Tag Archives: Geospatial ecology

Exercise 2: Possible Influence of ENSO Index on Dolphin Sighting Latitudes

Exercise 2

Question Asked: Are latitudinal differences in dolphin sightings in the San Diego, CA survey area related to El Niño Southern Oscillation (ENSO) index values on a monthly temporal scale?

  1. My previous question for Exercise 1 was: do the number of dolphin sightings in the San Diego, CA survey region differ latitudinally? I was finally able to answer this question with a histogram of sighting count by latitudinal difference. I defined latitudinal difference as the difference from the highest latitude of dolphin sightings (the Northernmost sighting point along the San Diego transect line) to the other sighting points, in decimal degrees. Therefore it becomes a simple mathematical subtraction in ArcMap. Smaller differences would be the result of a small difference and therefore mean more Northerly sighting, with large differences being from more Southerly areas. I used all sightings in the San Diego region (from 1981 through 2015). As you can see from below, there is an unequal distribution of sightings at different latitudes. Because I had visual confirmation of differences at least when all sightings are binned (in terms of all years from 1981-2015 treated the same), I looked for what process could be affecting these differences in latitude.

    Comparing the Latitudes with the frequency of dolphin sightings in San Diego, CA

ENSO is a large-scale climate phenomena where the climate modes periodically fluctuate (Sprogis et al. 2018). The climate variability produced by ENSO affects physical oceanic and coastal conditions that can both directly and indirectly influence ecological and biological processes. ENSO can alter food webs because climate changes may impact animal physiology, specifically metabolism. This creates further trophic impacts on predator-prey dynamics, often because of prey availability (Barber and Chavez 1983). During the surveys of bottlenose dolphins in California, multiple ENSO cycles have caused widespread changes in the California Current Ecosystem (CCE), such as the squid fishery collapse (Nezlin, Hamner, and Zeidberg 2002). With this knowledge, I wanted to see if the frequency of dolphin sightings in different latitudes of the most-consistently studied area was driven by ENSO.

Tool/Approach:

Primarily R Studio, some ArcMap 10.6 and Excel

Step by Step:

  1. 1.For this portion of the analysis, I exported my table of latitudinal differences within my attribute table for dolphin sightings from ArcMap 10.6. I saved this as a .csv and imported it into R Studio.
  2. Some of the sighting data needed to be changed because R didn’t recognize the dates as dates, rather as factors. This is important in order to join ENSO data by month and year.
  3. Meanwhile, I found NOAA data on a publicly-sourced website that had months as the columns and years as the rows for a matching ENSO index value of either: 1, 0, or -1 for each month/year combination. A value of 1 is a positive (warm) year, a value of 0 is a neutral year, and a value of -1 is a negative (cold) year. This is a broad-value, because indices range from 1 to -1. But, to simplify my question this was the most logical first step.
  4. I had to convert the NOAA data into two-column data with the date in one column by MM/YYYY and then the Index value in the other column. After multiple attempts in R studio, I hand-corrected them in Excel. Then, imported this data into R studio.
  5. I was then able to tell R to match the sighting date’s month and year to the ENSO data’s month and year, and assign the respective ENSO value. Then I assigned the ENSO values as factors.
  6. I created a boxplot to visualize if there were differences in distributions of latitudinal differences and ENSO index. (See figure)Illustrating the number of sightings grouped by ENSO index values (1, 0, and -1).
  7. Then I ran an ANOVA to see if there was a reportable, strong difference in sighting latitudinal difference and ENSO index value.

    Results:

     

    From the boxplot, it appears that in warm years (ENSO index level of “1”), the dolphins are sighted more frequently in lower latitudes, closer to Mexican waters when compared to the neutral (“0”) and cold years (“-1”). This result is intriguing because I would have expected dolphins to move northerly during warm months to maintain similar body temperatures in the same water temperatures. However, warm ENSO years could shift prey availability or nutrients southerly, which is why there are more sightings further south.  The result of the ANOVA, was a p-value of <2e-16, providing very strong evidence to reject the null of hypothesis of no difference. I followed up with a Tukey HSD and found that there is strong evidence for differences between both the 0 and -1, -1 and 1, and 1 and 0 values. Therefore, the different ENSO indices on a monthly scale are significantly contributing to the differences in sighting latitudes in the San Diego study area.

Tukey HSD output:

diff               lwr                        upr           p adj

0–1 0.01161047 0.004250827 0.01897011 0.0006422

1–1 0.04101170 0.030844193 0.05117920 0.0000000

1-0 02940123 0.020689737 0.03811272 0.0000000

 Critique of the Method(s):

These methods worked very well for visualization and finally solidifying that there was a difference on sighting latitude related to ENSO index value on a broad level. Data transformation and clean-up was challenging in R, and took much longer than I’d expected.

