This is the old Zanzibar Project web site.
Please visit the new one at: www.ZanzibarProject.org

X

Home | East African Coast mailing list | Progress | Gabriela Mayorga Adame

Zanzibar Channel Project: Progress

From April 17 to July 13, 2007, Gabriela Mayorga Adame, a coastal modeler of Theiss Research, was visiting the Institute of Marine Sciences (IMS) in Zanzibar to work with IMS researchers to develop modeling capabilities by setting up the coastal models ROMS and ADCIRC and conducting a series of studies. A general summary is given on the project's homepage. On this webpage, Mrs. Mayorga Adame posted her scientific report and three progress reports.

We would appreciate any comments or suggestions you might have. Please consider joining the East African Coast mailing list to which you could post your contribution and also use it as a forum for general discussions on coastal modeling in the East African region.

Scientific Report

Mayorga-Adame, C.G.: 2007, Ocean circulation of the Zanzibar Channel: A modeling approach. (pdf)

Third Progress Report: July 13, 2007 (end of 13th week)

During the last month of my three-month visit to IMS, we continued working on the local model and in parallel also tried to get the mesoscale model running. Please see my two previous progress reports below for more information about both models.

Progress on the mesoscale model had been slower than planned due to unexpected problems with installing Matlab under linux, which is needed to run ROMS_TOOLS, a suit of programs that extract the input data for ROMS from the global databases. Therefore the mesoscale model is not yet running. It will be a good challenge for the group after I will have left.

On June 26, we had a field trip to measure the temperature and salinity profile of a water column in a way the pioneers of oceanography might have made such measurements. We extracted water samples by diving to different depths using the very basic equipment shown in Fig. 1.

Fig. 1. Plastic bottles attached to a string were used to obtain water samples at various depths (left). The salinity was measured with a refractometer (right).

On the boat, we measured salinity with a refractometer. Bottom and surface temperature were measured with the thermometer included in the scuba gear. Two vertical profiles were measured, one to 10m depth and another one to 20m depth. In the 10m depth profile both salinity (35ppm) and temperature (26.8 degree Celcius) were constant along the profile or at least we did not register a variability. In the 20m profile there was a difference in temperature of 1 degree Celcius between the surface and the bottom and salinity was lower at the bottom (34ppm).

The temperature profile was added as an initial condition to the local model and we run it with SE wind conditions. The distribution of temperature agreed with the possibility of having upwelling condition along the coast of Zanzibar Island, that is, the warm surface water is carried towards the coast of the mainland (Fig. 2)

Fig. 2. Distribution of surface temperature (in degree Celcius) along the Zanzibar Channel with SE wind forcing.

Both the field trip and the model run was just to show how direct measurements can be included in a model and therefore to encourage measurements in the area. Obtaining monthly measurements at different locations in the Zanzibar Channel would be a good start to create a local database.

An analysis of the results of the global tide model TPX07, reveals a very interesting behavior of the tide in the Zanzibar Channel (Fig. 3), which agrees with the circulation pattern proposed by Shaghude et al. (2002). The maximum amplitude for the component M2 is observed in the center of the Zanzibar Channel close to the coast of Zanzibar Town and decreases towards both entrances and the coast of the mainland. According to TPX07 the principal components for the area (M2, S2 and N2) follow the same pattern. The tidal wave seems to be entering the Zanzibar Channel through both sides.

Fig 3. Distribution of the amplitude (m) and tidal ellipses of the M2 tide component in the Zanzibar Channel extracted from the TPX07 model solution.

We ran our local model with a tidal wave with an amplitude of 0.5m and the period of the M2 tide component, coming in through both entrances of the Zanzibar Channel.

The modelled circulation agrees with the circulation pattern described by Shaghude et al. (2002). Two current systems, one at the northern part of the Zanzibar Channel and the other one at the southern part are formed. These current systems converge during a flood and diverge during an ebb. Maxima in the velocity field (~10cm/s) are observed at the entrances of the channel and the minima in the central area, where the current systems meet (Fig 4).

Fig. 4. Circulation pattern observed during a flood (left) and an ebb (right) with the tidal wave coming in through both entrances of the channel (velocities are vertically integrated).

The group and I also started to learn the basics of ADCIRC, another coastal model. We began by learning how to generate the grids.

Acknowledgements

Working at IMS has been a great and challenging experience. I hope they continue on with this modeling effort.

I would like to say "Thank you" to all of those who answered my emails and helped me with the model. A special "Gracias" to Javier Zavala-Garay and "Asante Sana" to those who made my time in Zanzibar easier and enjoyable.

