Monte Carlo (CIESM) talk - 24 Sept. 2001


The text below reflects the commentary of Prof. Allan R. Robinson for the slides of his Monte Carlo CIESM lecture on "The Coupling of Physics and Biology in Evolving Ocean Science". Some of the material represents spoken material, other material is self-instruction.


Intro

  1. title


  2. interdisciplinary ocean science - interdisciplinary ocean science today involves: fundamental problems of ocean science now are feasible and research is underway; coupled processes and interactions; multiple interactive scales. Stress theory, experimentation and simulation


  3. biological processes - The euphotic zone, where light is available for photosynthesis, is inhabited by a complex food web. Starting with new primary production fueled by using nutrients re-mineralized in the deep sea as well as regenerated production extending up to higher trophic levels and importantly exporting particles and carbon to the deep sea


  4. physical processes - Superimposed on the complex biological dynamics of the proceeding slide is physical forcing from a variety of processes and on a variety of scales. Advection transports and traps, and various fluxes occur from both the atmosphere and the deeper sea


  5. coastal processes - an aggregate of these processes characterizes any region of the sea. Those collected here include additionally convection and langmuir circulation


  6. bottom processes - when the region of the sea of interest is in the relatively shallow coastal zone, sediment processes become of utmost importance. There is dissolved and particular matter. Both interact with the sediment and the biological and physical dynamical environment


  7. scale eggs - I have introduced the concept of scales of space and time. Multiple space and time scale are another way of characterizing what is going on in a marine system. Here we have a range of time from one second to ten years and a range in space from one millimeter to 1000 kilometers. Notice that we have important overlaps of physical and biological ovals. For example, individual movement and turbulent patch size, and, plankton patchiness and fronts and eddies


  8. field estimation - having schematized a wide range of dynamics in the sea which occur over a vast range of scales, I now want to introduce another point of view - the concept that we must know the distribution of state variables, e.g. physical fields of temperature and velocity, the distribution and concentration of nutrients and plankton, and so forth. State variables as a function of time and space are called fields. Information on the evolution of fields is called field estimation. 2-D estimation of fields must be appropriate for the processes and scales of particular interest. It is a very demanding and complex task which today can only be adequately accomplished by data assimilation


  9. data assimilation - read bullets


  10. data assimilation schematic - the concept of data assimilation is schematized in this diagram, where arrows indicate flow of information and important feedbacks. I do not want to attempt to discuss the details of this diagram but do want to focus on: the process involves the estimation of state variables, the error of state variables, parameters and the error of the parameters; that some variables are observed and some are not but all are estimated; dynamical linkages spread information throughout. Feedbacks include model improvement and smart adaptation of sampling of desired quantities.


  11. Gulf of Cadiz - Real time field estimation or forecasts on interdisciplinary marine systems are now feasible and of great value. Here is a real time forecast carried out in conjunction with the NATO Undersea Research Centre (SACLANTCEN) in the Gulf of Cadiz for NATO naval operations in February and March 1998. Here is the Gulf of Cadiz, the Strait of Gibraltar and the Alboran Sea. SST and SSC (interpreted as chlorophyll) were assimilated into forecasts of temperature and chlorophyll and other variables. Error forecasts and other knowledge were used to optimize the sampling pattern of an airplane dropping AXBTs.


Coupled scales and Processes

  1. interactive coupled processes (divider) - having introduced the concept of coupled physical/biological dynamics in the sea involving multiple scales and interdisciplinary dynamical interpolation and the necessity of field estimation in context, now illustrate these ideas in the context of interactive submeso-/mesoscale/large scale coupled biological/physical dynamical processes


  2. SST - the first illustration will involve mesoscale eddies, the injection of nutrients into the euphotic zone and will also illustrate the context of observations, simulations and theory going hand-in-hand. Here is a typical satellite SST image of the western north Atlantic. I call your attention to the frontal edge of the Gulf Stream and Gulf Stream rings. This area of interest lies in the region of the Sargasso Sea and there are indications of mid-ocean eddies.


  3. CZCS - now look at this region but at satellite observations from a SeaWiFS predecessor. The image is of color pigments interpreted as chlorophyll.


