Invasive Plant Spread Models for Future Climate

ISEF Category: Plant Sciences

Subcategory: Ecology  ·  Difficulty: Intermediate  ·  Setup: Home Setup  ·  Time: 1 to 2 Months

The Hook

A plant can look harmless today and become a serious problem a few decades from now. Climate change can open new territory for invasive species, like moving the goalposts after the game starts. With the right data, you can model where that plant may spread next. That gives your project a real conservation edge.

What Is It?

This project uses a species-distribution model, which is a way to predict where a plant could live based on the climate where it already appears. Think of it like a filter. The model asks, “What conditions keep this plant showing up, and where else do those same conditions exist?” MaxEnt is a common tool for this. It looks for the most likely habitat pattern from presence records, which are location points where the species has been observed.

You combine two big data sets. GBIF gives you occurrence records, and WorldClim gives you climate layers such as temperature and rainfall. Then you compare the plant’s current range with future climate scenarios. The result is a map that shows where suitable habitat may expand, shrink, or shift. You are not proving exactly where every plant will grow. You are testing a climate-based prediction with real ecological data.

Why This Is a Good Topic

This is a strong science fair topic because you can ask a clear question, use public data, and produce a map that changes when you change the inputs. You do not need a wet lab. You can study a real environmental problem, invasive spread, and learn how scientists turn raw records into predictions. You also get to practice data cleaning, GIS thinking, and model interpretation, which are useful skills in ecology and environmental science.

Research Questions

• How does the choice of climate variables change the predicted future range of the invasive plant?
• What is the effect of using different background regions on MaxEnt model output?
• Does adding elevation improve the model's ability to predict known occurrence points?
• To what extent do future climate scenarios shift the center of suitable habitat?
• Which climate layer contributes most to the predicted spread of the invasive plant?
• How does spatial filtering of GBIF occurrence records affect model accuracy?
• What is the effect of using one invasive species versus a closely related native species as a comparison?

Basic Materials

• Laptop or desktop computer with internet access.
• Free GBIF occurrence data for the chosen plant species.
• Free WorldClim climate layers.
• MaxEnt software or another accessible species-distribution modeling tool.
• QGIS for mapping and raster viewing.
• Spreadsheet software such as Google Sheets or Excel.
• Digital notebook for tracking data cleaning decisions and model settings.

Advanced Materials

• Laptop or desktop computer with strong memory and storage.
• R or Python for batch processing and model comparisons.
• QGIS for map production and spatial checks.
• Occurrence data from GBIF, cleaned for duplicates and spatial bias.
• WorldClim current and future climate layers.
• Elevation and land cover layers for covariate testing.
• MaxEnt software or comparable ecological niche modeling package.
• Statistical tools for cross-validation and model evaluation.

Software & Tools

• GBIF: Downloads species occurrence records for your target plant and nearby comparison species.
• WorldClim: Provides current and future climate layers for habitat modeling.
• MaxEnt: Builds a species-distribution model from presence-only data.
• QGIS: Lets you map occurrence points, inspect raster layers, and make final figures.
• R: Helps you clean data, compare model runs, and graph evaluation metrics.

Experiment Steps

1. Define one invasive plant species and one clear prediction about where climate may allow it to spread.
2. Gather occurrence records, then clean them for duplicates, obvious errors, and spatial clustering.
3. Choose climate variables and decide which ones you will hold constant across all runs.
4. Build a baseline model for the current climate, then compare it with future scenario layers.
5. Plan controls that test whether the model is better than random and whether output changes when you alter background selection.
6. Compare maps, summary metrics, and variable importance to answer your research question.

Common Pitfalls

• Using unfiltered GBIF records, which can include duplicate points, cultivated plants, or bad coordinates.
• Choosing too many correlated climate layers, which can make the model favor redundant variables.
• Skipping spatial bias checks, which can make the model overfit heavily sampled areas.
• Treating a future climate map as a real certainty instead of a scenario-based prediction.
• Changing the background region between runs, which makes model comparisons unfair.

What Makes This Competitive

A stronger project tests more than one modeling choice and explains why the predictions change. You can compare climate scenarios, variable sets, and background selection instead of reporting one map. You can also add a native species or a second invasive species as a comparison. Clear validation, careful data cleaning, and a thoughtful ecological explanation make the project much stronger.

Project Variations

• Use a different invasive plant species with a similar habitat, then compare how climate sensitivity changes.
• Test whether adding land cover or elevation improves prediction more than climate alone.
• Compare current climate predictions with two future emissions scenarios to see how the spread pattern shifts.

Learn More

• GBIF: Search for species occurrence records and metadata on the Global Biodiversity Information Facility site.
• WorldClim: Find current and future climate layers and documentation on the WorldClim site.
• NASA Earthdata: Explore climate and land data products that can support habitat modeling, then search by variable type.
• QGIS documentation: Learn how to inspect raster layers, clip maps, and visualize model output on the QGIS help pages.
• PubMed: Search for review articles on invasive species distribution modeling and MaxEnt validation methods.
• USGS invasive species resources: Look for plant invasion background, spread patterns, and management context on USGS pages.