 

References:

Barber, Richard T., and Francisco P. Chavez. 1983. “Biological Consequences of El Niño.” Science 222 (4629): 1203–10.

Sprogis, Kate R., Fredrik Christiansen, Moritz Wandres, and Lars Bejder. 2018. “El Niño Southern Oscillation Influences the Abundance and Movements of a Marine Top Predator in Coastal Waters.” Global Change Biology 24 (3): 1085–96. https://doi.org/10.1111/gcb.13892.

Contact information: this post was written by Alexa Kownacki, Wildlife Science Ph.D. Student at Oregon State University. Twitter: @lexaKownacki

The Biogeography of Coastal Bottlenose Dolphins off of California, USA between 1981-2016

Background/Description:

Common bottlenose dolphins (Tursiops truncatus), hereafter referred to as bottlenose dolphins, are long-lived, marine mammals that inhabit the coastal and offshore waters of the California Current Ecosystem. Because of their geographical diversity, bottlenose dolphins are divided into many different species and subspecies (Hoelzel, Potter, and Best 1998). Bottlenose dolphins exist in two distinct ecotypes off the west coast of the United States: a coastal (inshore) ecotype and an offshore (island) ecotype. The coastal ecotype inhabits nearshore waters, generally less than 1 km from shore, between Ensenada, Baja California, Mexico and San Francisco, California, USA (Bearzi 2005; Defran and Weller 1999). Less is known about the range of the offshore ecotype , which is broadly defined as more than 2 km offshore off the entire west coast of the USA (Carretta et al. 2016). Current population abundance estimates are 453 coastal individuals and 1,924 offshore individuals (Carretta et al. 2017). The offshore and coastal bottlenose dolphins off of California are genetically distinct (Wells and Scott 1990).

Both ecotypes breed in summer and calve the following summer, which may be thermoregulatory adaptation (Hanson and Defran 1993). These dolphins are crepuscular feeders that predominantly hunt prey in the early morning and late afternoon (Hanson and Defran 1993), which correlates to the movement patterns of their fish prey. Out of 25 prey fish species, surf perches and croakers make up nearly 25% of coastal T. truncatus diet (Hanson and Defran 1993). These fish, unlike T. truncatus, are not federally protected, and neither are their habitats. Therefore, major threats to dolphins and their prey species include habitat degradation, overfishing, and harmful algal blooms (McCabe et al. 2010).

This project aims to better understand that distribution of coastal bottlenose dolphins in the waters off of California, specifically in relation to distance from shore, and how that distance has changed over time.

Data:

This part of the overarching project focuses on understanding the biogeography of coastal bottlenose dolphins. Later stages in the project will require the addition of offshore bottlenose sightings to compare population habitats.

Beginning in 1981, georeferenced sighting data of coastal bottlenose dolphin off the California, USA coast were collected by R.H. Defran and team. The data were provided in the datum, NAD 1983. Small boats less than 10 meters in length were used to collect the majority of the field data, including GPS points, photographs, and biopsy samples. These surveys followed similar tracklines with a specific start and end location, which will be used to calculate the sighting per unit effort. Over the next four decades, varying amounts of data were collected in six different regions (Fig. 1). Coastal T. truncatus sightings from 1981-2015 parallel much of the California land mass, concentrating in specific areas (Fig. 2). Many of the sightings are clustered nearby larger cities due to logistics of port locations. The greater number of coastal dolphin sightings is due to the bias in effort toward proximity to shore and longer study period. All samples were collected under a NOAA-NMFS permit.Additional data required will likely be sourced from publicly-available, long-term data collections, such as ERDDAP or MODIS.

Distance from shore will be calculated in a program such as ArcGIS or R package. These data will be used later in the project to compare to additional static, dynamic, and long-term environmental drivers. These factors will be tested as possible layers to add in mapping and finally estimating population distribution patterns of the dolphins.

Figure 1. Breakdown of coastal bottlenose dolphin sightings by decade. Image source: Alexa Kownacki.

 

 

 

 

 

 

 

 

 

 

 

Hypotheses:

I predict that the coastal bottlenose dolphins will be associated with different bathymetry patterns and appear clustered based on a depth profile via mechanisms such as prey distribution and abundance, nutrient plumes, and predator avoidance.

Approaches:

My objective is to first find a bathymetric layer that covers the coast of the entirety of California, USA to import into ArcMap 10.6. Then I need to interpolate the data to create a smooth surface. Then, I can add my dolphin sighting points and create a way to associate each point with a depth. These depth and point data would be exported to R for further analysis. Once I have extracted these data, I can run a KS-test to compare the shape of distribution based on two different factors, such as points from El Niño years versus La Niña years to see if there is a difference in average sighting depth or more common sighting depths based on the climatic patterns. I am also interested in using the spatial statistic analysis tool, Moran’s I, to see if the sightings are clustered. If so, I would run a cluster analysis to see if the sightings are clustered by depth. If not, then maybe there are other drivers that I can test, such as distance from shore, upwelling index values, or sea surface temperature. Additionally, these patterns would be analyzed over different time scales, such as monthly, seasonally, or decadally.

