Background motivation:

Interest in the controls on biological patchiness has motivated many of the investigations of plankton populations since the late 1800's.  The debate over whether plankton populations are essentially uniform over large areas, making patchiness "noise" in the distribution, or essentially    patchy, making it critical to understand the scales over which patchiness is relevant, hearkens back to an intense debate started over 100 years ago.  Today the advent of remote sensing-- moorings, tow bodies, ROV/AUV platforms, satellites, radar/sonar-- gives us the data to tackle the patchiness question in more detail than before. The motivation for renewed interest stems from efforts to make complex biophysical models of the ocean predictive, since these models must be built to adequately  characterize the relevant controls on plankton populations.  In addition, the fields that are used to initialize the model can heavily influence the results, as seen in the case of 100% fish larvae mortality in uniform fields with the "average" food supply.

Abstract:

Data are needed over a wide range of scales to assess and characterize biological patterns in the ocean.  Most of the data that provide information over a sufficient range of scales to characterize the patchiness of biological fields relevant to large scale ocean simulations are obtained from in situ platforms making proxy measurements (e.g. sonar, fluorescence).  The volume of these data, and the lack of continuous calibration in most cases, presents challenges for traditional methods of data analysis.  Wavelet analysis provides one approach to extracting pattern information that is localized in both signal space (or time) and in scale space (or frequency).  It has shown particular utility when applied to high frequency sonar data.  Sonar data is a highly frequency dependent convolution of physical and biological scattering sources--sound scatters off of density discontinuities, which include turbulence generated microstructure in T-S fields as well as plankton patches. Using internal waves as an example, pattern-based analysis of sonar data shows that distributions of plankton can either follow the distributions of physical fields, or diverge sharply depending on the circumstances. Moreover, patterns are often well characterized by fractal or multifractal distributions, which in turn allow straight-forward generation of realistic fields appropriate for model initialization that are true to the observations over the resolved scales. This talk aims to demonstrate that:
1) patchiness varies in quantitatively detectable ways in different hydrographic regimes, plankton communities, and seasons; 2) signatures of 2-D and 3-D turbulence can be identified in acoustic records; 3) wavelets are a tractable way to go about this, once you find the right references; and 4) fractals provide a useful framework for characterizing stochastic systems.

Light refreshments are served in the Interaction Area (4th floor of the Oceanography/Physics Building) at 4:00 p.m.

All seminars begin at 3:00 p.m. and are held in room 200 of the Oceanography/Physics building.