QIAO Xudong, WANG Baodong, SUN Xia and LIANG Shengkang

1) State Key Laboratory of Satellite Ocean Environment Dynamics, The Second Institute of Oceanography, State Oceanic Administration of China, Hangzhou 310012, P. R. China

2) Marine Ecology Research Center, The First Institute of Oceanography, State Oceanic Administration of China, Qingdao 266061, P. R. China

3) College of Chemistry and Chemical Engineering, Ocean University of China, Qingdao 266100, P. R. China

Numerical Simulation of Nutrient and Phytoplankton Dynamics in Guangxi Coastal Bays, China

QIAO Xudong1), WANG Baodong2),*, SUN Xia2), and LIANG Shengkang3)

1) State Key Laboratory of Satellite Ocean Environment Dynamics, The Second Institute of Oceanography, State Oceanic Administration of China, Hangzhou 310012, P. R. China

2) Marine Ecology Research Center, The First Institute of Oceanography, State Oceanic Administration of China, Qingdao 266061, P. R. China

3) College of Chemistry and Chemical Engineering, Ocean University of China, Qingdao 266100, P. R. China

The increasing riverine pollutants have resulted in nutrient enrichment and deterioration of water quality in the coastal water of Guangxi Province, China. However, the quantitative relationship between nutrient loads and water quality responses, which is crucial for developing eutrophication control strategies, is not well studied. In this study, the riverine fluxes of nutrients were quantified and integrated with nutrient cycling and phytoplankton dynamics by using box models for Guangxi coastal bays. The model concepts and biogeochemical equations were the same; while most model parameters were specific for each bay. The parameters were calibrated with seasonal observations during 2006–2007, and validated with yearly averaged measurements in 2009. The general features of nutrient and phytoplankton dynamics were reproduced, and the models were proved feasible under a wide range of bay conditions. Dissolved inorganic nitrogen was depleted during the spring algal bloom in Zhenzhu Bay and Fangcheng Bay with relatively less nutrient inputs. Phosphorus concentration was high in spring, which decreased then due to continuous phytoplankton consumption. Chlorophyll-a concentration reached its annual maximum in summer, but was the minimum in winter. Eutrophication was characterized by both an increase in nutrient concentrations and phytoplankton biomass in Lianzhou Bay. Either about 80% reduction of nitrogen or 70% reduction of phosphorus was required to control the algal bloom in Lianzhou Bay. Defects of the models were discussed and suggestions to the environmental protection of Guangxi coastal bays were proposed.

ecosystem model; nutrient; phytoplankton; seasonal variation; coastal bay

1 Introduction

In recent years, eutrophication and water quality degradation in coastal waters due to excessive anthropogenic nutrient inputs have received worldwide concerns (Kuusisto et al., 1998; Wang et al., 1999; Savchuk, 2002; Chau et al., 2004; Park et al., 2005; Los and Wijsman, 2007; Han et al., 2011). Since coastal ecosystems are complex, dynamic and non-linear, it is difficult to fully understand the problem just by limited field observations. Instead, the mechanics of coastal ecosystem can be explored by numerical modeling which integrates knowledge derived from process studies. Including the important nutrient cycles and plankton dynamics, the modeling may strengthen our understanding of marine system behaviors and optimize decision-making process (Baretta et al., 1995; Cloern, 2001; Stow et al., 2009). Although the general trend in ecosystem modeling is shifting from box models to three-dimensional models, fast-running box models can provide first approaches to long-term processes of ecosystem dynamics and give reliable results about seasonal and annual trends (Ménesguen et al., 2007).

