concentrations of SS and silicate observed in 2017, the parameters exhibited minimal variation, indicating a relatively stable ecological environment from 2016 to 2020. 3.2. Biological parameters Coral coverage increased markedly from 14.13% (2016) to 42.10% (2020), with intermediate values of 15.03% (2017) and 38.59% (2019) ( Table S2 ). The coral reef ecosystem’s trophic model, comprising 37–40 functional groups (including 103 fish species, 107 benthic organism species, phytoplankton, turf algae, zooplankton, and detritus), comprehensively represented energy flow dynamics in the ecosystem. Annual variations occurred in taxon composition, notably through expansions in carnivorous fish populations ( Appendix B ). Likewise, the food web structure comprised 38 functional groups in 2016 and 2017, 37 functional groups in 2019, and 40 functional groups in 2020. Coral biomass increased by 205% and zoobenthos biomass by 26%, while macroalgae and turf algae biomass decreased by 53% and 43%, respectively, over time. Biomass of phytoplankton and zooplankton exhibited fluctuations over time but remained below the initial value in 2016 with decreases of 22% and 18%, respectively. For fish assemblages, biomass also exhibited fluctuations, with some taxa below the initial values (Acanthuridae, decrease of 46%) and others above them (Muraenidae and Haemulidae, increases of 320% and 200%) ( Table S1 and Fig. 2 ). Fig. 2. Open in a new tab Changes in biomass of fish (a), and benthos and plankton (b) in Nansha coral reef ecosystem as a function of year. 3.3. Model balancing and verification According to PREBAL diagnostics ( Fig. S1 ), estimates of B , P / B, and Q / B consistently decreased with increasing trophic levels across all periods; meanwhile, Q / B , EE , R es / A s , and P / R es were below 1 for all ecosystem compartments ( Table S1 ), indicating the input parameters were within acceptable ranges for model validation. Ecopath-derived trophic levels demonstrated significant concordance with δ 15 N-based estimates (linear regression: r 2 = 0.83, p < 0.0001; Fig. S2 ), validating both the accuracy of model input parameters, including biomass and dietary matrices, and reinforcing the methodological reliability of the Ecopath framework for elucidating trophic dynamics. 3.4. Trophic structure and energy flows Changes in the coral reef food web structure were observed across different years ( Fig. S3 and Table 1 ). The trophic levels estimated by Ecopath ranged from 1.00 (primary producers and detritus) to 4.04 (piscivorous fish). The mean trophic level in the food web was similar across different periods (2.61−2.68). Likewise, CI (0.15−0.17), OI (0.16−0.18), and Shannon diversity index (2.73−2.96) showed no change from 2016 to 2020. Finn’s mean path length reached its highest value in 2016 (2.58), with lower values in 2020 (2.27). Table 1. Network properties after mass balance process by Ecopath from 2016 to 2020. Network metric 2016 2017 2019 2020 Units Sum of all consumption (Q) 2478 2490 2147 2361 t/(km 2 ·year) Sum of all exports (E) 1873 2593 5524 7646 t/(km 2 ·year) Sum of all respiratory flows (R) 1786 1301 1203 1276 t/(km 2 ·year) Sum of all flows into detritus (D) 3319 3283 6771 8938 t/(km 2 ·year) Total system throughput (TST) 9455 9667 15644 20222 t/(km 2 ·year) Sum of all production (P) 4351 4418 7303 9556 t/(km 2 ·year) Net primary production (NPP) 3659 3727 6726 8921 t/(km 2 ·year) Net system production (NSP) 1873 2426 5523 7645 t/(km 2 ·year) Total biomass (excluding detritus) (TB) 145.6 146.1 159.9 165.8 t/km 2 TB/total throughput (TB/TT) 0.02 0.02 0.01 0.01 year −1 Total primary production/total respiration (TPP/TR) 2.05 2.86 5.59 6.99 TPP/TB 25.13 25.50 42.07 53.79 Connectance index (CI) 0.16 0.15 0.17 0.15 Omnivory index (OI) 0.18 0.16 0.16 0.18 Shannon diversity index 2.82 2.96 2.73 2.84 Finn’s mean path length 2.58 2.48 2.33 2.27 Ascendency (A) 11974 12556 18123 23640 bit t/(km 2 ·year) Overhead (O) 24441 25159 32757 35148 bit t/(km 2 ·year) Development capacity (C) 36416 37714 50877 58788 bit t/(km 2 ·year) Average mutual information (AMI) 2.19 1.84 2.24 2.72 A/C 32.88 33.29 35.62 40.21 % O/C 67.12 66.71 64.38 59.79 % Mean trophic levels 2.68 2.61 2.64 2.68 Mean transfer efficiency 6.58 7.55 8.72 10.03 % Open in a new tab Metrics such as TST, net system production (NSP), total biomass (excluding detritus) (TB), net primary production (NPP), transfer efficiency (TE), and A/C increased over time ( Table 1 ). TST increased by a factor of 1.10 [from 9455 to 2,0222 t/(km 2 ·year)], NSP increased by a factor of 3.08 [from 1873 to 7645 t/(km 2 ·year)], NPP increased by a factor of 1.43 [from 3659 to 8921 t/(km 2 ·year)], TE increased by a factor of 0.52 (from 6.58% to 10.02%), while A/C increased by a factor of 0.22 (from 32.88% to 40.21%). Temporal A/C trends for individual taxa or functional groups revealed four distinct patterns ( Fig. 3 ). Phytoplankton, coral, and Acanthuridae exhibited a clear trend of increasing order, while turf algae and macroalgae showed a loss of order. For fish, large crabs, small crabs, shrimps, gastropods, and echinoderms, the