Overview

Brian Schlining sent out this message on January 4th 2010:

Happy new year everyone,

With the start of the new year, I'm kicking off the work for the Video Annotation and Presentation project (VAAP) (see https://mww.mbari.org/resources/2010_Proposal_Process/phase_I_pdfs/901022_Video_Annotation_Analysis_and_Presentation.pdf)

As per management request, my first task is to figure out what VARS-derived data products are useful for your 'Biodiversity Initiative' work. Since the original idea of the VAAP project was to use techniques that we've already developed, here's the grocery list of work-that-has-been-done-before:

• Separating benthic annotations from midwater annotations
• Generating grids (e.g. maps) from annotations that are useful for spatial analysis
• Creating vertical distributions (i.e. both altitude and depth, either aggregated or as a timeseries)
• Normalizing annotation data base on ROV effort (very roughly normalized)

We can slice-and-dice the annotations in any way you see fit. Some examples are:

• by taxa
• by functional group (if you provide me with a list of organisms in the group you are interested in)
• by spatial bounds

As a first cut, it might be best to generate products that give 'the big picture'. In my mind the big picture is:

• what is the horizontal and vertical spatial distribution of the critters (and for Charlie, the rocks
  -Do these distributions change over time.

My naive suggestions for a first cut of data products would be:

1) Spatial grids of benthic annotations by taxa or functional groups.
a) An aggregate grid for each taxa/group spanning all annotations within Monterey Bay. This would provide an overview of where a taxa/group has been found
b) Grids for each taxa/group representing discrete time periods (for example, one grid for each year or grids for each oceanographic season). This would give some indication of change in the community structure.

2) Vertical distributions of taxa/groups
a) Aggregate vertical distributions (i.e. annotations from any time)
b) vertical distributions for discrete time periods (again, either yearly or by oceanographic season)
c) Same as a and b above but normalized by ROV effort

3) A time series product, perhaps vertical distribution vs time?

Please let me know if these products are useful for you. If you have other ideas please let me know. Again, these just represent a starting point; other data products can and will be added later as you determine what questions you're interested in examining for the Biodiversity Initiative.

No scientist responded to the email.

Product relations

Artifact Relationships

Tentative Products

Aggregated over entire data series from surface to 4000m with Monterey Bay

What spatial bounds do scientists want?

EXPD

1. 1m Histogram of EXPD data as text file
2. 1m Histogram of EXPD data as plot
3. 1m Histogram of EXPD normalized by total counts as text file
4. 1m Histogram of EXPD normalized by total counts as plot

VARS

The following will use a Concept and all it's children for an analysis. So for example, an analysis of Squid would be an aggregate analysis of all squid.

1. 1m Histogram of VARS data for each concept as text file
2. 1m Histogram of VARS data for each concept as plot

Normalized products

1. Normalized vertical distribution
   1. 1st normalize the VARS distribution by total counts
   2. Then normalize by dive effort (using normalized EXPD vertical distribution)