Guangxi coastal bays are located in the northern Beibu Gulf, China and consist of five major semi-closed water bodies: Zhenzhu Bay, Fangcheng Bay, Qinzhou Bay, Lianzhou Bay and Tieshan Bay (Fig.1), which are characterized by diurnal tides. The hydrodynamics of the bays have been studied using a three-dimensional model by Zhao et al. (2002). Several studies demonstrated that the marine coastal ecosystems in the Qinzhou Bay and Lianzhou Bay have been deteriorating continuously as a result of nutrient enrichment arising from terrestrial inputs (Chen, 2001; Xin et al., 2010; Lan, 2011; Long et al., 2011; Zhuang et al., 2011). Eutrophication phenomenon with frequent algal blooms has been reported in Lianzhou Bay (He and Wei, 2009). However, most previous studies (Qiu, 2001; Zhuang et al., 2010; Lan and Peng, 2011; Long et al., 2011; Zhang et al., 2011) have been descriptive rather than mechanistic. Inputs from the rivers and other resources, which determine the biogeochemical cycling system to a large extent, are not well studied yet. Subsequently, little attention has been paid to the quanti-tative relationship between terrestrial inputs and water quality responses, which limited the development of nutrient control strategies. In this study, riverine nutrient loads to the bays were quantified and integrated with nutrient cycling and phytoplankton dynamics in ecosystem box models. The models were calibrated with seasonal field data, and validated with annual mean data for Guangxi coastal bays. Nutrient reduction targets for Lianzhou Bay were also suggested based on the model calculations.

Fig.1 Location of monitoring sites in Guangxi coastal bays. The polygons indicate the water bodies and dots illustrate the sites.

2 Monitoring Programs

Four seasonal investigations were carried out by The First Institute of Oceanography, State Oceanic Administration of China in July 2006, December 2006, April 2007 and October 2007, respectively. The water samples at surface and bottom layers were collected from 10 stations (Fig.1). Water samples for nutrient analysis were filtered through 0.45 μm cellulose acetate filter. Nitrate (NO3-N), nitrite (NO2-N), ammonium (NH4-N) and phosphate (PO4-P) were determined spectrophotometically with methods previously described by Wang et al. (2003). Dissolved inorganic nitrogen (DIN) was calculated as the sum of NO3-N, NO2-N and NH4-N. Chlorophyll-a (chl-a) was measured fluorometrically in accordance with Strickland and Parsons (1972). All data collected at stations in each bays during each cruise was treated as replicates and averaged, and was used for model calibration.

Table 1 Characteristics of Guangxi coastal bays and terrestrial nitrogen and phosphorus loads in 2006–2007

3 Ecosystem Models

The structure of ecosystem models is schematically illustrated in Fig.2. Seven variables were considered, which included DIN, PO4-P, dissolved organic nitrogen (DON), dissolved organic phosphorus (DOP), phytoplankton (PPT), zooplankton (ZPT) and detritus (DPT). Nitrogen was taken as flow parameter for the domination processes within these compartments. The model concepts and biogeochemical equations were the same for different bays. We assumed that the uptake and release of N and P by marine phytoplankton obeyed Redfield ratio of 16 (Redfield et al., 1963). Due to a lack of observational data for the sediment-seawater interface, detritus resuspension and benthic nutrient fluxes were not included in our models.

Fig.2 Schematic diagram of biogeochemical cycles in the lower tropic ecosystem model. 1, photosynthesis; 2, phytoplankton respiration; 3, zooplankton grazing; 4, phytoplankton mortality; 5-6, zooplankton excretion; 7, zooplankton grazing on detritus, defecation and mortality; 8–9, detritus decay and mineralization; 10, dissolved organic nutrients mineralization; 11, detritus sedimentation; 12, river inputs; 13, water exchange.

3.1 Evolution Equations and Forcing Variables

Temporal evolution of each nutrient compartment is described as follows:

where Ciis the arbitrary nutrient compartment; t is time; kOis the exchange rate with the outside water body; COis the concentration of the compartment in outside water. The averaged 1.4 mmol m?3for nitrogen and 0.03 mmol m?3for phosphorus in adjacent Beibu Gulf (Xin et al., 2010) were adopted as the model boundary conditions.

The exchange rate kO(shown in Table 2) is determined as described by Zhou et al. (1996）:

where QFis the total volume of water flowing into the bay during flood tide; γEis the fraction of ‘new’ ocean water entering on the flood tide (Parker et al., 1972); γFis the fraction of ‘new’ bay water exiting on the ebb tide (Kashiwai, 1984); V represents the mean bay volume (shown in Table 1). The values of γE, γFand QFfor the five bays have been described early (Table 6.3.1 of Zhao et al., 2002). The development equations for biological compartments are similar to Eq. 1 except for the absence of river inflow term. The exchange rate of particulates (kPO, Table 2) is assumed to be about half of kO.

