My objective was to see if the displacement of the birds showed particular patterns. For this, I decided to analyze the distribution of speed and rotation angles in space. Speed at a particular point is calculated as distance to previous point over time taken to move between points. Rotation angle refers to the angle between two consecutive movement lines (i.e., lines joining point A to B and B to C).
I first tried the Spatial Autocorrelation function, which indicated a clustered distribution of the values.
These results weren’t meaningful for me though, as I was interested in the variability within the observations. Studies on different animals species have shown that the analysis of variability within movement patterns can be used to infer behavioral patterns. I expected the birds would show varying speeds and rotation angles in response to the habitat where they were living (e.g., move slower inside the forest and quicker between forest patches; straighter movement lines in non-forest habitat). Thus, I decided to apply the Incremental Spatial Autocorrelation function, as this tool would indicate if the spatial clustering of values varied in the study area.
The results show mixed responses from each bird, with no clear interpretation for the observed patterns.
Most of them have non-significant z-scores, and those that do have no clear relationship to any environmental factor. Hot spot analyses don’t show a particular concentration of values at any point either.
In conclusion, speed and rotation angles are either A) not affected by the disposition of forest or B) bad indicators of behavioral changes associated to space use.
Very interesting that nothing showed up as significant! Do you think that prey distributions might be a better indicator of speed and rotation angles? I would seem that feeding events would be correlated with slower speeds and more diverse rotation angles but I might not know enough about hummingbird behavior to justify that. Could you map areas of preferred feeding habitat and correlate these parameters with that in mind? Maybe the entire forest is good feeding habitat and this would be irrelevant. Just an idea!
Thanks for your comment! Yes, I want to add information on food resource distribution to see if it correlates to any movement pattern. Hopefully that will help to clear things up!
Hi Noelia,
I’m going to disagree with your conclusions. It’s hard to tell from just your one screenshot of forest areas, but I’m going to assume that forest areas are non-uniform in size or shape. So strict clustering analysis like Moran I, which only takes into account distance values and ignores anything else, may not be appropriate. Birds in a large forest may behave similar to birds in a small forest, but that data will look spatially different.
For a question like “Do birds behave differently within and outside forests”, you can use basic analysis of variance without adding the complication of the spatial component.
This isn’t to say there aren’t other interesting spatial distribution questions you may ask and I’ll look forward to seeing how this dataset might work with some of the other tools in the Spatial Statistics toolset.
Hi Max,
thanks for your comment. I chose this analysis precisely because I am interested in the spatial component of the question. I’m not only interested in the general way they behave inside vs outside the forest, but on how the behavior changes while moving from one place to another. I would like to see if, as the birds move across the landscape, they change their behavior according to the type of habitat they are moving through. What I would expect to see, for example, is that, starting from a point in the forest, nearby points show similar speeds/turning angles but as the points become more distant and fall into the open space, the values will start to be different. Then, when the bird gets to the next forested area, the values should be similar again.
I’ll give a though to what you say about using basic analysis of variance though. I might be over-thinking this as you say.
Now, writing this I realize my data set might have not been appropriate for the analysis, as I don’t really have points falling on open areas (as birds moved so fast there it was very hard to catch up with them). The points I have are the end locations along a movement path, so even when the bird moved through an open space, the point is in the forest. Maybe I should transform the movement lines into lines of points with the corresponding speed values and run the analysis on that?
hi Noella,
I think that your results from the incremental Moran’s I are actually quite interesting. The first plot, of Moran’s I for speeds, as a function of distance among points, shows that points within about 50 m of one another were more likely to have similar speeds (Moran’s I > 0), whereas points separated by 100 m or so were more likely to involve a change in speed (Moran’s I < 0). This suggests to me that hummingbirds are changing speed in response to something (forest patches?) that have a scale of about 100 m. It would be interesting to create a histogram of forest patches and look for this.
The second plot, Moran's for angles, shows that points separated by about 10-20 m had different angles, (does this indicate rapid, abrupt turning, perhaps associated with foraging?) whereas points separated by 100 m or greater have more similar angles (perhaps indicating relatively straight flight paths over longer distances). To test whether my suggested interpretation is correct, you should create a plot of the turning angle as a function of distance (or time) for a single bird, and verify that over short distances the bird was turning abruptly but over longer distances it was traveling in more similar directions.
Keep up the great work! I look forward to your next post.
Julia
Hi Noelia,
You mentioned experimenting with Ripley’s K on your data. The results from Ripley’s K are really interesting in terms of interpreting data at varying scales. Last term, I had the students in GEO580 use Ripley’s K to look at distribution patterns of people’s hometowns within the United States. The results were very compelling, as we were able to “tease” out the clusters at varying scales.
I can’t seem to attach a screenshot of the Ripley’s K graph in this post, but I will try and send you one via email so that you can take a look at it. There are quite a few options in terms of permutations that you can run in order to narrow/widen the confidence interval. Your research is fascinating!
Doug