order fluctuated above the initial value, indicating a gain in order. Conversely, lobsters, bivalves, polychaetes, and zooplankton displayed fluctuating values below the initial value, indicating the loss of order. Relative order/entropy variance was computed across temporal cohorts using 2016 as the baseline. Temporal analysis revealed a shifting relationship between the relative change in mass and in order: while no significant correlation existed in 2017 ( p > 0.05), positive correlations emerged in both 2019 ( r = 0.38, p < 0.05) and 2020 ( r = 0.54, p < 0.01) ( Fig. S5 ). Fig. 4 delineated temporal shifts in relative entropy variance across taxonomic groups, revealing accelerating entropy reduction trajectories. The proportion of taxa exhibiting entropy decline escalated exponentially from 8% in 2017 to 38% in 2019 and 74% in 2020. Generalized linear modeling confirmed no trophic-level dependency, while the number of taxa showing entropy reduction increased progressively over time. Fig. 3. Open in a new tab Changes in the order (A/C) of fish (a), and benthos and plankton (b) in Nansha coral reef ecosystem as a function of year. Trajectories below the initial A/C value (2016) represent a loss of order while above it represented an increase of order. A/C, ascendency/development capacity. Fig. 4. Open in a new tab Entropy changes across different trophic levels from 2017 to 2020. Functional groups are labeled by numbers: 1, Scombridae; 2, Carangidae; 3, Haemulidae; 4, Muraenidae; 5, Serranidae; 6, Lutjanidae; 7, Balistidae; 8, Octopodidae; 9, Lethrinidae; 10, Nemipteridae; 11, Holocentridae; 12, Cirrhitidae; 13, Synodontidae; 14, Blenniidae; 15, Diodontidae; 16, Labridae; 17, Mullidae; 18, Chaetodontidae; 19, Pomacentridae; 20, Caesionidae; 21, Zanclidae; 22, Monacanthidae; 23, Tetraodontidae; 24, Ptereleotridae; 25, Pomacanthidae; 26, Scaridae; 27, Siganidae; 28, Acanthuridae; 29, Large crabs; 30, Small crabs; 31, Lobsters; 32, Shrimps; 33, Bivalves; 34, Gastropods; 35, Echinoderms; 36, Polychaetes; 37, Other Invertebrates; 38, Coral; 39, Zooplankton. R/A was used to estimate each taxon’s potential contribution to ecosystem resilience. To estimate the taxon’s energetic importance, taxon-specific R/A values were plotted against the taxon-specific Throughput ( Fig. 5 ). A significant negative correlation was observed between R/A and Throughput across all periods ( Fig. S4 ), indicating that taxa with higher potential flexibility were generally those with a lower energetic importance and vice versa. The highest number of taxa with R/A values exceeding the mean (61% of total taxa) occurred in 2017, suggesting that most species enhanced resilience in this period, while the lowest proportion (27% of total taxa) was observed in 2020. Four taxa (Muraenidae, Cirrhitidae, Blenniidae, and Pomacanthidae) consistently exhibited higher-than-mean R/A values, and five taxa (large crabs, small crabs, bivalves, gastropods, and polychaetes) demonstrated higher-than-average Throughput across all periods. Notably, Acanthuridae emerged as a taxon with both higher-than-average energetic importance and higher-than-grand mean informational flexibility across all periods. Fig. 5. Open in a new tab Taxon-specific relative flexibility from 2016 to 2020. The dotted green line indicates the mean R/A value (5.27), and the dotted red line indicates the mean Throughput (T) value (19.23) across all time periods. Energetic importance (flow), quantified as taxon-specific throughput, plotted on a log 10 scale. Numbers denote functional groups (see Fig. 4 ). 4. Discussion A trophic network model was developed by integrating data on taxa, temporal (years), and energy information. The simulation results from the model enabled a comprehensive analysis of the structure and function of coral reef ecosystems. The physicochemical characteristics of the coral reef ecosystem remained relatively stable from 2016 to 2020 ( Table S2 ), making the data in our study comparable. While there was a significant increase in coral coverage (from 14.13% to 42.10%), the biomass of most taxa in our study showed fluctuating changes. Previous research has demonstrated a unimodal, rather than linear, relationship between coral cover and taxon richness [ 58 ]. To maintain the organization of the coral reef ecosystem, a decrease in coral-dependent taxa was sometimes offset by an increase in algae-feeders when coral coverage was low. In contrast, very high coral coverage could lead to an increase in taxon richness, compensating for a decline in the richness of algae-feeders [ 58 ]. Our observations align with these findings: the biomass of coral-dependent fish, such as Chaetodontidae and Scaridae, increased with rising coral coverage, whereas the biomass of herbivorous fish, like Acanthuridae, decreased alongside reductions in turf algae and macroalgae ( Fig. 2 ). Some network metrics can provide insights into the maturity of the system [ 59 ]. The TST, defined as the cumulative sum of flows passing through the system, is one such metric. Ecosystem evolution was accompanied by an increase in ecosystem flow [ 