The ecosystem seasonality is assumed to be driven by changes in incident photosynthetically active radiation (PAR), temperature (Fig.3) and nutrient inputs. Estimates of annual N and P loads to the bays are shown in Table 1, which were analyzed on the data of river discharges and nutrient concentrations collected by Marine Environmental Monitoring Center of Guangxi Province during 2006–2007. The seasonal variations of river discharge and nutrient input are demonstrated in Fig.4. No data for atmospheric deposition and mariculture sources were available.

Fig.3 Annual course of seawater temperature (monthly mean data from Editorial Board of Annals of Bays in China, 1993) and incident PAR (PAR is assumed to be 48% of monthly mean total incident solar radiation data from China Meteorological Data Sharing Service System) of Guangxi coastal bays described by sinusoidal type functions.

Fig.4 Seasonal variations of river discharge (a), nitrogen (b) and phosphorus (c) inputs of Guangxi coastal bays.

3.2 Main Ecological Processes and Parameters

Nutrients enter the bays as riverine inputs. They are assimilated by the phytoplankton, and regenerated in water by phytoplankton respiration, zooplankton excretion and dissolved organic nutrients (DON and DOP) mineralization. Dissolved organic nutrients are derived from the excretion of phytoplankton and zooplankton, as well as the detritus decay process.

The rate of phytoplankton growth is mediated by water temperature, solar irradiation and nutrient (N and P) availability. Population losses result from respiration, natural mortality, zooplankton grazing and sinking. Phytoplankton concentrations are computed internally in nitrogen units and converted to chl-a using stoichiometric relationship (rChl/PN, Table 2). Zooplankton biomass grows via grazing phytoplankton and detritus, which is limited by temperature and food availability. Zooplankton abundance declines as a result of natural mortality, consumption by predators at high ranks, defecation and excretion.

The detritus compartment is comprised of fecal materials, dead phytoplankton and zooplankton. The detritus can be recycled within the systems through re-ingestion by zooplankton or mineralization by bacteria. Detrital sinking is modeled by specifying a simple sinking rate and the resulting flux is considered to be exported from the water column.

Details of ecosystem models have been previously described in Jiaozhou Bay, and the nutrient cycling equations can be found in Li et al. (2007). Model Maker 4.0 (Cherwell Scientific Ltd, UK), a computer software, is used to integrate the compartments together and solve the differential equations. The estimation of hydrodynamic exchange rates has been described in Section 3.1. The light attenuation coefficient of seawater (β0) was determined using the mean suspended particulate matter concentration measured during the cruises, as described in Irigoien and Castel (1997). Initial values of the other model parameters were originally adopted from the Jiaozhou Bay ecological model (Li et al., 2007) and the literature of other coastal bays (Xu and Hood, 2005; Park et al., 2005). The parameters were adjusted to obtainsimulations comparable with the observations, and the new parameters are given in Table 2.

Table 2 Parameters used in the ecosystem modeling for Guangxi coastal bays

4 Modeling Results

4.1 Model Calibration

The simulations were conducted from 1 January 2006 to 31 December 2007 and the observed values of winter cruise were taken as initial conditions. Field data of the five bays are compared with modeling outputs in Fig.5 and Table 3. The simulated data before June 2006 are not shown here for concision. Range bars are used to indicate spatial variability of the measurement and mean relative standard deviation (RSD) to quantify the errors in simulation. Both simulations and field data manifested strong seasonal variations.

