Evaluation of the Nutrient Cycle in the Seagrass Bed Yuko Oshima and Michio J. Kishi Ocean Research Institute, University of Tokyo 1-15-1, Minamidai, Nakano-ku, 164 Tokyo, Japan Abstract Seagrass bed plays an important role in the marine ecosystem in the estuarine area, For the purpose of conservation and management of these seagrass beds, it is necessary to know the systein and the role of the seagrass A tluee dimensional physical-biotogical model (MK-3) was modified in order te include seagrass bed and thea applied to the Akkeshi Lake, which is situated at a Pacific coast of Hokkaido, in the northern part of Japan. Nitrogen and phosphorus fluxes were evaluated to investigate the role of the seagrass on the nutrient cycle. At frst a simple box model was developed. We divided the lake into three boxes to estimate roughly the characteristics of the autrient cycle based on the observed data. The box model revealed that aitrate is supplied into the lake through the physical processes ie., river discharge, advection from outer region, and is consumed biologically inside the lake. On the other hand, phosphete is supplied from inside the lake and flows out, Second, in order to explain this different behavior of nitrate and phosphate, three dimensional physical-biological coupled model was applied to the Akkesbi Lake, The model was runt under the average metevrological conditions of April for 10 days, then calcuration was continued using, real time condition (wind, precipitation etc.) during 16-27th April 1996. The calcuration was cattied out for two cases, one was with eelgrass and the other was without eelgrass. The case without eelgrass, nitrate and phosphate were both decreased by biological and chemical processes inside the lake. The case with eelgrass, nitrate was decreased. bout phosphate was produced inside the lake which was estimated by the box model. Furthermore, the eelgrass was important than phytoplankton ¢o produce the particulate organic matter. ‘The daily variation of chlorophyll a. and. aitrate which were observed during 18-27th April 1996 couldn’t be appeared in the model results although the averaged. values coincided with the observed values. This may be because the effect of melting of the snow in the river discharge, its data is not included, 1. Introduction Seagrasses have been recognized to play an impor- tant role in the nutrient budget in the coastal marine ecosystem. They usually form the dense and highly productive meadows in shallow regions in Japan. They act as a primary producer, a provider of spawning, hid- ing, and feeding fields for fishes and a habitat for epi- phyte, and a resister for water current. For example, Zostera marina L. (eelgrass), which is characterized as, 2 wide eelgrass bed over shallow sandy aceas, can reach the same order of produetion as grassiands (Aioi,1981;, ‘Mann,1982). It can take up nvtrients by leaves and by roots (Iigumi and Hattori, 1982), Recentry, many seagrass beds are destroyed by ei- ther reclamation or coastal-protecting construction, The degree of these influence, however, has nat been well: estimated because the quantitative contribution of sea- _gtasies to the nutzient budget still remains unknown in the coastal zone of Japan. ‘There have been made some numerical models ia- cluding seagrass ecosystem. For example, Bach(1993) ‘nade @ one-dimensional general eutrophication model for coastal marine areas that had been extended to in- clude the seasonal variations in growth of eelgrass. The model had been developed and calibrated based on data from a shallow area in the southern part of Denmark, ‘The basic model described the growth of phytoplankton, and cooplankton and nutrient dysamics. The eelgrass submodel described seasonal and regional variations in production and biomass of above and below ground parts, This model was seen to be able to describe the 1748, seasonal variations in the data material, however, we cannot sce the nutrient budget in these area with this model. A three-dimensional numerical ecosystem model was developed by Kishi and Uchiyama (1993,1995) for the ‘quantitative