60 ]. In our study, as coral coverage increased from 14.1% to 42.1% during the post-El Niño recovery period (2016−2020), TST increased by a factor of 1.10 [from 9455 to 2,0222 t/(km 2 ·year), Table 1 ], reflecting a strengthening of overall energy flux within the ecosystem. Concurrently, total NPP rose by a factor of 1.43 [from 3659 to 8921 t/(km 2 ·year), Table 1 ], indicating enhanced primary production capacity, a key driver of improved energy supply to higher trophic levels. These increases aligned with observed rises in TE (factor of 0.52) and NSP (factor of 3.08), collectively demonstrating that the expansion of coral cover was accompanied by a more efficient and robust energy flow network. In terms of thermodynamics, ecosystem shifts from one state to another resulted in alterations in both its order and entropy [ 21 ]. Ascendency (A), a measure of system order calculated as the product of TST and AMI (AMI = A/TST), traditionally indicated lower system resilience with higher values [ 61 ]. In our models, ascendency showed a clear increase by around 0.9 times, indicating enhanced organization, while the mean mutual information increased by 0.24 times, suggesting TST was the main driver behind the ascendency change. Furthermore, examination of ascendency by taxon revealed that low trophic level taxa (zooplankton, phytoplankton, turf algae, macroalgae, and detritus) contributed over 80% of total ascendency, highlighting their significant role in the development and maturity of ecosystems. Thus, controlling eutrophication and detritus should be taken seriously when formulating management strategies. The relative ascendency (A/C) excluded the influence of TST, served as a convenient index of system order or organization, and was a suitable index for comparing different ecosystems [ 62 ]. Calderon-Aguilera et al. [ 63 ] compared the structural and functional characteristics of 11 representative ecosystems from the eastern Pacific and found that higher values of A/C were observed in the Gulf of Ulloa (61%) and Cabo Pulmo (52%), suggesting that these two ecosystems were relatively more mature and specialized than the others. In our simulations, the value of A/C increased from 32.88% to 40.21% ( Table 1 ) as coral coverage gradually increased, indicating that a clear trend toward increasing order emerged. Yet, A/C for each taxon exhibited different patterns of change as coral coverage increased. Some taxa, such as fish, small crabs, and echinoderms, exhibited fluctuations over time but remained above their initial value, suggesting an increase in order. On the other hand, taxa like lobsters, bivalves, and polychaetes also demonstrated fluctuations but remained below the initial value, indicating a loss of order ( Fig. 3 ). Lobsters, bivalves, and polychaetes primarily occupy benthic/infaunal habitats dependent on unconsolidated sediments, detritus, or low-coral-cover microhabitats [ 64 ]. Increasing coral coverage shifts reef structure toward denser coral-dominated formations, compressing niche space for these taxa and disrupting their energy acquisition and trophic organization. This directly impacted their A/C, a metric of energy flow efficiency and functional order [ 23 ]. A significant correlation was found in relative change between A/C and biomass for the taxa during the period from 2019 to 2020 (Spearman rank correlation, p < 0.05; Fig. S5 ). This pattern likely reflects the fact that changes in energy levels would result in alterations in biomass, with organisms/taxon potentially redistributing their biomass based on the availability of energy [ 65 ]. By controlling the relationship between biomass and energy, the storage and transfer of energy within the ecosystem might be enhanced, ultimately preserving ecosystem order. In our study, the A/C (40.21%) was lower than O/C (59.79%), indicating that less than half of its energy contributed to the development and evolution of the coral reef system, and more than half of the energy available was used to cope with El-Niño. O/C described a measure of ecosystem entropy and represented the resilience of the system [ 22 ]. As the ecosystem developed, the entropy gradually decreased to a minimum level upon maturation [ 66 ]. In the coral reef system of Meiji, we found that the entropy (O/C) decreased from 67.12% to 59.79% with increasing coral coverage, suggesting the ecosystem was moving towards a relatively mature and stable state. Not all taxa in an ecosystem reduce entropy when developing. Fig. 4 exhibited that taxa demonstrating entropy reduction increased from 8% to 74% with rising coral cover from 14.13% to 42.10%. Arreguín-Sánchez and Ruiz-Barreiro [ 23 ] believed that the slope between the relative entropy and trophic levels of the taxa could be interpreted as a measure of the relative ecosystem vulnerability and stability, with the slope magnitude inversely related to instability. Through comparison of five marine ecosystems, they found no relationship (slope was zero) between the relative change in entropy and the TL of the perturbed taxon in the coral ecosystem, meaning higher vulnerability and