Zhenzhu Bay: The measured DIN, PO4-P and chl-a concentrations ranged from 0.3 to 3.6 mmol m?3, from 0.03 to 0.09 mmol m?3and from 0.7 to 3.0 mg m?3, respectively. DIN was depleted when phytoplankton started to bloom in April, and was high in October and December. By contrast, PO4-P was low in July but high in April. Chl-a increased to the maximum in wet summer-autumn seasons (July–October) and decreased to the minimum in dry winter-spring seasons (December–March). PO4-P was reproduced well in the simulation (RSD=9.1%). Althoughthe biomass in winter-spring were underestimated, model phytoplankton outputs generally followed seasonal trends of the measurement (RSD=39.4%). However, the DIN simulation failed to reveal the lowest value, which resulted in a large RSD value of 146.4%.

Fangcheng Bay: The measured DIN, PO4-P and chl-a values varied from 1.3 to 4.5 mmol m?3, from 0.04 to 0.15 mmol m?3and from 2.2 to 5.4 mg m?3, respectively. DIN, PO4-P and chl-a exhibited similar trends to those in Zhenzhu Bay. The general trends were followed by the simulations, while the phytoplankton biomass in winter-spring were also underestimated. The mean RSDs for DIN, PO4-P and chl-a simulation were 30.8%, 19.3% and 34.0%, respectively.

Qinzhou Bay: The measured DIN, PO4-P and chl-a values ranged from 4.4 to 13.4 mmol m?3, from 0.04 to 0.35 mmol m?3and from 1.9 to 3.3 mg m?3, respectively. Despite the variability in the aggregated data, the seasonal trends were clear. DIN was low in December but high in October. PO4-P displayed a winter minimum and a spring maximum. Chl-a rose in spring, and peaked in summer then declined in autumn and winter. The model outputs generally revealed the observations. DIN was reproduced well in our simulation (RSD=7.0%), while PO4-P and chl-asimulations demonstrated some discrepancies with observations (RSD equaled 56.6% and 36.3%, respectively).

Fig.5 Model (–) and measured DIN (●), PO4-P (□) and chl-a (○) seasonal variations for five Guangxi coastal bays. Scaling on the y-axis are different.

Lianzhou Bay: The measured DIN, PO4-P and chl-a values varied from 3.5 to 23.1 mmol m?3, from 0.06 to 0.53 mmol m?3and from 1.3 to 32.9 mg m?3, respectively. DIN was low in October and December, but high in July. PO4-P was much higher in spring than in the other three seasons. Chl-a displayed a peak concentration far exceeding 10.0 mg m?3(Eutrophication; National Academy of Science, NAS, 1972) in July and a minimum in April. The chl-a concentration was 15.3 mg m?3high even in December. The model outputs represented the general trends of observed data. The mean RSDs for DIN, PO4-P and chl-a simulation were 74.2%, 87.4% and 45.1%, respectively.

Tieshan Bay: The measured DIN, PO4-P and chl-a values ranged from 2.7 to 13.1 mmol m?3, from 0.04 to 0.24 mmol m?3and from 0.8 to 5.3 mg m?3, respectively. The observations displayed similar trends to those in Qinzhou Bay. DIN, PO4-P and chl-a indicated acceptable coincidences between simulations and observations (RSD equaled 14.7%, 31.9% and 44.1%, respectively).

Table 3 Mean RSD of modeled state variables (DIN, PO4-P and chl-a) and observations

Moreover, the observed and simulated results appeared to exhibit annual cycles in the bays. During the spring, the intensifying solar radiation and rising water temperature (Fig.2) caused phytoplankton grow rapidly. DIN was depleted by the spring algal bloom in Zhenzhu Bay and Fangcheng Bay with relatively low nutrient supplies. Owing to continuous phytoplankton consumption, PO4-P decreased since late spring. The chl-a concentrations attained their annual maxima as freshwater discharge and nutrient loads increased during the summer months. The algal biomasses declined since autumn and reached their minima in winter.

