Ocean Tracer Transport Model (OTTM)

Pathways of Indonesian Throughflow
at surface traced using dye in OTTM


25 years of air-sea CO2 fluxes derived
using OTTM 4D-var carbon assimilation


IITM label rotated in OTTM using sphere rotation velocities.



What is OTTM?

OTTM stands for Ocean Tracer Transport Model. It may be also called as Offline Tracer Transport Model, but originally it was named as shown in the first form.


History of OTTM

OTTM is developed to conduct researches on ocean carbon cycle. It was originally developed in the modelling group of GOSAT-project operating at National Institute for Environmental Studies (NIES), Tsukuba, Japan. The first version of the model was developed and tested in 2006 and published in 2008. For any questions and comments on the model and development please contact valsala@tropmet.res.in


What does OTTM do?

OTTM calculates the evolution of a prognostic passive tracer in the ocean by solving the advection-diffusion-mixing-source-sink equation of any scalar tracer. OTTM operates on pre-determined currents and other physical parameters of the ocean as well as the atmosphere. Therefore, OTTM is an “offline” model, meaning that it requires a 4-dimensional data of ocean currents, temperature, salinity and other physical parameters to drive the model. OTTM solves only the scalar tracer equation. The following is the general form of the scalar tracer equation solved in OTTM.

In the above equation, ‘C’ represents the scalar tracer (or passive tracer). The ‘U’ represents the 4-dimensional ocean velocities, the ‘nabla’ represents the horizontal gradient operator, ‘Kz’ represents the vertical mixing coefficients, ‘Kh’ represents the horizontal diffusion coefficients, ‘Φ’ represents the tracer source or sink term. In addition to the above general form of the equation, OTTM specifically includes ‘other’ processes of tracer evolution in the ocean such as ‘isopycnal diffusion’ and ‘eddy-induced transports’. These two terms are calculated based on the following general form.

The above term is added as ‘bolus’ velocities into the first equation in order to account for the eddy induced transport. In a verbal format the OTTM solves totally the following processes,

Tracer evolution = advection + convection + vertical mixing + horizontal diffusion + eddy induced transport + isopycnal diffusion + source + sink.


What is the difference between OTTM and conventional Ocean General Circulation Model (OGCM)?

OTTM does not solve momentum (i.e. velocities) and active tracer (i.e. temperature, salinity and density) equations like in OGCM. Instead, OTTM ‘borrows’ these essential parameters from a ‘foreign’ model output or from an ‘ocean re-analysis’ data and utilize them to evolve a passive tracer or scalar in the ocean. Such models which operate on pre-determined circulations are generally called ‘offline’ models.


What are the variables required externally and what are diagnosed internally in OTTM?

OTTM require zonal velocity(u), meridional velocity (v), temperature (t), salinity (s) as 4-dimensional variables. Most of the foreign model outputs are available only in the above four parameters. Essential parameters such as seasonal vertical mixing coefficients are seldom available from the parent model. Therefore OTTM is designed to calculate its own mixing coefficients with the additional information of surface forcing parameters. They are zonal wind stress (Tx), meridional wind stress (Ty), surface shortwave flux, net heat flux and precipitation-evaporation fluxes. With the help of these OTTM calculate the vertical mixing coefficient based on KPP-parameterization. Vertical velocity is calculated within OTTM by respecting the principles of continuity equation.


How is OTTM coded?

OTTM is coded in simple Fortran-77 language with essential parallel environment. The model runs only in a parallel or cluster computers. The geometry of the model is in spherical co-ordinates with poles located exactly over the geometric poles the earth. It is a finite volume model in which the co-ordinates are in x,y,z directions. The Arakawa-B grid is used for horizontal staggering of grids. The vertical coordinates are in z-levels.


What are the applications of OTTM?

OTTM can be used to find 4 dimensional evolutions of any passive tracers (scalars) in the ocean. The tracers can be ‘artificial dye’, age tracer or any other meaningful biogeochemical elements. Therefore OTTM can be readily used as a biogeochemical model.


What are the environments required to run/use OTTM?

a.   Technically, a parallel computing environment or cluster of computers is required to run OTTM. It cannot be run in single processor. It requires multi-processor facility.

b.   An input offline data of ocean velocities, temperature, salinity and other essential parameters are required. Usually these can be monthly data from a foreign model output or from an ocean re-analysis data. OTTM interpolate the offline inputs into model time steps to find evolution of tracers.

c.   Input data are preferably from a model output which is also run under staggered B-grid configuration. Otherwise, OTTM calculates the data in B-grid, which may conserve mass in different volume cells than in the parent input data.

d.   All the input data should be converted to a model friendly format. OTTM accepts binary sequential access unformatted files created in Fortran. Outputs will be available in NetCDF format.