overall system stability. Similarly, there was no relationship between the relative entropy and TL of the taxon in our study, suggesting a relatively stable state of the Meiji coral reef ecosystem. As mentioned above, some taxa reduced entropy/order while others generated entropy/order when the ecosystem was developing. By estimating the taxon-level flexibility, we could identify its potential role in ecosystem resilience and organization, which would facilitate the design of precautionary management. In the present study, throughput—the capacity of a taxon/species to process energy and matter—reflected its role in ecosystem functioning [ 67 ]. High-throughput taxa likely exert a pronounced influence on energy flow, potentially sustaining ecosystem functionality [ 68 ]. By integrating R/A with throughput, we developed a Comprehensive Contribution Assessment Framework, which enables comprehensive resilience assessment to identify key taxon pivotal for ecosystem resilience ( Fig. 5 ). The R/A values in the Meiji coral reef ecosystem varied from 2.17 to 10.70, with higher values indicating a stronger influence of the taxon on system flexibility rather than organization. Taxa with higher potential flexibility tended to have lower energetic importance. These findings were consistent with results from the marine ecosystem of Dauphin Island, where R/A values ranged from 2 to 12, and a negative relationship between potential flexibility and energetic importance was observed [ 24 ]. Therefore, potential resilience might lie within fish assemblages, and throughput could be maintained in benthic consumers. The herbivorous fish Acanthuridae was the only taxon with R/A and throughput values above the grand means, suggesting it might play a crucial role in actively contributing to system resilience. Previous studies have highlighted the significance of herbivorous fish in coral reef ecosystems. By consuming algae and attached detritus, they helped maintain the ecological balance of coral reefs, preventing them from being overrun by algae and losing their functionality. Additionally, Lin et al. [ 55 ] discovered that Acanthuridae exhibited a capacity for active adaptation to environmental changes by increasing its biomass and enhancing their physical condition in the coral reef ecosystem of the Nansha Islands when coral loss reached moderate levels. This observation indicated that Acanthuridae possessed notable ecological adaptability and resilience, enabling them to maintain their population size and overall health status even when faced with disturbances in the coral reef ecosystem, thereby contributing significantly to coral reef resilience. Thus, maintaining the population of herbivorous fish should be prioritized when formulating management strategies. We evaluated the maturity level of the Meiji coral reef ecosystem by comparing ecosystem metrics with representative ecosystems from Xisha Islands and 17 foreign coral reefs ( Table S3 ). TST represented the size of the ecosystem and indicated the energy circulation within the system. Generally, a higher TST corresponded to a higher level of ecosystem maturity, as well as NPP [ 22 ]. The TST [2,0222 t/(km 2 ·year)] and NPP [8921 t/(km 2 ·year)] of the Meiji coral reef ecosystem in Nansha Islands were significantly higher than those of Qilianyu in Xisha Islands, suggesting a relatively high level of maturity in the Nansha ecosystem [ 32 , 33 ]. The total primary production to total respiration ratio (TPP/TR) also reflected ecosystem maturity. A TPP/TR ratio closer to 1 indicated a more mature ecosystem [ 65 ]. The TPP/TR ratios obtained in Nansha Islands (6.99) and Xisha Islands (2.47−3.79) ranged from 2.47 to 6.99, indicating an immature state of the coral reef ecosystems. Additionally, CI and OI were correlated with the complexity of the food web. Higher values of CI and OI indicated more complex ecosystems, reflecting greater stability. Despite the lower values of CI (0.15) and OI (0.16) in the coral reef ecosystem in Nansha Islands compared to Xisha Islands (CI 0.31−0.33, OI 0.21−0.24), the relatively low CI and OI (<1) suggested limited connections between trophic groups in both regions. The metric A/C represented ecosystem organization weighted by the total system flow. In mature systems, A/C values were high, while in resilient systems, they were low [ 22 ]. In our study, the A/C value (40.21%) was significantly higher than that of Qilianyu in the Xisha Islands (27.21%), indicating a relatively mature ecosystem in the Nansha Islands. These differences might be explained by varying degrees of human activity disturbance in the two regions. Nansha Islands, being relatively distant from the mainland and with fishing activities prohibited in coral reef areas, experienced less human disturbance [ 30 , 69 ]. This allowed for the growth of coral and the development of coral reef ecosystems. In contrast, the Xisha Islands, an important fishing ground for commercial fisheries, faced frequent disturbances from human activities such as overfishing and destructive fishing. The coral coverage of Xisha Islands decreased from 