4.2 Model Validation

The models were verified by the annual mean observing data in 2009, which were kindly supplied by Marine Environmental Monitoring Center of Guangxi Province. According to the monitoring data, the annual riverine nutrient load to Guangxi coastal bays in 2009 was heavier by a factor of 1.5–1.8 than in 2006–2007. The measurement of Zhenzhu Bay was missing in our validation data. The observed DIN, PO4-P and chl-a in Guangxi coastal bays ranged from 5.4 to 35.3 mmol m?3, from 0.40 to 1.02 mmol m?3and from 2.1 to 21.3 mg m?3, respectively (Fig.6). The highest nutrient values were measured in Qinzhou Bay, where the lowest annual phytoplankton biomass was observed. The highest chl-a concentration occurred in Lianzhou Bay, where the nutrient concentrations appeared relatively small. This contrast in chl-a and nutrient concentrations implied that there was a large difference between Qinzhou Bay and Lianzhou Bay in the pathways through which exogenous nutrients were converted into algal biomass. Fangcheng Bay and Tieshan Bay exhibited elevated levels of nutrient and chl-a in 2009. The alteration of nutrient loads and the spatial-temporal differences in sampling strategies might be responsible for the interannual variations.

Fig.6 compares the observed DIN, PO4-P and chl-a values with simulations in the other four bays. DIN and PO4-P were well replicated (the mean RSD were 7.7% and 3.6%, respectively). The chl-a concentrations were also reproduced appropriately with an averaged RSD of 20.6%. Overall, the calculations mimicked the measured data in the bays and represented reasonably the nutrient and phytoplankton dynamics. The results from the validation suggested that the choice of model state variables and parameters properly reflected the actual bay conditions.

Fig.6 Modeled and measured annual mean DIN (a), PO4-P (b) and chl-a (c) concentrations of Fangcheng Bay, Qinzhou Bay, Lianzhou Bay and Tieshan Bay in 2009.

4.3 Sensitivity Analysis

Aiming to identify the main processes and factors thataffect the simulations, sensitivity analysis of parameters was conducted as described by Li et al. (2008). The coefficient of variation (CV) was used to analyze the effect of parameters on the variables (H?kanson, 2000). The results (details are not shown for concision) suggested that the nutrient and phytoplankton simulations of Zhenzhu Bay and Fangcheng Bay were most sensitive to theparameters of water exchange processes and photosynthetic processes (CV-value > 0.5). The parameters related to respiration and mortality of phytoplankton, grazing, excretion and mortality of zooplankton, as well as detritus decay and sedimentation processes also influenced the simulations to some extent (0.1 < CV-value < 0.5). The organic nutrient mineralization rates had weak effects on the simulations (CV-value < 0.1).

The nutrient and phytoplankton simulations of Qinzhou Bay, Lianzhou Bay and Tieshan Bay were most sensitive to the parameters related to water exchange processes, photosynthetic processes, respiration and mortality of phytoplankton, as well as zooplankton mortality processes (CV-value > 0.5). The simulations were less sensitive to the parameters for zooplankton grazing processes (0.1 < CV-value < 0.5). The parameters related to zooplankton excretion, detritus decay and sedimentation, and organic nutrient mineralization processes had weak effects on the simulations (CV-value < 0.1).

Sensitivity analysis of the nutrient and phytoplankton simulations to different factors is shown in Table 4. The results indicated that the variations of simulation were mainly due to riverine nutrient loads. Lianzhou Bay was the most sensitive one among the five bays to external inputs, whereas Zhenzhu Bay was the least sensitive one.

Table 4 Mean sensitivity (% simulation change) of state variable simulations to different factors

5 Discussion

The simulated DIN, PO4-P and chl-a concentrations were compared with the measurements at five Guangxi bays where conditions varied considerably in terms of nutrient, phytoplankton and water exchange. Since the model concepts and equations were the same for the five bays, this highlighted the importance of correct boundary conditions and model parameters. Although the accuracy of nutrient input data and parameters ought to be improved to enhance the confidence of ecological modeling, the simulations appeared to capture the basic seasonal dynamics of the bay systems. Some deviations occurred probably due to the lack of nutrient fluxes across the top and bottom interfaces, especially the sediment-seawater exchange that plays an important role in biogeochemical cycling in shallow bay systems.