What are the essential tests for the OTTM?

OTTM can be tested for its spherical geometry based on a simple test of ‘sphere rotation’. OTTM is essentially coded in spherical co-ordinate with a constant radius of earth. Therefore, a test velocity can be generated, as if, a sphere is rotating with respect to an axle passing through its centre. For such rotations, each point on the surface of the sphere will have a specific velocity vector which can be written down in simple spherical co-ordinate trigonometric equations of sin and cos functions. After resolving u, v (i.e. zonal and meridional velocities), they are used to run the OTTM for a test case. In this sphere rotation case the horizontal velocities are non-convergent. In this 2-dimensional ‘sphere rotation’ experiment, any tracer initialized in OTTM should rotate as if it is a ‘sticker’ on a rotating solid sphere. This property is a powerful tool to evaluate the geometry of the model and its fidelity and estimate the numerical errors and other statistics.

Following figure shows the rotation of an ideal tracer initialized and run with a sphere rotation velocity fields. The maximum speed of rotation is 1 m/second.

Figure 1: Rotation of a sphere simulated in OTTM. The initial tracer is placed at the equator and allowed to evolve under velocities of surface of a rotating sphere with maximum rotation speed of 1m/sec. View from both sides of the sphere are shown separately.



How do we validate the results of OTTM in a real application case?

OTTM or any transport model needs a thorough validation of results. OTTM is validated usually by simulating the Chlorofluorocarbon (CFC) in the ocean. The CFC has been emitted to the atmosphere since 1930s. The atmospheric record of CFC is available from 1930s. Initially, the ocean had zero CFC concentrations. Since 1930s, the atmospheric CFC has been dissolving into the ocean, and CFC concentrations in the ocean started to develop. Since 1930 the CFC concentration in the ocean is increasing continuously, until recently due to the industrial emissions of CFC has been regulated.


OTTM results can be verified by CFC simulations. For this purpose, the OTTM is used to simulate the oceanic CFC from 1930 and continue the simulation until 1995 or 2000. A global survey of CFC during 1990s is available as observational data. The simulated CFC concentration of model year 1995 can be compared with the global mean observations. For an accurate transport model the model CFC should match with the observations. The model error can be estimated in this way.



What are the works so far done using OTTM?

OTTM has been used, so far, for the following purposes.

a.   CFC-11 simulation, validation and interannual variability of CFC-11 sinks

In Valsala et al., (2008) we validated the OTTM using CFC-11 simulations. Oceanic concentrations of CFC-11 were simulated from 1938 to 2000 using offline input data from GFDL re-analyses. This work reported the design and validation of OTTM by simulation of CFC-11.

In a following study Valsala and Maksyutov, (2010a) have used OTTM to investigate the interannual variability of CFC-11 sink in the ocean, and the long term trend and capacity of the ocean to sequestrate the atmospheric CFC. The major finding was that the CFC sink variability of the ocean is tied up to the climate variability such as El Nino and other major climate anomalies. On the whole, the ocean’s capacity of sequestrating CFC is reducing because of saturation of upper thermocline and ventilation pathways with the CFC which reject the further uptake. A slow transport of the ocean thermocline to subsurface of the ocean also causes a saturation of CFC11 in the thermocline and slowdown of continued uptake of CFC-11 by the ocean.


b.   Carbon cycle simulation and air-sea CO2 flux assimilation.


In Valsala and Maksytov (2010b), the OTTM was used to simulate the carbon cycle in the ocean based up on a simple biogeochemical model. The target was to find out the contemporary CO2 fluxes between the atmosphere and the ocean. We used the OCMIP-II type simple abiotic model for the carbon chemistry and a phosphate dependent one-component ecosystem model for the biology. In addition to the simulation of CO2 fluxes, we assimilated the surface ocean pCO2 of the model with the ship-based observations. We synthesized a 9 year of air-sea CO2 flux data from 1996 to 2004 in this study. The OTTM is used for assimilation in a 4D-variational approach. The 'adjoint' of the forward model was simply estimated with reversed currents of the offline input data and the integrations were carried out from the end point to the starting point. The assimilated air—sea CO2 flux data from 1996 to 2004 is available at http://cdiac.ornl.gov/oceans/CO2_Flux_1996_2004.html and a longer record of 1980 to 2010 at http://apdrc.soest.hawaii.edu/datadoc/co2_flux.php



c.   Pathways of water masses in the ocean and its variability.