70% to 15% between 2006 and 2019, leading to ecosystem degradation [ 70 , 71 ]. Compared with 17 foreign coral reefs, the TST ranked eleventh, and NPP ranked seventh in the Meiji coral reef ecosystem of the Nansha Islands. The mean transfer efficiency for the entire system, at 10.03%, was slightly lower than that observed in Isla del Coco and Darwin and Wolf Islands ( Table S3 ). However, it remained comparable to the 10% proposed by Lindeman [ 72 ], which has been demonstrated to be a reliable estimate of the average transfer efficiency across aquatic ecosystems. The values of CI and OI in the 17 foreign coral reefs were slightly higher than those in Meiji reef, yet these values remained relatively low (<1), indicating an unstable environment for the few trophic groups. The A/C ratio described the degree of organization and was low among selected ecosystems (24%−38%; Table S3 ), except for the Meiji Reef (40%), Holbox (41%), Campeche Bank Mexico (41%), Cayos (47%), and Cabo Pulmo (52%), suggesting that these five ecosystems were relatively more mature and specialized than the others. 5. Conclusion This study integrated Ecopath with Ecosim modeling, stable isotope analysis, and key parameter validation to elucidate the resilience mechanisms of the Nansha coral reef ecosystem. Our findings demonstrated that the ecosystem exhibited significant resilience to disturbance, characterized by the temporal stability of core structural metrics (CI, OI, Finn’s mean path length, mean trophic level) and the preservation of foundational food web structure. Crucially, concurrent recovery capacity was evident through sustained annual increases in functional metrics (TST, NSP, TB, NPP, A/C, and TE), signifying enhanced energy utilization efficiency, resource processing, and overall productivity post-disturbance. A pivotal thermodynamic mechanism underpinning this resilience was identified: rising coral coverage fostered “global order with local flexibility”. Increased order in most taxa reflected optimized energy allocation and structural stability (e.g., dominant species), while localized entropy in specific groups (e.g., opportunistic species) signified adaptive flexibility to environmental fluctuations. This dynamic interplay ensured functional redundancy and mitigated systemic collapse. Furthermore, we developed a Comprehensive Contribution Assessment Framework, establishing a quantitative link between R/A and Throughput to identify key species pivotal for ecosystem resilience. The research framework established here—encompassing thermodynamic drivers, quantitative functional weighting of species, and recovery trajectory analysis for immature systems—provides a robust theoretical foundation. It not only deepens the understanding of the Nansha reefs but also offers a valuable tool for assessing resilience and formulating restoration strategies for tropical coral reefs globally, particularly those in vulnerable recovery phases. For instance, concerning the Acanthuridae (surgeonfish) family, identified in this study as not yet fully recovered, specific restoration plans can be developed based on their pivotal role in energy transfer within the food web. This would accelerate the recovery of its population, thereby preventing delays in the maturation of the overall system due to the lag in recovery of a single taxonomic group. We acknowledge limitations inherent in our approach. Constructing functional groups at taxa level, while necessary for model tractability, may introduce deviations from species-specific dynamics. Challenges also remain in capturing the inherent temporal variability of diet matrices, although nitrogen stable isotope analysis helped mitigate this gap. Future research should prioritize species-level modeling where feasible, develop methods for high-resolution temporal diet tracking, and apply this integrated framework to diverse reef systems under varying disturbance regimes to further validate and refine our understanding of coral reef resilience. CRediT authorship contribution statement Yanyan Zhou: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Xianzhi Lin: Investigation. Haoxuan Cheng: Investigation. Shuo Zhai: Investigation. Sen Du: Investigation. Lizhao Chen: Investigation. Li Zhang: Writing – review & editing, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This work was supported by the National Key Research and Development Program of China (2022-20), Hainan Provincial Natural Science Foundation of China (423CXTD392), Science and Technology Planning Project of Guangdong Province, China (2023B1212060047), Science and Technology Planning Project of Guangzhou (202201010198), Special Fund of South China Sea Institute of Oceanology of the Chinese Academy of Sciences (SCSIO2023QY04). Footnotes Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.eehl.2026.100232 . Appendix A. Supplementary data The following are the Supplementary data to this article: Multimedia component 1 mmc1.docx (2MB, docx) Multimedia component 2 mmc2.xlsx (44KB, xlsx) References 1. Odum H.T., Odum E.P. Trophic structure and productivity of a windward coral reef community on eniwetok atoll. Ecol. Monogr. 