In order to control the phytoplankton blooms in highly eutrophicated Lianzhou Bay, strict nutrient reductions are needed. The response curves for the nutrient reductions in our model are shown in Fig.7. The results showed the reduction of phosphorus was more effective, indicating that phosphorus rather than nitrogen was the main limiting factor to phytoplankton growth in Lianzhou Bay. There was a fairly linear relation between chl-a decrease and phosphorus reduction. With low levels of N reduction (0–20% reduction), there was little response in chl-a concentration. To keep the average summer chl-a concentration below 10.0 mg m?3, nearly 70% of P reduction or 80% of N reduction in annual load was required.

Fig.7 Predicted average summer chl-a concentrations resulting from P reduction (dotted lines) and N reduction (solid lines) for Lianzhou Bay.

6 Conclusions

In this study, box models that mechanistically represent key ecological processes were applied to the simulation of nutrient and phytoplankton dynamics in Guangxi coastal bays. In general, there was a good agreement between models and observations, which indicated that the models were valid for the studied bays. In order to control the frequent algal blooms in Lianzhou Bay, nutrient reduction strategy was proposed based on the model calculations. The spatial distribution of nutrient and phytoplankton in the bays was not explored because of the attribute of box model. For future research of Qinzhou Bay, which has a large spatial variation, a three-dimensional ecosystem model may be necessary. Besides intensive monitoring programs in the bays, more efforts are needed to investigate the nutrient loads from rivers, sewages, atmospheric and maricultural sources, as well as bottom sediments, which may provide a better basis for understanding the dynamic behaviors of nutrients and developing protection strategies for the coastal bays.

Acknowledgements

The authors thank all the participants during the cruises. We also thank Beihai Marine Environmental Monitoring Center, State Oceanic Administration of China for providing necessary data for computing and verification the models. We are very grateful to the reviewers for their constructive and valuable comments. This research was supported by National Key Technology Research and Development Program of the Ministry of Science andTechnology of China (2010BAC69B00), Basic Scientific Research Fund of the Second Institute of Oceanography, SOA (JG1210, JG1209 and JT1005), National Basic Research Program of China (2011CB409803) and Special Expenses Program of Scientific Research in Marine Commonweal Industry (201205015).

Baretta, J. W., Eben?h, W., and Ruardij, P., 1995. The European Regional Seas Ecosystem Model: A complex marine ecosystem model. Netherlands Journal of Sea Research, 33: 233-246.

Chau, K. W., Cheng, C. T., and Li, C. W., 2002. Knowledge management system on flow and water quality modeling. Expert System with Applications, 22 (4): 321-330.

Chau, K. W., 2004. A three-dimensional eutrophication modeling in Tolo Harbor. Applied Mathematical Modelling, 28: 849-861.

Chen, Q. Y., 2001. Assessment on situation of water quality in Lianzhou Bay of Guangxi. Marine Environmental Science, 20 (2): 62-66 (in Chinese with English abstract).

Cheng, C. T., Wang, W. C., Xu, D. M., and Chau, K. W., 2008. Optimizing hydropower reservoir operation using hybrid genetic algorithm and chaos. Water Resources Management, 22 (7): 895-909.

Cloern, J. E., 2001. Our evolving conceptual model of the coastal eutrophication problem. Marine Ecology Progress Series, 210: 223-253.

Editorial Board of Annals of Bays in China (Ed.), 1993. Annals of Bays in China, 12. Ocean Press, Beijing, 335pp, (in Chinese).

H?kanson, L., 2000. The role of characteristic coefficients of variation in uncertainty and sensitivity analyses, with examples related to the structuring of lake eutrophication models. Ecological Modeling, 131: 1-20.

Han, H. Y., Li, K. Q., Wang, X. L., Shi, X. Y., Qiao, X. D., and Liu, J., 2011. Environmental capacity of nitrogen and phosphorus pollutions in Jiaozhou Bay, China: Modelling and assessing. Marine Pollution Bulletin, 63 (5-12): 262-266.

He, B. M., and Wei, M. X., 2009. Reasons and mechanism caused red tide in Beihai Bay. Marine Environmental Science, 28 (1): 62-66 (in Chinese with English abstract).