Expected Outcome:

Ideally, I would produce multiple maps from ArcGIS representing different spatial scales at defined increments, such as by month (all Januaries, all Februaries, etc.), by year or binned time increment (i.e. 1981-1989, 1990-1999), and also potentially grouping based on El Niño or La Niña year. Different symbologies would represent coastal dolphin sightings distances from shore. The maps would visually display seafloor depths in a color spectrum by 10 meter difference. Because the coastlines of California vary in terms of depth profiles, I would expect there to be clusters of sightings at different distances from shore, but similar depth profiles if my hypothesis is true. Also, data with the quantified values of seafloor depth would be associated with each data point (dolphin sighting) for further analysis in R.

Significance:

This project draws upon decades of rich spatiotemporal and biological information of two neighboring long-lived cetacean populations that inhabit contrasting coastal and offshore waters of the California Bight. The coastal ecotype has a strong, positive relationship with distance to shore, in that it is usually sighted within five kilometers, and therefore is in frequent contact with human-related activities. However, patterns of distances to shore over decades, related to habitat type and possibly linked to prey species distribution, or long-term environmental drivers, is largely unknown. By better understanding the distribution and biogeography of these marine mammals, managers can better mitigate the potential effects of humans on the dolphins and see where and when animals may be at higher risk of disturbance.

Preparation:

I have a moderate amount of experience in ArcMap from past coursework (GEOG 560 and 561), as well as practical applications and map-making. I have very little experience in Modelbuilder and Python-based GIS programming. I am becoming more familiar with the R program after two statistics courses and analyzing some of my own preliminary data. I am experienced in image processing in ACDSee, PhotoShop, ImageJ, and other analyses mainly from marine vertebrate data through NOAA Fisheries.

Literature Cited:

Bearzi, Maddalena. 2005. “Aspects of the Ecology and Behaviour of Bottlenose Dolphins (Tursiops Truncatus) in Santa Monica Bay, California.” Journal of Cetacean Research Managemente 7 (1): 75–83. https://doi.org/10.1118/1.4820976.

Carretta, James V., Kerri Danil, Susan J. Chivers, David W. Weller, David S. Janiger, Michelle Berman-Kowalewski, Keith M. Hernandez, et al. 2016. “Recovery Rates of Bottlenose Dolphin (Tursiops Truncatus) Carcasses Estimated from Stranding and Survival Rate Data.” Marine Mammal Science 32 (1): 349–62. https://doi.org/10.1111/mms.12264.

Carretta, James V, Karin A Forney, Erin M Oleson, David W Weller, Aimee R Lang, Jason Baker, Marcia M Muto, et al. 2017. “U.S. Pacific Marine Mammal Stock Assessments: 2016.” NOAA Technical Memorandum NMFS, no. June. https://doi.org/10.7289/V5/TM-SWFSC-5.

Defran, R. H., and David W Weller. 1999. “Occurrence , Distribution , Site Fidelity , and School Size of Bottlenose Dolphins ( Tursiops T R U N C a T U S ) Off San Diego , California.” Marine Mammal Science 15 (April): 366–80.

Hanson, Mark T, and R.H. Defran. 1993. “The Behavior and Feeding Ecology of the Pacific Coast Bottlenose Dolphin, Tursiops Truncatus.” Aquatic Mammals 19 (3): 127–42.

Hoelzel, A. R., C. W. Potter, and P. B. Best. 1998. “Genetic Differentiation between Parapatric ‘nearshore’ and ‘Offshore’ Populations of the Bottlenose Dolphin.” Proceedings of the Royal Society B: Biological Sciences 265 (1402): 1177–83. https://doi.org/10.1098/rspb.1998.0416.

McCabe, Elizabeth J.Berens, Damon P. Gannon, Nélio B. Barros, and Randall S. Wells. 2010. “Prey Selection by Resident Common Bottlenose Dolphins (Tursiops Truncatus) in Sarasota Bay, Florida.” Marine Biology 157 (5): 931–42. https://doi.org/10.1007/s00227-009-1371-2.

Wells, Randall S., and Michael D. Scott. 1990. “Estimating Bottlenose Dolphin Population Parameters From Individual Identification and Capture-Release Techniques.” Report International Whaling Commission, no. 12.

——-

Contact information: this post was written by Alexa Kownacki, Wildlife Science Ph.D. Student at Oregon State University. Twitter: @lexaKownacki