Focus on Wildlife -- Cleveland Metroparks

The Problem

Cities have become the primary habitat for humans on our planet. The year 2008 marked the first time in human history that the majority of people lived in urban places. Over 1.5 million people are added to our population each week, and the majority of this growth is absorbed into cities. Our great urban migration has two major consequences for natural communities. First, the incredible strain we place on natural resources is increasingly concentrated in and around urbanized centers. Second, for the majority of humans worldwide, the city has become the primary context for interacting with and experiencing nature and wildlife. Thus, our cities are placing unprecedented demands on natural resources while simultaneously serving as humankind’s primary access point to nature. Despite this accelerating pace of urbanization and its consequences for both natural systems and human perceptions of nature, the discipline that studies these issues – urban ecology – is a relatively new and undeveloped branch of science. As a result, nearly all ecological theory regarding wildlife communities has been established in wild, rural, or non-urbanized contexts. This means that we know very little about urban wildlife ecology and even less about the best practices for conservation and the human-nature connection in the context of cities.

The Solution

To address these challenges, we have formed a partnership between Cleveland Metroparks and Michigan State University to launch Focus on Wildlife, a large-scale, multidisciplinary, long-term urban ecology project in Cleveland, Ohio. Collaboratively, we seek to understand the mechanisms shaping the distribution and abundance of wildlife throughout the park system. We are testing some of the fundamental hypotheses of ecology in an extensive urban ecosystem to see if theories derived in the wild hold true in urban settings. These include the theory of island biogeography, which models the relationship between biodiversity and geographical area, and predator-prey theory, including the ‘landscape of fear’ concept. We are likewise investigating human-wildlife conflicts in an effort to enable humans and wildlife to thrive side-by-side in the city. We are also examining how human perception of the parks relates to their ecological characteristics so that we can improve our education and outreach strategies for connecting people with nature in the city. Collectively, these research efforts will advance the nascent field of urban ecology and ensure that urban parks are ecologically healthy refuges for both wildlife and people.

The Cleveland Metroparks is an astounding and unique urban ecosystem, perfectly suited to study wildlife community ecology and human-nature interactions in an urban context. The Metroparks are composed of 18 parks with over 23,000 acres of natural habitat, including rolling rivers, breath-taking waterfalls, and miles and miles of forest. Deer, coyote, fox, raccoon, and many other mammals call the Metroparks their home. The parks are also a haven for hundreds of bird species, host amazing plant diversity, and serve as a destination for millions of park visitors seeking to connect with nature. In close partnership with our Metroparks collaborators, we have deployed hundreds of wildlife cameras throughout this system in an effort to study this ecosystem without disturbing it. We have paired each of these cameras with a long-term vegetation monitoring plot, which will enable us to model human-wildlife-vegetation relationships with an unprecedented level of detail at multiple scales.

Phase 2

In the early stages of this project, we learned that the volume of images from this project can be overwhelming. Volunteers fatigued very quickly when they had to look through large numbers of images with no animals present or at large numbers of pictures of the same squirrel. We set about trying to find a way to use machine learning approaches (https://en.wikipedia.org/wiki/Machine_learning) to help eliminate the tedious parts of the workflow. We enlisted the help of Matt Sochor and Josiah Olson, computer programmers working for Progressive Insurance, who volunteered to help us build a machine learning model. They built a model that can quickly sort through images and eliminate the "nothing there" images while classifying species as well. We trained the model with a subset of the over 200,000 images classified by Zooniverse volunteers and undergraduate volunteers from the RECaP lab at Michigan State. In phase two of the project we will be putting up on this Zooniverse project images from 10 species that the model had difficulty classifying correctly in initial training rounds. This will help us produce a second training set so that we can further improve the model and allow us to focus volunteer efforts on different questions such as finding the rare species that the model does not know how to classify or asking whether dogs are on or off leash, etc.

Citizen Science and Underrepresented Groups

We are enthusiastic about involving both the public and groups traditionally underrepresented in the natural sciences in this project. Many of our sites in the Metroparks are maintained by dedicated volunteers and we are developing a Zooniverse citizen science project so that anyone with a computer and an interest in urban wildlife can take part in Focus on Wildlife. Conducting research in the city affords access to a diversity of students, citizens, and researchers. We are promoting diversity via various avenues, such as mentorship of high school interns through the Multicultural Apprenticeship Program at Michigan State University.