management of mariculture which includes seagrass uptake of nitrogen, This model focused on the dissolved oxygen budget of bay area where the mati- calture of fish, shellfish, seagrass and seaweed is carried out. ‘Their models include seagrass and seweed as 2 ‘external forcing function the biomass of which is given following the observation, and which consume nutrient from ambient water. Their models give an averaged features of ecosystem in the bay area. Here we developed a new concept model which can follow the daily variation of the ecosystem. For this purpose, a three dimensional physical-ecological model hhas been developed and applied to the Akkeshi Lake, which is situated at pacific coast of Holekaido, in the northern part of Japan(Figure 1). The nitrogen aud phosphorus Guxes are evalucted based on the simulated results to investigate the role of seagrass in the autrient budget quantitatively. 2. Model Description "his ecosystem model is developed by coupling 2. three-dimensional physicat submodel with a biological submodel that based on nitrogen and phosphorous. The physical model isa three-dimensional free surface model, (MK-3) developed as a physical submodel(Kawamiya et al,1096). The biological model is ten-compartment, nitrogen and phosphorous-based model. This model isN Bekanbeushi Riv. Kkeshi Bay Figure 1: Schematic view of Akkeshi Bay and Akkeshi Lake, Hokkaido Prefecture, Japan, expanded after the KICYS model(Kawamiya et al,1995) which is seven-compartment nitrogen-based model, and 1we have been added the biological processes of eelgrass ‘The detailed formula of these two models were given by Kawamiya et al(1998) and Kawamiya et al(1995), and a brief description is given here. 2.1 Physical model ‘The governing equations are as follows: ff Sova: —v0av + VN AnVan) + 2 av “8 Get fk xa a ? 9 = 0T,3) (8) whore » ip the horizontal velocity veetor, w the ver- tical velocity, f the Coriolis parameter, g the gravity constant, Uj the horizontal divergence operator, ¢ the sea surface elevation, H the depth, 4y and Ay are the horizontal and vertical eddy viscosities, Ky and Ky the horizontal and vertical diffusivities, and T and $ the water temperature and salinity, respectively. We ap- ply Knudsen’s equation for calculation of density of sea water(eq.(5)} ‘We divide the water column into 8 levels vertically, and che horkzontal grid size and other parameters used in the physical and biological model are shown in Table 1 1749 Figure 2: The material lows based on nitrogen and phosphorous. ‘The Bekanbeushi River is considerd as the resource of fresh water to the Aldkeshi Lake, The averaged river discharge is estimated from observed data{ Hokkaido, 1981), and the effect of the temporal precipitation is added im- mediately according to Takahashi(1978) by lopalS)* seu ( 3) (8) where Q is the river discharge due to precipita- tion(ton/s}, C3 the outflow coefficient, R the precipi. tation (mm/hour), A the Bekanbeushi River basin, 3 the average slope of the basin, P the constant and these values are shown in Table 1, too. The vertical diffusion coofficient and viscosity coeflicient are obtained from formulas by Pacanowski and Philander(1981). Q 2.2 Biological model The governing equation of passive tracers is ~(09}C WE + Val, TAC) + E(«Z) +a where C is the biological compastment and Q(C) the biologieal processes, ‘The horizontal diffusion coe Scent (Z(,) is shown in Table 1 m7 ‘Ten biological compartments based on nitrogen and phosphorous flow are considered as phytoplankton (CHL), zooplankton (Z00), nitrate (NO3), ammonium (NE4), phosphate (POs), particulate organic nitrogea (PON), particulate organic phosphorous (POP), dissolved or- ganic niteogen (DON), dissolved organic phosphorous (DOP) and eelgrass (2M). ‘The parenthesized characters are their mathemat- ical symbols and ace used in the text, too. Nitrite is included in NO3. Figure 2 shows a schematic view of the ecological model ‘We assumed the ealgrass only as an external forcing, since the time evolution of eelgrass biomass is difficult‘Table 1: Parameter values in the physical and ecological