Bibliography

Shaghude, Y.W., K.O. Wannäs, and S.B. Mahongo, 2002: Biogenic Assemblage and Hydrodynamic Settings of the Tidally Dominated Reef Platform Sediments of the Zanzibar Channel. Western Indian Ocean J. Mar. Sci., 1, 107-116. (pdf)

Praise from Alfonse Dubi, Director of IMS (June 8, 2007)

Gabriela is wonderful and her group is well motivated. I am looking for possibilities for Gabriela to come back and attend the 5th WIOMSA Symposium in late October this year in Durban, South Africa. We want to set up a small demo and training programme for participants during the Symposium period.

Second Progress Report: June 8, 2007 (end of 8th week)

At the end of my 8th week here at IMS, my group and I gave a presentation to IMS staff on the current state of our project. The presentation is thus also a progress report and therefore posted here:

Presentation:

• as Powerpoint file, 1.5MB.
  
• as pdf file, 1.0MB.
  The animations have to be downloaded and viewed separately: sea_surface_elevation.avi, 0.6MB and velocity_vectors.avi, 1.3MB. They might have to be saved to a local disk first and then played.

Dr. Dubi suggested that for the remainder of my time at IMS I should continue developing the ROMS mesoscale model, which complements the ROMS local model explained in the presentation, set up ADCIRC (SMS), and teach my group more about linux.

The presentation

Praise from Ehrlich Desa at IOC (May 16, 2007)

I find that Gabriela is doing an excellent job based on the very detailed timetable drawn up [in her progress report below].

Thank you for keeping us updated on the progress in Zanzibar. It is really great that we are building on each others interventions and in alignment with the stated needs of IMS. Things look good to date.

First Progress Report: May 13, 2007 (end of 4th week)

After a very long trip from San Diego to Zanzibar four weeks ago, I have now settled in at IMS. I was given the Visiting Scientist Room where I work on the computer I brought with me from San Diego specifically for this project. Because Dr. Alfonse Dubi, the Director of IMS, would like to develop own modeling capabilities at his institute, he informed me during our first meeting that training local researchers should be the highest priority of my work here at IMS. A group of four researchers was therefore assigned to me to learn as much as possible about coastal modeling so that they would be able to continue working with the models ROMS and ADCIRC after I leave IMS in eight weeks. In the following, I would like to introduce the members of the group.

In order to increase the group's motivation for coastal modeling, I presented them the results of my previous work at the Universidad del Mar on the coast of Oaxaca, Mexico, where I developed a ROMS application of Huatulco Bays, a system of five bays. In turn, the group presented to me the results of a two-week coastal modeling workshop that was hosted by IMS a few weeks earlier and organized by IMS and the Intergovernmental Oceanographic Commission (IOC). It gave me a good understanding of what the group has already learned about coastal modeling. At this workshop, Dr. M.T. Babu of the National Institute of Oceanography in India introduced coastal modeling to IMS using the model MIKE21. Due to the lack of time, however, only an idealized configuration was implemented, using default values or approximations for most of the required variables.

As the group's members come from various backgrounds and have no prior experience with ocean modeling, we felt the need to start from the basics. Dr. Dubi thought that it would be a good idea if I first taught the group some of the basics of the software that we will be using. I therefore started to teach them some Matlab which we will be using as a pre- and post-processing tool to create the input files and analyze the results, respectively. Matlab will also be useful to process the local observational data. They are now familiar with creating small Matlab scripts using basic functions that are useful to read and plot data, and with the Matlab interface Seagrid which is used to generate the grid input file for ROMS. They also learned about the data bases that we are going to use to configure the mesoscale model, which are COADS, WOA, SeaWifs, Etopo2, Tpx06, NCEP, SODA and QuikSCAT. Now they have a better understanding of what is available in terms of global data, how these data were collected and whether they are useful for our particular mesoscale application.

During my current, fourth week here, the group is obtaining a deeper understanding of the basics of ocean modeling and ROMS. I am also teaching them some basic commands in linux, we are running a ROMS test case and are analyzing the results.

Since Dr. Dubi also wanted everyone else at IMS to have the chance to know about our project, I again presented the results of my previous work at the Universidad del Mar (see above) to a wider audience at IMS. In particular, I gave details on the local observational data that was used in my study. I hope served as a good illustration of the accuracy and resolution that would also be required from the local observational data of the Zanzibar Channel in order to develop a realistic model. I hope that through my presentation I will get to know further interested IMS researchers and to find out which data or useful information they might be able to offer for the local model of the Zanzibar Channel. For the next few weeks we have planned that I continue developing the mesoscale model and at the same time advise the group on developing the local model for which the emphasis will be on working with local observational data. Below is a timetable for the remaining time that I developed in close consultation with Dr. Dubi