1955;25(3):291–320. [ Google Scholar ] 2. Guan Y., Hohn S., Wild C., Merico A. Vulnerability of global coral reef habitat suitability to ocean warming, acidification and eutrophication. Glob. Change Biol. 2020;26(10):5646–5660. doi: 10.1111/gcb.15293. [ DOI ] [ PubMed ] [ Google Scholar ] 3. Shaver E.C., McLeod E., Hein M.Y., Palumbi S.R., Quigley K., Vardi T., et al. A roadmap to integrating resilience into the practice of coral reef restoration. Glob. Change Biol. 2022;28(16):4751–4764. doi: 10.1111/gcb.16212. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Chi-Espínola A.A., Vega-Cendejas M.E. Trophic dynamics and properties of the marine ecosystem of Campeche bank, Mexico. Mar. Biol. 2021;169(1):14. [ Google Scholar ] 5. Cáceres I., Ibarra-García E.C., Ortiz M., Ayón-Parente M., Rodríguez-Zaragoza F.A. Effect of fisheries and benthic habitat on the ecological and functional diversity of fish at the cayos cochinos coral reefs (honduras) Mar. Biodivers. 2020;50(1):9. [ Google Scholar ] 6. Martin C.W., Lewis K.A., McDonald A.M., Spearman T.P., Alford S.B., Christian R.C., et al. Disturbance-driven changes to northern Gulf of Mexico nekton communities following the deepwater horizon oil spill. Mar. Pollut. Bull. 2020;155 doi: 10.1016/j.marpolbul.2020.111098. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Capitani L., de Araujo J.N., Vieira E.A., Angelini R., Longo G.O. Ocean warming will reduce standing biomass in a tropical Western Atlantic reef ecosystem. Ecosystems. 2022;25(4):843–857. [ Google Scholar ] 8. Ulanowicz R.E. Biodiversity, functional redundancy and system stability: subtle connections. J. R. Soc. Interface. 2018;15(147) doi: 10.1098/rsif.2018.0367. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Ulanowicz R.E. Complexity in Ecological Systems Series. Columbia University Press; New York: 1997. Ecology, the ascendent perspective; p. 201. [ Google Scholar ] 10. Robinson J.P.W., Benkwitt C.E., E M., Morais R., Schiettekatte N.M.D., Skinner C., et al. Quantifying energy and nutrient fluxes in coral reef food webs. Trends Ecol. Evol. 2024;39(5):467–478. doi: 10.1016/j.tree.2023.11.013. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Sinnickson D., Harris H.E., Chagaris D. Assessing energetic pathways and time lags in estuarine food webs. Ecosystems. 2023;26(7):1468–1488. [ Google Scholar ] 12. Ruesink J.L., Hodges K.E., Krebs C.J. Mass-balance analyses of boreal forest population cycles: merging demographic and ecosystem approaches. Ecosystems. 2002;5(2):138–158. [ Google Scholar ] 13. Weijerman M., Gove J.M., Williams I.D., Walsh W.J., Minton D., Polovina J.J. Evaluating management strategies to optimise coral reef ecosystem services. J. Appl. Ecol. 2018;55(4):1823–1833. [ Google Scholar ] 14. Christensen V., Walters C.J. Ecopath with ecosim: methods, capabilities and limitations. Ecol. Model. 2004;172(2–4):109–139. [ Google Scholar ] 15. Ulanowicz R.E. Treatise on Estuarine and Coastal Science. Elsevier; Amsterdam: 2011. Quantitative methods for ecological network analysis and its application to coastal ecosystems; pp. 35–57. [ Google Scholar ] 16. Costanza R. Ecosystem health and ecological engineering. Ecol. Eng. 2012;45:24–29. [ Google Scholar ] 17. Bacalso R.T.M., Wolff M., Rosales R.M., Armada N.B. Effort reallocation of illegal fishing operations: a profitable scenario for the municipal fisheries of Danajon Bank, central Philippines. Ecol. Model. 2016;331:5–16. [ Google Scholar ] 18. Cáceres I., Ortiz M., Cupul-Magaña A.L., Rodríguez-Zaragoza F.A. Trophic models and short-term simulations for the coral reefs of cayos Cochinos and Media luna (Honduras): a comparative network analysis, ecosystem development, resilience, and fishery. Hydrobiologia. 2016;770(1):209–224. [ Google Scholar ] 19. Ibarra-García E.C., Ortiz M., Ríos-Jara E., Cupul-Magaña A.L., Hernández-Flores Á., Rodríguez-Zaragoza F.A. The functional trophic role of whale shark (Rhincodon typus) in the northern Mexican Caribbean: network analysis and ecosystem development. Hydrobiologia. 2017;792(1):121–135. [ Google Scholar ] 20. Hermosillo-Núñez B.B., Ortiz M., Rodríguez-Zaragoza F.A., Cupul-Magaña A.L. Trophic network properties of coral ecosystems in three marine protected areas along the Mexican Pacific Coast: assessment of systemic structure and health. Ecol. Complex. 2018;36:73–85. [ Google Scholar ] 21. Ulanowicz R.E., Goerner S.J., Lietaer B., Gomez R. Quantifying sustainability: resilience, efficiency and the return of information theory. Ecol. Complex. 2009;6(1):27–36. [ Google Scholar ] 22. Ulanowicz R.E. Springer-Verlag; New York, USA: 1986. Growth and Development: Ecosystem Phenomenology. [ Google Scholar ] 23. Arreguín-Sánchez F., Ruiz-Barreiro T.M. Approaching a functional measure of vulnerability in marine ecosystems. Ecol. Indic. 2014;45:130–138. [ Google Scholar ] 24. Lewis K.A., Christian R.R., Martin C.W., Allen K.L., McDonald A.M., Roberts V.M., et al. Complexities of disturbance response in a marine food web. Limnol. Oceanogr. 