model ‘Hovwontal grid size 300 "ime interval 20 | see. ‘Average Amplitude of Fide | 0.5 | Period of Tide 12 | Bours TS x0" [ems LO 10" ews (5x10 emt/s Averaged River Discharge | 10 | ton/s Ge os a 309.1 [at [ 5 GOL 7 I P 4 = ‘Table 2: The average physical conditions in April. Sea surface temperature [6.0 [°C] Soler radiation maximus [1.29 | Wy/min~ River discharge 0.0] ton/s Wind speed 30 | wife to caleulate based on the physical conditions and am- bient nutrient concentration, and its biomass incleases twice linearly, through April to June, because eelgrass Biomass in June was twice in April at Odawa Bay and Nabeta Cove (Aloi,1981}. Based on the observation data of eelgrass biomass in June in Akkeshi Lake (Re- search and development af the coast Co. Ltd.,i991), celgrass biomass in June is sot as 1600 g/m#(wet weight) and That in Apel is supposed as 800 g/t? (suet weight). We also assume that cry weight is quarter of wet weight and its height is 1m, Nitrogen and phosphorous con- centrations in eelgrass are converted using the Cidry weight catio of O.f(by weight; Mukai et.al,1979), C:N ratio of 15( Aoi and Mukai, 1980) and N:P ratio of 18(Atkin- son and Smith, 1983}. Biological processes related with eelgr loves Photosysthesis: The uptake by leaves is taken into consideraion as a photosynthesis of eelgrass, because ve don’t include the benthic ecosystem in our model Nitrogen uptake by eelgrass leaves is set asa linear fune- tion with the ambient nutrient concentration following to Short and MeRoy(1984). We applied the Michaelis: Menten equation for Thalassia hempeichii by Stapel et 421,(1996) on the phosphate uptake by eelgrass leaves, because we don’t have the data, Decrease: The decrease of eelgrass is separated into two parts, depending on depth and temperature(Bach, 1993), ‘The part depending on depth is supposed as mortality of eelgrass, and the other is eospiration, First physical-biological coupled model was run for ton days under the average physical conditions in April in order to get a steady state, secondary it was run using the actual boundary conditions (solar radiation, sea surface temperature, wind stress and precipitation) feom Lek April to 27th Apsil 1996. The case studies ‘with and without eelgrass were eazried out in order to discuss the tole of eelgrass on the material budget ia te Lake ware as fol- 1750 * Observed Station 3 F\_ Station 2 Figure 3: Observed stations in Akkeshi lake from 15 to 26 April 1996, Chlorophyit a (ug/l) i4 2B 12, I 10. 5 a] aI en pesan 46 20 25 (date) Figure 4: Time-dependent values of chlorophyll a ob- tained by the model (solid line) and observations (sur- face, circles and 1 m depts, bares), 3. Observed Data In Akkeshi Lake we collected samples eight times from 15th to 26th April 1996. Weter semperature, salinity, chlorophyll a, nitrate, phosphate and PON con- centration were analysed ance a day at 3 stations shown in Figure 3. Water samples were collected at 0 m and I'm depth at Stations 1 and 2, and five samples (0,1,2,8,5,7-5 m depth) at Station 3. At each layers, temperature and salinity were measured by T-S meter (Model 33 §-C-T meter, YSI). Chlorophyll a was de- termined fiuorometrically with a Turner Designs 10R by using N,N-Dimethylformamide extraction method. Nutrient was measured by Technicon Auto Analyser {1 with DISMIC.25 (0.45 um diameter} filtered sea water which was flozen immediately after sampling. PON was analyzed by Dr-Hicumi at Hokkaido National Fisk cries Research Institute. Furthermore, at the mouth of Bekanbeushi River, we observed water temperature, nitrate and phosphate concentration 4 times during that period.NO3 (umol/ly (date} Figure 9: Time-dependent values of nitrate obtained by the mode! (solid line) and observations (surface, circles and 1 m depth, boxes) PO, (umolA) 06: 05+ 04a 03 02. Op a ot 46 20 (date) Figure 6; Time-dependent values of phosphate obtained by the model (solid line) and observations (surface, cir- cles and Lm depth, boxes). PON (umol/l) & TT oT 7 4 j | 6 1 ole t* | 5 ol i | 4 re aa Pee tt | on | | ot Ppt ttt a6 20 2 (date) Figure 7: Time-dependent