Irigoien, X., and Castel, J., 1997. Light limitation and distribution of chlorophyll pigments in a highly turbid estuary: The Gironde (SW France). Estuarine, Coastal and Shelf Science, 44 (4): 507-517.

Kashiwai, M., 1984. The concept of tidal exchange and the tidal exchange ratio. Journal of the Oceanographical Society of Japan, 40: 135-147 (in Japanese with English abstract).

Kuusisto, M., Koponen, J., and Sarkkula, J., 1998. Modelled phytoplankton dynamics in the Gulf of Finland. Environmental Modelling & Software, 13: 461-470.

Lan, W. L., 2011. Variation of organic pollution in the last twenty years in Qinzhou Bay and its potential ecological impacts. Acta Ecologica Sinica, 31 (20): 5970-5976 (in Chinese with English abstract).

Lan, W. L., and Peng, X. Y., 2011. Variation characteristics of nutrient concentrations in the Tieshangang Bay. Guangxi Sciences, 18 (4): 380-384 (in Chinese with English abstract).

Lan, W. L., Wang, X. H., and Li, M. M., 2011. Phytoplankton community structure in Qinzhou Bay during flood season by analysis of HPLC photosynthetic pigment signatures. Acta Ecologica Sinica, 31 (13): 3601-3608 (in Chinese with English abstract).

Li, K. Q., Wang, X. L., Shi, X. Y., Zhu, C. J., and Li, R. X., 2007. Modeling nitrogen and phosphorus transport and transformation in pelagic ecosystem in mesocosm in Jiaozhou Bay. Acta Oceanologica Sinica, 29 (3): 76-83 (in Chinese with English abstract).

Li, Y. B., Wang, X. L., Han X. R., Li, K. Q., Zhao, X. X., and Shi, X. Y., 2008. An ecosystem model of the phytoplankton competition in the East China Sea, as based on field experiments. Hydrobiologia, 600: 283-296.

Long, X. H., Jiang, Q. H., and Deng, Y., 2011. Study on nutrient salt variation trend and control countermeasures of offshore marine area in Qinzhou Bay during ‘the 11th Five-Year Plan’. Environmental Science and Management, 36 (8): 65-68 (in Chinese with English abstract).

Los, F. J., and Wijsman, J. W. M., 2007. Application of a validated primary production model (BLOOM) as a screening tool for marine, coastal and transitional waters. Journal of Marine System, 64: 201-215.

Ménesguen, A., Cugier, P., Loyer, S., Vanhoutte-Brunier, A., Hoch, T., Guillaud, J., and Gohin, F., 2007. Two- or threelayered box-models versus fine 3D models for coastal ecological modelling? A comparative study in the English Channel (Western Europe). Journal of Marine Systems, 64: 47-65.

Muttil, N., and Chau, K. W., 2007. Machine learning paradigms for selecting ecologically significant input variables. Engineering Applications of Artificial Intelligence, 20 (6): 735-744.

National Academy of Science (NAS), 1972. A Report of the Committee on Water Quality Criteria. NAS, Washington, D. C., 19-23.

Park, K., Jung, H. S., Kim, H. S., and Ahn, S. M., 2005. Threedimensional hydrodynamic-eutrophication model (HEM-3D): Application to Kwang-Yang Bay, Korea. Marine Environmental Research, 60: 171-193.

Parker, D. S., Norris, D. P., and Nelson, A. W., 1972. Tidal exchange at Golden Gate. Journal of the Sanitary Engineering Division, 98 (2): 305-323.

Qiu, S. F., 2001. Investigation of marine environmental quality in the offshore of four bays of Guangxi. Guangxi Sciences, 8 (3): 236-240 (in Chinese with English abstract).

Redfield, A. C., Ketchum, B. H., and Richards, F. A., 1963. The influence of organisms on the composition of seawater. In: The Sea. Vol.2, Hill, M. N., ed., Interscience, New York, 26-77.

Savchuk, O. P., 2002. Nutrient biogeochemical cycles in the Gulf of Riga: Scaling up field studies with a mathematical model. Journal of Marine System, 32: 253-280.