OTTM can be used to simulate ideal dye tracer in response to a source kept at any point in the ocean. This property is used to trace the Indonesian Throughflow (ITF) in the Indian Ocean from 1950 to 2000. The target was to find out the interannual variability of pathways of the ITF in the Indian Ocean. In this case two set of simulations were done using OTTM. In the control case, an ideal tracer from ITF origin region is traced for 50 years using monthly circulations of offline input data. In a second run, the OTTM is simulated for the same source of tracer but with climatological monthly mean circulations repeatedly for 50 years. The difference between the control run and the climatological run gives the oscillating patterns of tracers in the ocean which comes from the interannual variability in the circulation. An EOF method is employed to finding ‘major dancing pattern’ of the dye tracers. They are interpreted as the interannual variability in the pathways of the ITF. The results were published in Valsala et al., (2010c).


In a complementary study, OTTM was used in ‘adjoint’ mode to detect the origin and pathway variability of ITF in the Pacific Ocean. The adjoint mode means the backward simulation of tracers with a reversed currents (input current data is multiplied by negative 1 to reverse the direction of flow) and integrate the model from end point (year-2000) to starting point (year-1950). This gives the origin of ITF in the Pacific Ocean. The above sets of experiment were done in the ‘adjoint’ OTTM model and pathway variability of origin of ITF in the Pacific Ocean was found. Results were published in Valsala et al., (2011).




What are the potential (future) usages of OTTM?


1.   Dispersion of dye in the ocean. This can be used to find ‘pollution’ transport in the ocean such as oil-spillage.

2.   Dispersion of biological elements such as Larvae, microbial and fish-eggs.

3.   Identifying a suitable location in the ocean for waste material dumping from where they never or least surfaces after a disposal. This can be used in artificial carbon sequestration projects.

4.   Biogeochemistry and Carbon cycle simulations of global oceans.

5.   Adjoints and variational data assimilation.


Where to contact for obtaining OTTM code or related data for research use?

Contact valsala@tropmet.res.in for obtaining model code and any related data for research usage for free of cost.


OTTM related publications


1.   Valsala V., S. Maksyutov and M. Ikeda, (2008): Design and validation of an offline Oceanic Tracer Transport Model for Carbon Cycle Study, J. Climate , Vol. 21, 2752-2769.



2.   Valsala V., (2009): Different spreading of Somali and Arabian coastal upwelled waters in the northern Indian Ocean: A case study, J. Oceanography, Vol. 65, 803-816.



3.   Valsala V., H. M. Alsibai, M. Ikeda and S. Maksyutov, (2010a): Interannual variability of CFC-11 absorption by the ocean: an offline model study, Climate Dynamics, doi:10.1007/s00382-010-0784-4.

4.   Valsala V. and S. Maksyutov, (2010b): Simulation and assimilation of global ocean pCO2 and airsea CO2 fluxes using ship observations of surface ocean pCO2 in a simplified biogeochemical offline model, Tellus-B, Vol. 62B, 821-840.

5.   Valsala V., S. Maksyutov and R. Murtugudde, (2010c): Possible Interannual to Interdecadal variabilities of the Indonesian Throughflow (ITF) water pathways in the Indian Ocean, J. Geophysical Research, Vol. 115, C10016, doi:10.1029/2009JC005735.

6.   Valsala V., and S. Maksyutov, (2010d): A short surface pathway of the subsurface Indonesian Throughflow Water from the Java coast associated with Upwelling, Ekman transport and subduction, Int. Journal of Oceanography, vol. 2010, doi:10.1155/2010/540783.



7.   Valsala, V. S. Maksyutov, R. Murtugudde, (2011): Interannual to Interdecadal Variabilities of the Indonesian Throughflow Source Water Pathways in the Pacific Ocean. J. Phys. Oceanography., 41, 1921–1940.



8.       Valsala, V., S. Maksyutov, (2013): Interannual variability of the air–sea CO2 flux in the north Indian Ocean, Ocean Dynamics,  doi:10.1007/s10236-012-0588-7, 1-14

9.   Valsala, V., S. Rahul, A. Lenton and R. murtugudde, (2013), Interannual to interdecadal variability of south Indian Ocean MOC inferred from CFC-11 inventories, J. Geophysical Research, (under review)