values of PON obtained by the model (solid line) and observations (surface, and 1 m depth, boxes) eles 1751 Figure 8: Simulated horizontal distribution of phos. phate, (a)without eelgrass and (b)with eelgrass, Figure 9: Devided three areas of Aldkesbi lake; riverine, estuarine and marine area, 4, Results and Discussion 4.1 Time dependent features at station 2 ‘Time series data using a coupled model at the point corresponding to upper layer of Station 2 were described in Figures 4-7 together with observed data. The magni- tude of the concentrations show a good agreement with observed data although simulation cannot follow a de~ tailed daily ductuation. Especially the sudden inclease of chlorophyll aon 22nd April cannot be followed by simulation, Snow fall was observed from 18-19th April, however, the effect of melted snow is not cousidered in this model. ‘This is one possibility for the explanation of descrepacy. 4.2 Horizontal distribution ‘We can see the difference between the case without eelgrass and the case with eelgrass of horizontal distri- bution especially in phosphate( Figure 8).Table 2.(anol/day) Nutrient due without eelgrass in Box = NOs PO; c Advection__|_172_| L383 | g a Poplakion Photosynthesis [ase] ‘Bin foo Peplankion Respieation || 14509610 | 1410] Zocnannon I Niteifieation [Cis “T0807 [fast ad { Nutrient Release - 35300 _| 3530 foo #20 [Decomposition of BON 6500_| 1060. Eearae [Decomposition of PON Ti Frooo (xcretion by 232, mdf [[Sumrof Other Processes _] He ase “sis | ‘Table 4: Nutrient flax with eelgrass in Box 2,(mol/day) NOx TNE c [Lass (P-plankton Photosynthesis “8730 } Pplankton Respiration 150 Nitrification ~~ Nuiient Release 3100 [Decomposition of DOM. mc Decomposition of POM [1880 Excretion by Zooplanlcton [ass Eelgrass [35 B82 1070 Based upon the observed salinity data in Akkeshi Lake by Tiznmi et «1(1995), itis easy to elasify the lake into three region, Le., riverine (Box 1 hereafter), eatu- arine (Box 2 hereafter) and marine area (Box 3 here alter)(Figure 9} 4.8 Inflow and Outflow of nutrient in Akkeshi Lake All biological and physical terms were integrated in the box. All processes in Box 2 were averaged during 16th Apeil and 26th April and indicated in Tables 3 and 4. In these tables, the item of “Sum of Other Processes” indicates the sum of all terms of Biological and chemical processes Tn the case without eelgrass, the sum of other pro- cesses are negative in all quttient (NOS, NH and PO4) budgets, ie, all nutrient are consumed in the Box 2 e5- peciaily through photosynthesis by phytoplankton. On the other hand, ia the case with eelgrass, the sign change to positive in NH and PO4. This is mainly caused by nutrient release fom sediment, If production by phytoplankton using ammonium is named ‘reproduction’ and that using nitrate is named ratio is decided as division of ce production by total production. ‘The fratio in the case with celgrass (0.07) is half of that in the case without vvligrass (0.4). This result shows that the aitzogen re- production is accelerated by the eelgrass Figure 10 shows the simulated nutrient Sow in Box 2 during one day of 23rd April with eelgrass. This gure supports the above dliscussion "new productio 1752 Figure 10: Nutrient flow with eelgrass. ‘The effect of the role of eelgrass in the estuarine ecosystem have been thought as large, but few people said so quantitatively. ‘The mutrient uptake of phytoplankton and eelgrass show little competition, In the case with eelgrass, though, eelgrass uptake nutrient, repreduction of nutrieat is ac- celerated by the effect of eelgrass. 