2022;67(S1):S352–S364. [ Google Scholar ] 25. Zhao H.T. Science Press; Beijing: 1996. Physical Geography of Nansha Islands. [ Google Scholar ] 26. Tan F., Zhang Y., Fu G., Shi Q., Zhang X., Zhou S., et al. Triggering of a 2500-year coral shutdown in northern South China Sea by coupled East Asian monsoon and El Niño–Southern oscillation. Global Planet. Change. 2025;245 [ Google Scholar ] 27. Lin Y., Chen T.-R., Leonard N.D., Zhao J.-X. Millennial-scale episodic coral growth on the northern margin of the South China Sea. Global Planet. Change. 2025;253 [ Google Scholar ] 28. Zhang T., Chen T., Liu S., Lin X., Li S., Yan W. Coral reef resilience persisted for a millennium but has declined rapidly in recent decades. Front. Mar. Sci. 2023;10 [ Google Scholar ] 29. Mo S., Chen T., Chen Z., Zhang W., Li S. Marine heatwaves impair the thermal refugia potential of marginal reefs in the northern South China Sea. Sci. Total Environ. 2022;825 doi: 10.1016/j.scitotenv.2022.154100. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Gong Y., Zhang J., Chen Z., Cai Y., Yang Y. Taxonomic diversity and interannual variation of fish in the lagoon of meiji reef (Mischief reef), South China Sea. Biology. 2024;13(9):740. doi: 10.3390/biology13090740. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Zhao J., Wang T., Li C., Shi J., Xie H., Luo L., et al. Seven decades of transformation: evaluating the dynamics of coral reef fish communities in the Xisha Islands, South China Sea. Rev. Fish Biol. Fish. 2024;34(4):1261–1281. [ Google Scholar ] 32. Zhang X., Li Y., Du J., Qiu S., Xie B., Chen W., et al. Effects of ocean warming and fishing on the coral reef ecosystem: a case study of Xisha Islands, South China Sea. Front. Mar. Sci. 2022;9 [ Google Scholar ] 33. Hong X., Chen Z., Zhang J., Jiang Y., Gong Y., Cai Y., et al. Construction and analysis of a coral reef trophic network for Qilianyu Islands, Xisha Islands. Acta Oceanol. Sin. 2022;41(12):58–72. [ Google Scholar ] 34. Huang H., Chen Z., Huang L.T. Ocean Press; Beijing: 2021. Status of Coral Reefs in China (2010-2019) [ Google Scholar ] 35. Ferrario F., Beck M.W., Storlazzi C.D., Micheli F., Shepard C.C., Airoldi L. The effectiveness of coral reefs for coastal hazard risk reduction and adaptation. Nat. Commun. 2014;5:3794. doi: 10.1038/ncomms4794. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Chen T., Tan X., Zhang T., Liu S., Zhao J. Exploring branching corals as high-resolution paleo-SST archives. Quat. Sci. Rev. 2025;349 [ Google Scholar ] 37. Wang M., Ye J., Zhang X., Tan F., Shi Q., Sun F., et al. Dynamic evolution of coral reef ecosystems in the South China Sea under global change: a comprehensive multidimensional numerical simulation. Coral Reefs. 2025;44(1):309–319. [ Google Scholar ] 38. Carrer S., Opitz S. Trophic network model of a shallow water area in the northern part of the lagoon of venice. Ecol. Model. 1999;124(2–3):193–219. [ Google Scholar ] 39. Allen K.R. Relation between production and biomass. J. Fish. Res. Board Can. 1971;28(10):1573–1581. [ Google Scholar ] 40. Bohnsack J.A., Harper D.E. Length-weight relationships of selected marine reef fishes from the southeastern United States and the Caribbean. NOAA Tech. Mem. NMFS-SEFC-215. 1988:31. [ Google Scholar ] 41. McClanahan T.R. A coral reef ecosystem-fisheries model: impacts of fishing intensity and catch selection on reef structure and processes. Ecol. Model. 1995;80(1):1–19. [ Google Scholar ] 42. Zheng L.X., Li B.Q., Wang H.F., Wang S.Q., Wang J.B., Zhang B.L. Macrobenthic community characters of zhubi reef, Nansha Islands, South China Sea. Acta Zool. Sin. 2007;53:83–94. [ Google Scholar ] 43. Pauly D., Soriano-Bartz M.L., Palomares M.L.D. In: Trophic Models of Aquatic Ecosystems, ICLARM Conference Proceedings. Christensen V., Pauly D., editors. 1993. Improved construction, parameterization and interpretation of steady-state ecosystem models; pp. 1–13. [ Google Scholar ] 44. Charpy L., Blanchot J. Photosynthetic picoplankton in French Polynesian atoll lagoons: estimation of taxa contribution to biomass and production by flow cytometry. Mar. Ecol. Prog. Ser. 1998;162(February 12 1998):57–70. [ Google Scholar ] 45. Pauly D., Christensen V. Mass-balance models of north-eastern Pacific ecosystems. Fish Centre. Res. Rep. 1996;4:1–131. [ Google Scholar ] 46. Pauly D. On the interrelationships between natural mortality, growth parameters, and mean environmental temperature in 175 fish stocks. ICES J. Mar. Sci. 1980;39(2):175–192. [ Google Scholar ] 47. Palomares M.L., Pauly D. A multiple regression model for prediction the food consumption of marine fish populations, mar. Freshwater Res. 1989;40(3):259. [ Google Scholar ] 48. Wu C.Y., Zhang J.L., Huang L.M. Primary productivity in some coral reef lagoons and their adjacent sea areas of Nansha Islands in spring. J. Trop. Oceanograph. 2011;20:59–67. [ Google Scholar ] 49. Tan Y.H., Huang L.M., Yin J.Q. Estimation of secondary zooplankton productivity and transformation efficiency in Nansha islands sea areas. J. Trop. Oceanograph. 2003;22:29–34. [ Google Scholar ] 50. Chen Z., Qiu Y., Jia X., Xu S. Using an ecosystem modeling approach to explore possible ecosystem impacts of fishing in the Beibu Gulf, northern South China Sea. Ecosystems. 