Breaking down the clusters: exploring human influences on spatial patterns in invasive species data

Maps are powerful tools for understanding and tracking the spread of invasive species, with observations showing up as simple dots on a screen. But how do we interpret the story that these points tell us? Are observed patterns reflective of true ecological distributions or do human factors such as population, land protection status, and ease of access shape what we see on the map?

This semester, I had the great opportunity to work as an intern with the New York Natural Heritage Program, an organization dedicated to conserving New York's biodiversity through data collection, analysis, and research. I worked with the iMapInvasives data in ArcGIS Pro to examine potential spatial bias in invasive species observations. The main objective of the internship was to explore the geographic patterns present across NYS to better understand how human activity and accessibility may influence where observations are reported.

To analyze these patterns, invasive observation locations were compared with several datasets, including:

●        NYSDEC Trails (NYS GIS clearinghouse)

●        NYS Streets (NYS GIS Clearinghouse)

●        Population Density (U.S. census data)

●        New York Protected Areas Database (www.nypad.org)

Using GIS mapping and a simple overlay analysis of layers, I examined how invasive species observations are spatially related to human factors such as ease-of-access and protection status.

A particular example of this is the Adirondack region, where many of the observations are near or closely follow the major streets and trails. This thus raises the important question of whether these are true ecological hotspots or if it's simply because they are areas where people are more likely to travel or have access to and record observations.

Map of New York with invasive species presence and not-detected data (top left). The three map cutouts show data clusters in the Adirondack Park (top right), Willowbrook Park and Clove Lake Park in Staten Island (bottom right), and Marine Park in Brooklyn (bottom left).

I had also examined invasive species that overlapped with protected areas, using the New York Protected Areas Database (NYPAD), managed by the New York Natural Heritage Program (learn more at www.nypad.org). In multiple areas, the observations seem to be concentrated in places that are either within or near accessible and protected lands.

In fact, invasive species records disproportionately occur within NYPAD parcels. Over half of the presence and not-detected records in New York State occur within NYPAD parcels, despite these parcels taking up a minority of New York’s land area (focusing on directly owned NYPAD parcels, excluding lands protected via easement or proclamation).

This clustering of records within NYPAD parcels may partly reflect the increased recreational use, visitation, and monitoring or research efforts occurring within the protected lands.

Map of New York with invasive species presence data, not detected data and protected areas (top left). The zoomed in maps show data centered on Livingston/Ontario Counties (top right), Cattaraugus County (bottom left), and Jefferson County (bottom right).

However, when comparing the distribution of records to population density, the pattern was quite different. Observations were generally more frequent in moderately populated regions of NYS, and in less-populated regions surrounding high population areas, but fewer within the most densely populated areas. The pattern was variable in sparsely populated areas, with many records in some areas (often within NYPAD parcels) and few in others. While highly populated areas may provide more opportunities for reporting of invasive species, highly urbanized areas such as New York City contain highly developed land, limiting the amount of natural land. As a result of this, many of the observations in high-density areas were concentrated in green spaces such as public parks rather than being distributed throughout the urban landscape.

New York census data and invasive species presence and not detected data (bottom left). The zoomed in maps show the population centers of New York City (top left), Rochester (top right), and Buffalo (bottom right).

Throughout this project, I utilized ArcGIS Pro to organize, map, analyze, and compare multiple datasets in New York State. Conducting spatial analysis allowed me to examine the invasive species observations along with the trails, streets, population density layers, and protected land datasets. Rather than focusing on a single direct cause of spatial bias, this project’s aim was to primarily explore the visual patterns and spatial relationships within the data. By mapping these relationships, we can reveal how observation density may be influenced by protected land use, accessibility, and human population patterns. Understanding this information is important when interpreting the iMap data, because understanding spatial bias is essential to addressing it in future spatial analyses of the data.

This internship helped develop a simple framework for exploring potential bias patterns within invasive species datasets. By identifying how human activity and access may influence observation, this work will contribute to improving how invasive species distributions are mapped and interpreted.

Moving forward, future work could expand this analysis method to better help distinguish the human-driven sampling bias and true ecological patterns. Considering ecological and human factors together in a single model to quantify the relative influences of each layer on spatial clustering could improve our understanding. Ultimately, refining the methods and improving the analytical framework will support more accurate interpretation of invasive species distribution data, ensuring that future management and conservation decisions incorporate ecological data while recognizing human influences on those data.