Arabian Sea Micom
Progress report - April, 2001
Incident solar radiation
The radiation flux originally calculated by MICOM is not compatible with that required by the NPZD model, because MICOM calculates a net radiation flux to and from the surface and takes into account the upper mixed layer rather than solar radiation reaching the ocean surface. Therefore solar radiation at the surface boundary and initial conditions have been set for the physical model (MICOM) using data from the Max Plank Institute Heat Flux and Surface Radiation Climatology at http://ingrid.ldeo.columbia.edu/SOURCES/.OBERHUBER/.solr/, which is consistent with the radiation/heat flux climatology originally used in MICOM. Anastasia Romanou and Josefina Olascoaga converted these data as well as cloud cover data (Oberhuber data set) from a 1.4š grid spacing to that of the model's 0.35š grid size by linear interpolation. In order to develop a consistent light field used by the NPZD model, missing data that were present in the surface solar radiation dataset at various grid points for certain months were filled in by linear spatial interpolation by Nasseer Idrisi. A time interpolation scheme for forcing data smoothing is done in MICOM. Light extinction at depth is calculated with a subroutine supplied by Victoria Coles of the University of Maryland and modified for the Arabian Sea MICOM by Anastasia Romanou. Nasseer Idrisi developed code to simulate the diurnal variation in incident solar radiation dependent on time of day and day in year; this routine was then implemented in MICOM by Anastasia Romanou. Rupert Minnett has been extracting and processing surface chlorophyll and surface photosynthetically active radiation (PAR) from SeaWiFS data for comparison with model output.
State variables including nitrate, chlorophyll fluorescence, ADCP-derived biomass and total organic nitrogen (TON) from the 1994-96 Joint Global Ocean Flux Study (JGOFS) Arabian Sea Expedition were used to initialize the NPZD model. Robin Kovach processed nitrate, fluorescence and ADCP-derived biomass using objective analysis, and a mean value was calculated for 10 m depth intervals from 0-200 m. Detritus as total nitrogen was taken from the data at http://usjgofs.whoi.edu/jg/dir/jgofs/arabian/, using all cruises and stations where TON was measured and a mean value calculated at each depth and interpolated to 10 m intervals. For nitrate and detritus, values were calculated to 3000 m at 100 m intervals from 300 to 3000 m.
Additions/modifications to the physical model
The physical model developed by Oleg Esenkov for the Arabian Sea in MICOM is being used as the foundation in this project effort.
1. Various time steps were tested to determine a convenient and stable system. The original time step of 19.2 minutes was not practical for output writing, especially for the biological variables, which are dependent on time of day (phytoplankton dynamics). The most convenient, as well as stable time step to use was that of 15 minutes. Lower than this time step, e.g., 10 minutes led to instabilities in u-v velocities near the northern boundaries of the model domain.
2. A depth scheme was developed to accommodate the NPZD model coupling to MICOM that converts isopycnal layers to z-coordinates when NPZD as well as PAR routines are being accessed by MICOM.
3. As mentioned earlier, the solar radiation and cloud cover inputs and depth PAR routine has been added to MICOM for use by the NPZD routines.
4. Josefina Olascoaga is in the process of developing HYCOM for the Arabian Sea for eventual migration of the biological model from MICOM to HYCOM.. HYCOM will allow greater flexibility in depth-resolving in the upper mixed layer of the physical model.
The biological model
The original NPZD model model used for this project was first developed by Ivan Lima for POM in a generic model configuration. Additions and modifications are by Nasseer Idrisi for use in the Arabian Sea MICOM; Anastasia Romanou made necessary modifications to merge the routines in MICOM. To date, routines for nutrient dynamics (nitrogen) and phytoplankton growth have been successfully coupled to MICOM. Initialization and coupling of the dynamical NPZD state variables will be completed by the end of April. Changes to the POM-NPZD include (note: parameterizations are open to further modifications depending on model behavior and stability):
1. The model has been reparameterized using nitrogen (nitrate and ammonia) uptake kinetics specific for the Arabian Sea from McCarthy et al. (1999: DSR II 46: 1623-1664). Although this study was conducted during the Northeast Monsoon, uptake parameters are used for all seasons, assuming uptake depends on nutrient concentrations alone.