2008;11(8):1318–1334. [ Google Scholar ] 51. Liu P.J., Shao K.T., Jan R.Q., Fan T.Y., Wong S.L., Hwang J.S., et al. A trophic model of fringing coral reefs in Nanwan Bay, southern Taiwan suggests overfishing. Mar. Environ. Res. 2009;68(3):106–117. doi: 10.1016/j.marenvres.2009.04.009. [ DOI ] [ PubMed ] [ Google Scholar ] 52. Ma W.G., Yin H.Y., Sun C.Y., Wang Z.G., Wei Y.F., Feng B.X., et al. The ecological carrying capacity of Stichopus monotuberculatus and ecological effect prediction in a tropical coral reef island marine ranching area. Oceanol. Limnl. Sin. 2000;53:1573–1584. [ Google Scholar ] 53. Jia N., Zhou T.C., Hu S.M., Zhang S., Huang H., Liu S. Difference in the feeding contents of three hermit crabs in coral reefs of Nansha Island, South China Sea. J. Trop. Oceanograph. 2023;43:109–121. [ Google Scholar ] 54. Ricker W.E. Committee on Resources and Man. US National Academy of Sciences, E. H. Freeman; San Francisco, CA: 1968. Food from the sea; pp. 87–108. Resource and Man. Chapter 5. [ Google Scholar ] 55. Lin X., Hu S., Liu Y., Zhang L., Huang H., Liu S. Disturbance-mediated changes in coral reef habitat provoke a positive feeding response in a major coral reef detritivore, Ctenochaetus striatus. Front. Mar. Sci. 2021;8 [ Google Scholar ] 56. Post D.M. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology. 2002;83(3):703–718. [ Google Scholar ] 57. Jones H.J., Swadling K.M., Butler E.C.V., Barry L.A., Macleod C.K. Application of stable isotope mixing models for defining trophic biomagnification pathways of Mercury and selenium. Limnol. Oceanogr. 2014;59(4):1181–1192. [ Google Scholar ] 58. Wilson S.K., Dolman A.M., Cheal A.J., Emslie M.J., Pratchett M.S., Sweatman H.P.A. Maintenance of fish diversity on disturbed coral reefs. Coral Reefs. 2009;28(1):3–14. [ Google Scholar ] 59. Heymans J.J., Coll M., Link J.S., Mackinson S., Steenbeek J., Walters C., et al. Best practice in ecopath with ecosim food-web models for ecosystem-based management. Ecol. Model. 2016;331:173–184. [ Google Scholar ] 60. Salaün J., Raoux A., Pezy J.-P., Dauvin J.-C., Pioch S. Structural and functional changes in artificial reefs ecosystem stressed by trophic modelling approach: case study in the Bay of Biscay. Reg. Stud. Mar. Sci. 2023;65 [ Google Scholar ] 61. Ulanowicz R.E., Jørgensen S.E., Fath B.D. Exergy, information and aggradation: an ecosystems reconciliation. Ecol. Model. 2006;198(3–4):520–524. [ Google Scholar ] 62. Baird D., McGlade J.M., Ulanowicz R.E. The comparative ecology of six marine ecosystems. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 1991;333(1266):15–29. [ Google Scholar ] 63. Calderon-Aguilera L.E., Reyes-Bonilla H., Olán-González M., Castañeda-Rivero F.R., Perusquía-Ardón J.C. Estimated flows and biomass in a no-take coral reef from the eastern tropical Pacific through network analysis. Ecol. Indic. 2021;123 [ Google Scholar ] 64. Nelson H.R., Kuempel C.D., Altieri A.H. The resilience of reef invertebrate biodiversity to coral mortality. Ecosphere. 2016;7(7) [ Google Scholar ] 65. Odum H.T. Explanations of ecological relationships with energy systems concepts. Ecol. Model. 2002;158(3):201–211. [ Google Scholar ] 66. Saint-Béat B., Baird D., Asmus H., Asmus R., Bacher C., Pacella S.R., et al. Trophic networks: how do theories link ecosystem structure and functioning to stability properties? A review. Ecol. Indic. 2015;52:458–471. [ Google Scholar ] 67. Fath B.D., Scharler U.M., Baird D. Dependence of network metrics on model aggregation and throughflow calculations: demonstration using the sylt–rømø bight ecosystem. Ecol. Model. 2013;252:214–219. [ Google Scholar ] 68. Christian R.R., Voss C.M., Bondavalli C., Viaroli P., Camacho-Ibar V. In: Coastal Lagoons: Critical Habitats of Environmental Change. Kennish M.J., Paerl H.W., editors. CRC Press; 2010. Ecosystem health indexed through networks of nitrogen cycling; pp. 73–90. [ Google Scholar ] 69. Zhou Y., Du S., Liu Y., Yang T., Liu Y., Li Y., et al. Source identification and risk assessment of trace metals in surface sediment of China sea by combining APCA-MLR receptor model and lead isotope analysis. J. Hazard. Mater. 2024;465 doi: 10.1016/j.jhazmat.2023.133310. [ DOI ] [ PubMed ] [ Google Scholar ] 70. Wu Z., Li Y., Liang J., Zhao J., Chen S. Analysis on the outbreak period and cause of Acanthaster planci in Xisha Islands in recent 15 years. Chin. Sci. Bull. 2019;64(1):1–7. [ Google Scholar ] 71. Li Y.J., Chen Z.Z., Zhang J., Jiang Y.E., Gong Y.Y., Cai Y.C., et al. Species and taxonomic diversity of Qilianyu island reef fish in the Xisha Islands. J. Fish. Sci. China. 2021;27:815–823. [ Google Scholar ] 72. Lindeman R.L. The trophic-dynamic aspect of ecology. Bull. Math. Biol. 1991;53(1–2):167–191. [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Multimedia component 1 mmc1.docx (2MB, docx) Multimedia component 2 mmc2.xlsx (44KB, xlsx) Articles from Eco-Environment & Health are provided here courtesy of Elsevier ACTIONS View on publisher site PDF (3.2 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top