Click to edit Master text styles
Second level
Third level
Fourth level
Fifth level
Thank you, let’s finish things up
I think everyone is pretty aware that the scope and span of data collection designed and conducted during SO GLOBEC was pretty amazing. We’ve heard from several sources about all of these variables for collecting data, and I’d like to talk briefly about some of the work that will be done on the cetacean observations that Deb just mentioned
Until recently, few marine research cruises in the Southern Ocean have attempted to simultaneously collect data on both cetaceans and their prey with the objective of integrating these and other biological and physical data to investigate linkages at fine and meso scales.
 However, where this has been attempted recently, significant insights into ecosystem processes have been made.
SO-GLOBEC studies provide the ideal platform for such long-term studies, where scientists from a range of disciplines can conduct intensive focused studies, within the framework of international collaboration and long term synthesis of data and planning.
Current research efforts are directed at understanding how human activities (particularly climate change) and physical and biological processes in cetacean habitat and ecosystems affect the distribution, movements, social behaviour and health of individuals and populations.
These are the stated goals of the collaboration between the International Whaling Commission and SO GLOBEC With this directive in mind, I am interested in trying to use the data that were collected to define or characterize habitat for cetaceans in this part of the world. The GLOBEC data will support this effort at a range of spatial scales during fall and winter months. To do this, we will be using GIS and I’ll spend the rest of  the time briefly going over the methods that we will use and the questions that we are trying to answer
Some of the hypotheses I would like to test are relative to understanding how top predators interact with their environment. Several studies, including work done by Dr. Frazer, have shown that top predator populations in Antarctica respond differently to augmented climate change and variability
To date, there are very few data to address this question for cetaceans
It would stand to reason that animals that are larger, more mobile, and less dependent on an ephemeral prey source would not respond as quickly or dramatically as say penguins Eventually, we will have comparable data sets for all top predators in the region, and over time, we can elucidate how the interannual variation that we saw in 2 years is manifested in top predator distributions, fitness, and population growth
In order to make these connections we need to understand what determine the spatial patterns of the animals we are interested in We’ve heard a great deal about the physical environment and how it structures the biological environment and the spatial patterns of some species We will eventually analyze how much influence both the physical and biological environment have on the distribution of cetaceans in this part of the world at this time of year
In order to achieve this goal, we need the data.  On the surface, collecting the whale sightings data from direct observations may seem like a pretty simple and easy process, but it certainly has it’s difficulties just like all of the other more high-tech operations.
Direct observations that will define our spatial framework come from:
Static oceanographic data like the bathymetry
The bulk of the data were collected from the vessels
And in the case of the penguin and seal teams only thus far, observations were made using satellite tracking instruments.  Lucky dogs The synthetic data that we will be using will come as instantaneous surface data from ice images And most of the other data collected from the ships will be interpolated across the survey grid Then the are other data like distance data from shore or the shelf break or fronts etc
And finally, gradient data mostly from the bathymetry charts
So the cetacean data will be stacked on top of all of the other layers that were collected
And measured against a smattering of randomly placed points
To look for correlation
Some of these layers include distance from shore, depth, slope, and distance from other features
And here are what the sightings data from this past year look like overlayed with bathymetry I have to thank Tom Bolmer who isn’t here for working with me to get the bathymetry data into a format that was usable in GIS. And if there are other folks that are going to be using GIS, I’ll give these files to Bob Beardsley and I think he can make them available through one of the GLOBEC or WHOI web sites
And here are humpback whale sightings relative to the slope of the bottom in and around Marg. Bay
And finally distance from shore
And as we get the data processed, we will also have distance from shelf break
And distance from the ice edge from satellite images
When we look though at 2002, we see a very different pattern
Ice was forming much earlier in 2002 and there was fast ice present throughout the eastern and southern portion of marguerite bay.  Relative to humpback whales, there seems to be some pattern relative to this ice edge both across the mouth of the bay, and around adelaide island
Hydrographic data will be used to create layers to show gradients and fronts, which are well known to aggregate prey species
And then we will have the biological data from BIOMAPER to use as well
The first tests will describe the spatial structure of the sightings data.
And will tell us if there are positive- clumping associations, or negative associations between the cetacean sightings
Next we will look for relationships to environmental factors using Mantel’s tests
This is a useful tool for analysis of autocorrelated data
It will define the effects of space on animal distributions and spatial structures of environmental factors
And define whether an environmental factor has a significant relationship after accounting for the spatial autocorrelation
Finally, we will generate predictive habitat models using…
I am sure folks are aware of these stats, but as I am learning them, you’ll have to bear with me for a few seconds This cartoon shows how a Mantel’s test visualizes the correlation between space and these variables And which of these variables correlates with the distribution of minke whales off the east coast of korea in this case
As our spatial coverage of the survey area was not complete due to the time of year that the cruises took place, we’ll have to look at the conditions and levels of each variable at each sighting or patch of sightings.  We can then look for where these combinations occur throughout the study area to predict where we would likely see animals. And we can then test this predicted habitat against our observations to see how robust this assessment actually is
As we go down the road of these analyses, we have to realize that the models that we generate are only going to be as good as the data that we put into them
That the models we are creating are temporally explicit
The questions addressed with empirical models should be primarily targeted at building better hypotheses not creating predictive maps. With all of these caveats though, we are excited because we should be able to map out relationships we observe in our data and test them objectively
We will be able to define and test stronger hypotheses
We will be able to develop better sampling strategies and monitoring plans
And we will be able to show spatial interactions between cetaceans and their habitat, which while it sounds pretty basic, is really pretty substantial towards understanding the overall goals that I mentioned at the beginning.
And with that, I am done.
Here are humpback whale sightings from