Stow, C. A., Jolliff, J., Mcgillicuddy, D. J., Doney, S. C., Allen, J. I., Friedrichs, M. A. M., Rose, K. A., and Wallhead, P., 2009. Skill assessment for coupled biological/physical models of marine systems. Journal of Marine System, 76: 4-15.

Strickland, J. D. H., and Parsons, T. R. (eds.), 1972. A Practical Handbook of Seawater Analysis. Fisheries Research Board of Canada, Ottawa, Canada, 185-202.

Sun, H. L., Liao, Q. Y., and Huang, W. M., 2002. Field experiment of the solute dilute diffusion and numerical simulation of the COD tidal diffusion near the coast of Guangxi, China. Acta Oceanologica Sinica, 24 (5): 115-124 (in Chinese).

Wang, B. D., Wang, X. L., and Zhan, R., 2003. Nutrient conditions in the Yellow Sea and the East China Sea. Estuarine, Coastal and Shelf Science, 58: 127-136.

Wang, P. F., Martin, J., and Morrison, G., 1999. Water quality and eutrophication in Tampa Bay, Florida. Esturine, Coastal and Shelf Science, 49: 1-20.

Wu, C. L., and Chau, K. W., 2006. Mathematical model of water quality rehabilitation with rainwater utilization–A case study at Haigang. International Journal of Environment and Pollution, 28 (3-4): 534-545.

Xie, J. X., Cheng, C. T., Chau, K. W., and Pei, Y. Z., 2006. A hybrid adaptive time-delay neural network model for multistep-ahead prediction of sunspot activity. International Journal of Environment and Pollution, 28 (3-4): 364-381.

Xin, M., Wang, B. D., Sun, X., and Xie, L. P., 2010. The distribution characteristics and seasonal variation of nutrients in Guangxi coastal area. Marine Sciences, 34 (9): 5-9 (in Chinese with English abstract).

Xu, J., and Hood, R. R., 2005. Modeling biogeochemical cycles in Chesapeake Bay with a coupled physical-biological model. Estuarine, Coastal and Shelf Science, 69: 19-46.

Zhang, C. X., Huang, L. L., Zhong, Q. P., Dai, Z. R., Shen, T. M., and Yang, B., 2011. The analysis and evaluation of seawater quality in Qinzhou Port. Guangzhou Chemical Industry, 39 (15): 140-141 (in Chinese with English abstract).

Zhao, J. S., Wang, G. Y., Chen, Z. S., and Geng, S. J. (eds.), 2002. Analysis of Effective Utilization of Marine Environment in Guangxi Coastal Area, Beibu Gulf. Ocean Press, Beijing, 77-131 (in Chinese).

Zhao, M. Y., Cheng, C. T., Chau, K. W., and Li, G., 2006. Multiple criteria data envelopment analysis for full ranking units associated to environment impact assessment. International Journal of Environment and Pollution, 28 (3-4): 448-464.

Zhou, S. L., Chen, J. F., and Ma, S. S., 1996. Preliminary study on the laws of seawater exchange in Dingzi Bay. Marine Fisheries Research, 17 (1): 49-56 (in Chinese with English abstract).

Zhuang, J. L., Jiang, F. J., Ke, K., Xu, M. B., Zhang, R. C., Wang, Y. B., Peng, Y., and Chen, B., 2011. Monitoring and analysis on a seawater anomaly in Qinzhou Bay. Guangxi Sciences, 18 (3): 321-324 (in Chinese with English abstract). Zhuang, J. L., Xu, M. B., Zhang, R. C., Jiang, F. J., Wang, Y. B., and He, B. J., 2010. Anniversary variation of phytoplankton abundance in Fangchenggang Bay of Guangxi. Guangxi Sciences, 17 (4): 387-390 (in Chinese with English abstract).

(Edited by Qiu Yantao)

(Received June 26, 2012; revised August 6, 2012; accepted April 22, 2013)

? Ocean University of China, Science Press and Springer-Verlag Berlin Heidelberg 2014

? Corresponding author. Tel: 0086-532-88962016

E-mail: wangbd@fio.org.cn