5. Acknowledgements ‘We would like to thank Prof, H.Mulai of Alekeshi Marine Biological Station of Hokkaido University and Dr. Hlizumi of Hokkaido National Fisheries Research Institute for their help of our research and observation ‘Thanks are also given to Dr. M.Kawamiya of Center for Climate System Research(CCSR) for his help to recon. struct a three dimensional physical-biological coupled. model. Dr. T.Komatsu and Dr. K.Aioi of Ocean Re search Institute helped us by providing the knowledge of eelgrass 6. References Aioi.K., Produetion of Belgrass, Zostera ma- sina L., and Ite Growth Types in Sub- tidal Environments in Odawa Bay, Cen- tral Japan. A Dissertation, Presented to the Faculty of Science, Tokyo Metropol tan University for the Degree of Doctor of Philosophy, 1981 AioiK., The grassy plain in submarine. O:- Suchi Marine Research Center Report, 13, 38-40,1987. (in Japanese) Aiei,K., Ecology and protection of seagrass Saishucto-Shilk, 51-8, 392-350, 1989, (in Japanese) BachH.K., A dynamic model describing the seasonal variations in growth and the distribution of eelgrass (Zostera masioa L.) L. Model theory. Ecological Mod- elling, 5, 31-30, 1999. Research and development of the coast Co.Ltd., The existing condition of fishing environment of Akkeshi area. 1991.(in Japanese) Hokkaido, Report on Comprehensive Devel- opment Plan of Coastal Water in the Vicinity of Akkeshi, The Holkside Gov- ernment Office, Sapporo. 1981.(in Japanese) Hizumi,H., Interaction between the root of lgrass and soil, Otsuchi Marine Re- search Center Report, 13, 49-52, 1987. (Gn Japanese) Higumi,H., Nutrient uptake of Seagrasses. Monthly, Ocean, Vol.21, No.6, 317-221, 1989. (in Japanese) TizumiH. and A.Hattori, Growth and Or ganic Production of Belgrass (Zostera ma- ina L.) in Temperate Waters of The Pa- cifie Coast of Japan.IIT. The Kinetics of Nitrogen Uptake. Aquat. Bot., 12, 245- 256, 1982, Sizumi,H, AHattori and C.P.McRoy, Ni- trate and Nitrite in Interstitial Waters of Eelgrass Beds in Relation to The Rhizo- sphere. J. exp. mar. Biol. Beol., Vol-AT, 191-201, 1980. Tizumi,}, AHattori and C.P.McRoy, Am- onium Regeneration and Assimilation in Belgrass (Zostera marina) Beds. Ma- ine Biology, 68, 59-65, 1982, Hauoigt, $.Taguchi, T.Minami, HMukai and S.Maekewa, Distribution and vari ability of nutrients, chlorophyll a, par- ticulate organic matters, and their car- bon and nitrogen contents, in Akkeshi- Ko, an estuary in northern Japan, Bul Hokkaido Nail. Fish, Res, Iast., No.59, 49-87, 1995 KawamiyaM., MJ-Kishi, MD-Kawser A. and T.Sugimoto, Causes and consequences of spring phytoplankton blooms in Ot- suchi Bay, Japan. Continental Shelf Re- search, Vol.16, No.13, 1683-1698, 1998, Kawamiya,M., M.J.Kiehi, Y. Yamanaka and N.Suginohara, An Ecological - Physical Coupled Model Applied ta Station Papa. Journal of Oceanography , Vol.5l, 635- 664, 1995. KishiMJ., and M.Uchiyama, Nitrogen Cy- cle Numerical Model of Mariculture and its Application to Oyster Culturein Nanao- Sei-Bay, Japan. Proceedings of The Sec- ‘ond International Conference on Pish- eries and Ocean Industrial Development lor Productivity Enhancement of the Coastal Waters (FOID'93), 23-40, 1993 Kishi,MJ.,and M.Uchiyama, A three-dimensional numerical model for a maricultuse nitro- ren cycle: case study in Shizugawa Bay, Japan, Fish. Oceanogr., 4:4, 203-316, 1995, MeRoy,C.P, and RJ.Darsdate, Phosphate Absorption in Eelgrass. Limaol. Oceznogr., 1: 6-13, 1970. McRoy.C.P. RJ Barsdate and M.Nebert, 1753 Phosphorous Cycling in an Belgrass (Zostera marina L.) Ecosystem. Limnol. Oceanogr., 17(1), 88-67, 1972. Mulai,HL, K.Aioi, LKoike, H.ligumi, M.Ohtsu and A-Hattori, Growth and Organie Pro- duction of Eelgrass (Zostera marina L.) in Temperate Waters of The Pacific Coast of Japan. I. Growth Analisys in Spring- Summer. Aquat. Bot., 7, 47-58, 1979. Pacanowski,R.C. and $.G.H.Philander, Pa- rameterization of vertical mixing in nu- merical models of tropical oceans. J. Phys. Oceanogr. 11, 1443-1451, 1981 Short,P.T, and C.PMeRoy, Nitrogen Up- take by Leaves and Roots of the Sear grass Zostera marina L.. Botanica Ma- Fina, Vol XXVIL, 547-555, 1984. Stapel,J., T-L-Aarts, B.H.M.van Duynhoven, ‘4.D.de Groot, P.H.W.van dea Hogen and M.A.Hemminga, Nutrient uptake by leaves and roots of the seagrass Thalassia hemprichil the Spermonde Archipelago, Indone- sia. Mar. Ecol. Prog. Ser., vol.134, 195-206, 1986. ‘Takahashi, Y., Hydrology of Rivers. Kyoritsu Syuppan Pub. Co, Ltd, 70-73, 1978. ‘Tatsukawa,K., T-Komaten, K.Aioi and K.Morita, Distribution of Seagrasses ia Funakoshi Bay. Otsuchi Marine Research Center Report, 21, 38-47, 1996. (In Japanese)