  4. mixed layer T - turning from observations to simulations carried out which were tuned to all available synoptic mesoscale data for this region. Here we see the mixed layer temperature field. Notice the pair of warm and cold eddies which are entering, propagating westward and interacting


  5. nitrate flux - the simulation allows the estimate of fluxes of nutrients - here N - by injection process into the euphotic zone. The dark contours indicate the mixed layer temperature of the previous slide and the red and yellow the strong flux of N into the euphotic zone induced by eddy/eddy interactions occurring here.


  6. injection mechanism - the general process schematized here is from a multi-authored which appeared in Nature. The eddy injection mechanism was shown to account for the primary productivity of the Sargasso Sea, which was larger than be could accounted for by other estimates.


Advective theory

  1. intro - we will move to a theoretical study of how upwelling nutrients stimulate biological production. Read two bullets: general theoretical approach ...; analytical theory allows ...


  2. the model - the general model includes physical advection which occurs in a set of coupled equations for biological reactions. I don't expect to communicate all the details on this slide. A general theoretical solution is achievable via the theory of characteristics.


  3. biological dynamics in steady upwelling - a specific solution is of interest here for the following situation: initially there is no nutrient available in the euphotic zone. As upwelling of nutrients and seed plankton from the deep occurs, biological activity commences as the now active water pushes the lifeless water before it.


  4. dependence on gamma - I am now going to show profiles of nutrient (N), phytoplankton (P) and zooplankton (Z) in the biologically active region. Here we are looking at the situation after a long time the biologically active water fills the euphotic zone. If we focus only on the solid lines, there is a decrease of nutrients, a growth of P, which is then grazed down by Z. When mortality is considered, the Z themselves decay.


  5. Summary - characterization of these processes by non-dimensional parameters has been achieved. I would like to point out the nature of these parameters which is the ratio of biological rates and biological physical rates. As examples, alpha is the ratio of P uptake rate to N advection rate and beta is the ratio of Z grazing rate to uptake rate.


Jets and eddies

  1. barotropic jet - I next want to turn to the processes of meandering currents and jets, the formation of eddies and turbulence, and the primary productivity, population dynamics and ecosystem dynamics which occur in the meandering environment. The border between theory and simulation is fuzzy. Theoretical simulations study idealized processes and realistic simulations attempt to reproduce ocean processes. This slide shows the time evolution of an idealized barotropic jet. We are looking down on the eastward flowing jet in a north-south channel. The field displayed is vorticity. The jet instability grows into a vortex street and eventually secondary instabilities, turbulence, eddies and filaments fill the channel.


  2. copepod transport in jet - a cohort of copepod eggs is located near the south shore as the jet instability develops. The copepods are entrained into the vortex street and subsequently into the field of turbulence, eddies and filaments. A variety of patterns for the distribution of copepods would occur depending on the location of the release of eggs and the time at which they are released in the development of the jet instability.


  3. LAA assimilation comparison - we now move from an idealized simulation to a realistic simulation of the Gulf Stream front, which meanders and forms interactions with rings. Recall that earlier I pointed out the location of the Gulf Stream in the SST image of the western north Atlantic. This is a simulation which includes data assimilation. The top row depicts temperature for days seven and ten of the simulation. The second row is phytoplankton concentration, also for days seven and ten. In this simulation, real physical data has been assimilated with a compatible biological data, simulated offline in order to avoid assimilation shocks or false biological activity due to the assimilation procedure. For comparison, the bottom row also depicts phytoplankton for days seven and ten from a simulation in which only physical data alone was assimilated. Note the shape and strength of structures induced by the assimilation shock. Now we turn to population dynamics and ecosystem dynamics at the edge of the front.


Fronts

  1. euphausiid in frontal zone - You are looking at a vertical section of the edge of a front at a time of strengthening due to converging horizontal flow, which gives rise to a downward vertical flow along the front. Here we see a predator zooplankton and a prey micro-nekton. Biological behavior processes include swimming and taxis response. The net result of the physical and biological processes is an aggregation of the communities in the frontal region.


  2. swordfish and squid - Higher trophic level processes along the Gulf Stream front are illustrated here by predator swordfish and prey squid. Here the horizontal swimming of animals is important. The biological activity in interestingly limited to a 3D volume along the front defined by the minimum temperature preferred by swordfish and the maximum daytime light tolerated by the animal.


