Are National Parks Islands? A Comparative Biodiversity Approach

Thumbnail Image



Journal Title

Journal ISSN

Volume Title

Repository Usage Stats



This project was a case study undertaken on behalf of the National Parks Conservation Association (NPCA). NPCA is a national non-profit organization that advocates for the protection and enhancement of the National Park System through issues like climate change, park funding, historical preservation, and wildlife conservation. With a more recent organizational shift to focusing on biodiversity conservation within National Parks, NPCA has become interested in ways to evaluate National Park units through the lens of biodiversity. This case study evaluates a single National Park, Yosemite, to see how well the biodiversity within the park represents the biodiversity of the larger surrounding area. In doing so, I develop a method for NPCA to use in future evaluations of other park units. Additionally, in this report I explore how well the park represents species of cultural importance to Indigenous tribes with ties to the Yosemite area.

Yosemite National Park is located in the southern stretch of the Sierra Nevada mountains of California, close to the Nevada border. To answer the question of how representative Yosemite’s biodiversity is of the surrounding area, I calculated an index of compositional dissimilarity, Bray-Curtis distance, for samples inside as compared to outside the Park (with “outside” defined by a 50-km buffer around the Park boundary). While the Bray-Curtis Index normally uses species occurrence data (counts of how many individuals of each species are present in each community), this approach proved untenable for this analysis because species occurrence data within Yosemite’s boundaries showed acute spatial bias, with observations clustered around the park’s most popular attractions and missing from the difficult to access backcountry regions of the park. Because of this, I turned to a dataset created by the California Department of Forestry and Fire Protections (CALFIRE). This dataset, which covers the entire state of California, maps 65 different Wildlife Habitat Relationships (WHR’s) at high thematic and spatial resolution. These categories range from single-species types (such as red fir) to more general habitat types (such as Sierran Mixed conifer forest or barren rock). This map allowed me to measure both habitat variety and abundance (calculated by the area covered by each WHR type) across the study site.

Because my study site spans an elevation range of almost 4,000 meters, it was necessary to subdivide the site to make ecologically relevant comparisons. Based on information provided by Yosemite National Park, five ecologically distinct “vegetation zones” exist within the park boundary. Upon further exploration, it became clear that the lowest elevation zone (500-900 m above sea level) was present in such small amounts within the park as to warrant exclusion from this analysis.

The study area was then randomly sampled with 2,000 cells measuring 1 km2 each (1,000 cells inside the park and 1,000 cells in the buffer area). The cells were then intersected with the California vegetation map, and the area of each WHR Type within each cell was calculated. The average area of each WHR Type was calculated for each elevation band within the park. The average composition (area per vegetation type) for each elevation band was then used to calculate the Bray-Curtis distance to each cell, both inside and outside the park. Boxplots were generated to summarize the results. Finally, the Bray-Curtis distances displayed as four maps, with each cell color coded by its compositional dissimilarity from the mean composition for each of the elevation bands.

The results of this analysis show that overall, Yosemite National Park and its surrounding area contain a large diversity of wildlife habitat types. The Park is representative of the surrounding area’s biodiversity, with two key exceptions: (1) the paucity of the lowest elevation zone (and its biodiversity) within park boundaries and (2) a key difference in the biodiversity present on the eastern side of the Sierra Nevada mountains. Because the mountain range cuts diagonally through the park, the biodiversity present on the eastern side (outside of the park boundary) was dissimilar from that of the western side (inside the park boundary), due to a rain shadow effect. This dissimilarity was particularly evident in the Upper Montane Forest and, to a lesser degree, the Subalpine vegetation zones.

I attempted to address the question of how well the park represents those species of cultural significance to Indigenous tribes, but I was unable to gain access to a key database curated by the Yosemite National Park, and therefore had to make do with publicly available information. This data, like other the occurrence data, was geographically clustered within the park, and I was therefore unable to make meaningful comparisons. In order to do evaluations of this kind in the future, it will be necessary to either have access to higher quality data, or collect that data in the field.

Throughout this case study I have documented every step of the data exploration and analysis. This includes writing scripts in R and Python and creating tools in ArcGIS Pro that allow the user to input their own data and perform similar analyses. While this study only covers one park, these tools will allow my client to evaluate other park units in a similar fashion, providing them with key insight into how well parks across the nation are representing their surrounding biodiversity.






Bukovnik, Isabel (2022). Are National Parks Islands? A Comparative Biodiversity Approach. Master's project, Duke University. Retrieved from

Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.