May 7-11 (week 4)
Monday	Tuesday	Wednesday	Thursday	Friday
Another Matlab exercise: Plotting the local bathymetry data of the Zanzibar Channel	Introduction to linux and learning basic commands	Running a ROMS test case	Analyzing the output of the test case using some Matlab programs; Developing an oceanographic interpretation of the obtained results	Presentation of my previous work to IMS: "Ocean model of Huatulco Bays and data collection needed for it" (9:00am); Continuation of previous activity.
May 14-18 (week 5)
Monday	Tuesday	Wednesday	Thursday	Friday
Collecting all the useful observational data that may be available from the Zanzibar Channel: CTD casts, tides, current profiles, sediments and pollutant characteristics	Putting together and in the right format the available data for the local model	Analyzing the available data	Continuation of previous activity	Continuation of previous activity
May 21-25 (week 6)
Monday	Tuesday	Wednesday	Thursday	Friday
Generating the input files for the local model	Continuation of previous activity	Configuring ROMS for the local semi-idealized model	Continuation of previous activity	Doing some preliminary runs of the local model
May 28 - June 1 (week 7)
Monday	Tuesday	Wednesday	Thursday	Friday
Continuation of previous activity	Running the local model	Continuation of previous activity	Analyzing the model output	Continuation of previous activity
June 4-8 (week 8)
Monday	Tuesday	Wednesday	Thursday	Friday
Developing an oceanographic interpretation of the obtained results; Comparing it with publications and measured data if available	Continuation of previous activity	Continuation of previous activity	Writing up the results for a presentation to the IMS staff	Presentation
June 11-15 (week 9)
Monday	Tuesday	Wednesday	Thursday	Friday
Generating the input files for the mesoscale model	Analyzing the input files to check if the data obtained from the global data bases is consistent with the local data	Configuring ROMS for the mesoscale model	Continuation of previous activity	Conducting some preliminary runs of the mesoscale model
June 18-22 (week 10)
Monday	Tuesday	Wednesday	Thursday	Friday
Running the model	Analyzing the model output	Continuation of previous activity	Continuation of previous activity	Developing an oceanographic interpretation to the obtained results; Comparing it with publications and measured data if available
June 25-29 (week 11)
Monday	Tuesday	Wednesday	Thursday	Friday
Continuation of previous activity	Writing up the results for a presentation	Continuation of previous activity	Continuation of previous activity	Presentation
July 2-6 (week 12)
Monday	Tuesday	Wednesday	Thursday	Friday
Discussing possibilities to improve both the mesoscale and the local model	Setting further tasks for the team to continue to work on

April 15, 2007 (end of preparatory San Diego visit)

From April 2 to April 15, 2007, I was at Theiss Research in La Jolla, California, to install and configure the coastal models ROMS and ADCIRC on the computer that I will take to Zanzibar today. Together with Jurgen Theiss, I also talked to the following experts about the best approach to model the Zanzibar Channel, which is described below the list. We also talked about possible future extentions of my project and about my trip to Zanzibar in general.

Aurelien Ponte	Scripps Institution of Oceanography	meeting
Nancy Knowlton	Scripps Institution of Oceanography	meeting
M. T. Babu	National Institute of Oceanography, India	phone call
Emanuele Di Lorenzo*	Georgia Institute of Technology	phone call
Brian Arbic*	University of Texas at Austin	phone call
Chris Reason*	University of Cape Town, South Africa	phone call
Nicolas Le Dantec	Scripps Institution of Oceanography	meeting
Flavia Nunes	Scripps Institution of Oceanography	meeting
Art Miller	Scripps Institution of Oceanography	meeting
Glenn Ireley	Scripps Institution of Oceanography	meeting
Alexander Shchepetkin	UCLA	meeting
Francois Colas	UCLA	meeting
Clint Winant*	Scripps Institution of Oceanography	meeting
Joannes Berque	Intergovernmental Oceanographic Commission	phone call

* Advisory Committee member

Special thanks to Alexander Shchepetkin, who spent an entire day and evening with us configuring our computer and ROMS. We also want to thank in particular Chris Reason for offering to make the output of his Western Indian Ocean ROMS run available and Emanuele Di Lorenzo for offering processing time on his cluster.

Through the many discussions with the above experts we came to the conclusion that a mesoscale model that includes the Zanzibar Channel and that features a realistic seasonal variability will be best for a first approach. This is because it will show us the dominant processes influencing the dynamics of the Zanzibar Channel and provide us with boundary conditions for a second, local model that only contains the Zanzibar Channel. This approach has also the advantage that the required data are readily available on the internet from global datasets while the availability and accuracy of local data needed for a local model is not entirely clear yet. The need for good data for a local model is especially crucial because a local model is even more sensitive towards data than a mesoscale model.

The mesoscale model will have a resolution of 10 km, covering an area of about 1000 km^2, and the bathymetry will be provided by ETOPO2. It will be forced by winds and climatology from global data bases such as NCEP-QuickScat, COADS, SODA, and WOA.

Upon my arrival at IMS in Zanzibar, I would like to analyze together with the local researchers the available local data and then evaluate the possibilities of setting up the above mentioned second, local model. This smaller scale model could be set up at first as a semi-idealized model and as local data becomes available it can be included in the model to generate a more realistic scenario.

---