HOPS/LOOPS/AFMIS

  1. HOPS schematic - for the remainder of the talk I will present multi-scale interdisciplinary studies carried out by my Harvard group for regions of the western north Atlantic. I will illustrate the use of realistic field estimation in such research and the implications of that research for practical applications. An important focus is real time field estimation or ocean forecasting carried out with the Harvard Ocean Prediction System, which is comprised of a set of coupled dynamical model and data analysis and assimilation schemes.


  2. ESSE schematic - the advanced data assimilation scheme utilized by HOPS is called ESSE. This slide is included not to show al the technical detail but as an aesthetic slide for those who enjoy mathematics.


  3. nested domains - a system of multiple 2-way nested domains have been set up in the western north Atlantic off the east coast of US and Canada. Here is Nova Scotia, the Gulf of Maine, Georges Bank (with its rich fisheries), Cape Cod, Long Island and New York. Versions of these domains have and are being used for specific purposes. This is a prototype system for multi-purpose real time applications.


  4. AFMIS models - let me first talk about forecasting in the Gulf of Maine for fisheries research. We see here the HOPS schematic with the biogeochemical module used for this application. The physical and ecosystem models are coupled to a fish model. The fish stock abundance is based on an environmental preference of a particular species; for example, the temperature tolerance ranges of Cod on the bottom.


  5. web page cover - this is the index page of the web site which presents the real time forecasts of this system which was applied as a demonstration of concept in April/May 2000.


  6. RTDOC product - here I show one example of many of a set of set of forecast products. The fields shown are bottom temperature and bottom Cod abundance. As the temperature field evolves, the Cod move to follow the evolution of their preferred temperature.


  7. MBST-98 circulation - we next move to a Massachusetts Bay study carried out in real time in the early Fall of 1998. The first row of figures shows the variability of circulation in Massachusetts Bay. Read bullets. Below we see maps of temperature and zooplankton distribution. Read bullets.


  8. MBST-98 errors - physical forecasts with data assimilation were carried out in real time during the Massachusetts Bay experiment. Subsequently, biological hindcasts with data assimilation have been carried out successfully. A basic skill metric for forecasting is to provide a better estimate than persistence. Indicate persistence, forecast and forecast after melding with observations. The forecast was significantly better than persistence and good enough to use data for a very accurate melded estimate.


  9. ASCOT-01 objectives - I will end by presenting briefly work in progress. A real time forecast experiment was carried out in June 2001 called Assessment of Skill for Coastal Ocean Transients (ASCOT) in collaboration with the NATO Undersea Research Centre (SACLANTCEN) and the UMass. campuses in Dartmouth and Boston. Specific objectives were: read bullet 1 and bullet 2. Point to bullet 3. Oversampling is necessary to provide data sets for specific purposes.

  10. ASCOT-01 GOM forecast - a distributed real time ocean observing and predicting system was demonstrated during ASCOT-01 as physical forecasts were carried out at sea, the biological forecasts were carried out at Harvard and real time adaptively sampling for three ships was carried out aboard the NRV Alliance. Here are sample web site products from forecasts for the Gulf of Maine. Notice on the SeaWiFS image the strong biological activity on Georges Bank, in Massachusetts Bay, on Nantucket Shoals and as part of the coastal current. Many studies have noted similar generic processes occurring the Gulf of Maine and in the Mediterranean Sea.


  11. ASCOT-01 Mass. Bay forecast - finally I show coupled biological and physical forecasts for June 23-25 for Massachusetts Bay. These illustrations from the web page are of the temperature and chlorophyll fields. A wind-driven upwelling event has stimulated a bloom of activity which is being advected both along-shore and offshore. An additional episodic event can be seen intensifying in western Cape Cod Bay.


Conclusions

  1. bibliography-1 - coupled physical/biological dynamical processes have been presented here to provide a general overview based on a number of recent specific detailed studies which are simply listed here on three slides


  2. bibliography-2


  3. bibliography-3


  4. interdisciplinary ocean science - in conclusion, I remind you of the ideas and issues with which I introduced my presentation.


  5. Conclusions - in particular - bullet 1, bullet 2, bullet 3