Survival and Reproductive Success of Florida-Scrub Jays
Florida Scrub-jays are endemic to Florida and are classified as Threatened under the Endangered Species Act. These birds are considered cooperative breeders because the young delay dispersal after maturing to adults and help raise siblings. The jays are also known for their sentinel behavior where they take turns watching for predators and foraging in a pyrogenic landscape. The Florida Scrub-Jays living near Archbold Biological Station are banded and observed throughout their lives. The study began in 1969 and continues to this day. Additionally, there is an extensive pedigree available for this population. Learn more here.
A large part of my dissertation involved the long-term data set on a marked population of Florida Scrub-Jays, subject to study since 1970. Due to their social biology, there is an extensive pedigree of all families in the population in addition to extensive data on all aspects of life history. Analyses from these data will help us better understand the effects of genetics and common environment on survival and breeding success; important metrics underlying the growth of this rare and threatened species.
Statistical Methods and Visualization
I use a variety of modeling and statistical tools to estimate annual survival and reproductive success of Florida Scrub-Jays. I extract the relevant data from an Access database using SQL and I use R for most analyses, taking advantage of several different packages. I frequently use survival, ggplot2, car, and popbio among many others.
A picture is worth a thousand words, and good graphics can go a long way in communicating the meaning of results. Humans like aesthetically pleasing images. As such, data visualization is a research priority. I want to improve how we share and communicate results of complex analyses in ways that are understandable.
More broadly, I’m interested in predictive modeling, classification and visualization of data. Throughout my career as an ecologist and a quantitative biologist, I have learned several different methods of data analysis. I learned methods like regression and clustering in the context of biology but they can be applied to many applications outside the field.
I’m interested in package development and object-oriented programming. I am currently developing a package on data classification in collaboration with Dr. A Friedman of USF’s School of Information department. I have made use of knitr and R Markdown to generate reports that include code snippets, tables, graphs and text. I used these reports to communicate research results and further questions to my research group.
I’ve been using R since 2010, and currently learning Python and SAS to supplement my tool kit.
Stage-based Population Models
Individuals within populations (including humans!) vary in phenotype, sometimes in ways that affect fitness (survival + reproduction) effecting birth and death rates for the population as a whole. We developed a stage-based model to understand how the rate of growth to maturity affects population growth rate based on previous research of the life cycles of sessile, iteroparous species. We used analytical and numerical methods in Mathematica and R, respectively, to understand how changing survival, growth and reproduction affect population growth rate, generation time and reproductive output. This model is simple and serves as a starting point to understand the affect of changing parameter values on the end result, population metrics.