Performing spatial stats on turtle trawl data

Discussion created by boyntonj78 on Dec 6, 2011
Latest reply on Mar 6, 2012 by lscott-esristaff
I have a point dataset for sea turtles.  Each point represents a single trawl where 0-10 turtles were collected.  There are 4500+ points, most of which have 0 catch.  The study area encompasses the southeast.  We want to know if 1.) there is clustering of turtles and 2.) if there is clustering of the high catch numbers.  We want to know this for the overall area and by region (there are four smaller regions).  1.) Would Moran's I and Hot Spot Analysis be the best tools in the spatial stats toolset to answer these questions?

We tried aggregating the data into 1 x 1 min grids, but when we ran the 'Incremental Spatial Autocorrelation' the Z score never peaked and kept going up and up.  We ran it again on the
individual points, and got a peak.  2.) Is it appropriate to run either of these tools on incident data or would it be better to try to find a finer scale on which to aggregate?

Some of the Hot Spot Analysis information I read suggested that for datasets with 3000+ features, it would be advisable to construct a spatial weights matrix file.  3.) How is this better than choosing one of the other spatial relationships listed (e.g. fixed distance)?

Jessica B.