2. Photosynthesis vs. irradiance parameters used to calculate primary productivity are developed from Toon et al. (2000: DSR II 47: 1249-1277) and calculated as N-production by converting from C-production.
3. The primary productivity equation for phytoplankton growth using the reparameterizations has been changed to a form that uses a temperature-dependent function for maximum photosynthetic rate, as well as nutrient- and irradiance-dependent growth rate. The routine for phytoplankton growth has been successfully coupled and run in MICOM.
4. A temperature-dependent growth and metabolism function has been added to the zooplankton component of the NPZD model. This is necessary to simulate seasonal variation in growth rates for the zooplankton.
5. Since there are no parameters developed for Arabian Sea zooplankton dynamics, parameterization are initially taken from Anderson et al. (2000: DSR I 47: 1787-1827) for the North Atlantic Gulf Stream, and further tuned to the Arabian Sea.
6. Also in the process of development is the bioenergetic IBM specific to Calanoides carinatus using a model developed by Hal Batchelder for Metridia longa, and C. carinatus population parameters from the Bengula Current. Two Lagrangian particle routines (off-line and on-line) are being developed for implementation in the Arabian Sea MICOM. The effort is also comparing the ONR/URIP derived models of fully structured copepod models by Olson/Pascual/Davis and Batchelder et al. The first model is continuous in time and uses a Huntley/Lopez metabolic scaling while the Batchelder model is discrete with a slightly different metabolic structure.
7. Lagrangian implementation of the structured biological codes are necessary because these inherently involve the history of conditions a population experiences. The formulation of the Lagrangian code in the MICOM model is being reviewed in collaboration with Chassignet's and Cowen's research groups. The issue involves different schemes for parameterization of mixing at the subgrid scale. The final tests should be completed by the end of June 2001. This test and the completion of the basic biological models will allow the completion of the first simulation by fall 2001.
Physical Data Analysis
A set of seven day composite images of the Somalia coast have been used to produce an analysis of the upwelling fronts and the area of habitat accessible for Calanoides carinatus. Of particular interest is the collapse of the system at the end of the monsoon where significant portions of upwelled water enter the Gulf of Aden. This marginal sea has the potential of being a major retention area for C. carinatus. Undercurrents coming from the Gulf back onto the Somali coast potentially provide a means of seeding the Somali populations. Currents eastward onto the Oman coast are currently being investigated in the MICOM model output. A paper on the undercurrents off Somalia has just been revised for Deep-Sea Research (Esenkov and Olson, 2001). A basic analysis of the surface drifter data is also complete. Methods for T/S models for water mass mixing developed with ONR funding have been modified to consider the dilution of the diapausing population.
Laboratory and data analyses of Calanoides carinatus
As of April, 2001, Calanoides carinatus and Eucalanus crassus abundance data have been extracted from copepod abundance data files generated by the Institute for Biology of the Southern Seas (IBSS) for samples collected during US JGOFS Arabian Sea cruises TN039 and TN043. Partial data extractions have been done for TN045, TN050 and TN054. This process is ongoing and will eventually also include data collected during the GLOBEC Arabian Sea cruises MB9503 and MB9506. Over two hundred individual C. carinatus have been sorted from preserved samples collected during the US JGOFS cruises for measurements of dry weight and prosome length. We've also estimated lipid sac volumes on over one hundred of these specimens. We are currently comparing fresh dry weight measurements made during the US JGOFS field work with preserved dry weights of C carinatus stage 5 copepodites to allow the estimation of fresh dry weight of individuals based on analyses of preserved samples. Copepod sorting and measurements will continue through the summer to acquire data necessary for hypothesis development, biological modeling parameterization and validating model output. An undergraduate student (Lundsten) who was assisting with this work has departed the University of Miami; however, we have another undergraduate who will begin working in May in the Honors Summer Research Program. This student will also be trained to sort, measure and weigh individual copepods from Arabian Sea samples for model input and validation. In addition to these measurements, we will look for indications of gonad development in deep, diapausing stage 5 copepodites near the end of the diapausing season in an effort to determine if these C5's molt to adult stages prior to or after their annual upward migration coincident with the onset of the Southwest Monsoon.