Exercise 3: Are bottlenose dolphin sightings distances to nearest lagoon related to ENSO indices in the San Diego, CA survey site?
1. Question that you asked
I was looking to see a pattern at more than one scale, specifically the relationship with ENSO and sighting distributions off of San Diego. I asked the question: do bottlenose dolphin sighting distributions change latitudinally with ENSO related to distance from the nearest lagoon. The greater San Diego area has six major lagoons that contribute the major estuarine habitat to the San Diego coastline and are all recognized as separate estuaries. All of these lagoons/estuaries sit at the mouths of broad river valleys along the 18 miles of coastline between Torrey Pines to the south and Oceanside to the north. The small boat survey transects cover this entire stretch with near-exact overlap from start to finish. These habitats are known to be highly dynamic, experience variable environmental conditions, and support a wide range of native vegetation and wildlife species.
Distribution of common bottlenose dolphin sightings in the San Diego study area along boat-based transects with the six major lagoons.
| FID | NAME |
| 0 | Buena Vista Lagoon |
| 1 | Agua Hedionda Lagoon |
| 2 | Batiquitos Lagoon |
| 3 | San Elijo Lagoon |
| 4 | Tijuana Estuary |
| 5 | Los Penasquitos Lagoon |
| 6 | San Dieguito Lagoon |
2. Name of the tool or approach that you used.
I utilized the “Near” tool in ArcMap 10.6 that calculated the distance from points to polygons and associated the point with FID of that nearest polygon. I also used R Studio for basic analysis, graphic displays, and ANOVA with Tukey HSD.
3. Brief description of steps you followed to complete the analysis.
- I researched the San Diego GIS database for the layer that would be most helpful and found the lagoon shapefile.
- Imported the shapefile into ArcMap where I already had my sightings, transect line, and 1000m buffered transect polygon.
- I used the “Near” tool in the Analysis toolbox, part of the of the “proximity toolset”. I chose the point to polygon option with my dolphin sightings as the point layer and the lagoon polygons as the polygon layer.
- I opened the attribute table for my dolphin sightings and there was now a NEAR_FID and NEAR_DIST which represented the identification (number) related to the nearest lagoon and the distance in kilometers to the nearest lagoon, respectively.
- I exported using the “conversion” tool to Excel and then imported into R studio for further analyses (ANOVA between the differences in sighting distances to lagoons and ENSO indices).
4. Brief description of results you obtained
After a quick histogram in ArcMap, it was visually clear that the distribution of points with nearest lagoons appeared clustered, skewed, or to have a binomial distribution, without considering ENSO. Then, after importing into R studio, I created a box plot of the distance to nearest lagoon compared to the ENSO index (-1, 0, or 1). I ran an ANOVA which returned a very small p-value of 2.55 e-9. Further analysis using a Tukey HSD found that the differences between ENSO states of neutral (0) and -1 and neutral and 1 were significant, but not between 1 and -1. These results are interesting because this means the sightings of dolphins differ most during neutral ENSO years. This could be that certain lagoons are preferred during extremes compared to the neutral years. Therefore, yes, there is a difference in dolphin sightings distances to lagoons during different ENSO phases, specifically the neutral years.
Histogram comparing the distance from the dolphin sighting to nearest lagoon in San Diego during the three major indices of El Niño Southern Oscillation (ENSO): -1, 0, and 1.
Violin plot showing the breakdown of distributions of dolphin sighting distances to lagoons (numbered 0-6) during the three different ENSO indices.
5. Critique of the method – what was useful, what was not?
This method was incredibly helpful and also was the easiest to apply once I got started, in comparison to my previous steps. It allowed to both visualize and quantify interesting results. I also learned some tricks for how to better graph my data and to symbolize my data in ArcMap.