Tropical Cyclone Rapid Intensification Trends

ISEF Category: Earth and Environmental Sciences

Subcategory: Climate Science  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

A hurricane can jump from manageable to devastating faster than many forecasts can react. That jump is called rapid intensification, and it can change lives in one day. You can study whether that jump is happening more often, and which ocean or atmosphere factor seems to drive it most.

What Is It?

Rapid intensification means a tropical cyclone gets much stronger in a short time. Think of it like a car that suddenly hits the gas on a wet road. The storm does not just get a little bigger, it can quickly become far more dangerous.

You would study this with two public data sources. IBTrACS gives storm track and intensity records. ERA5 gives environmental features such as sea surface temperature, which is the ocean skin temperature, and vertical wind shear, which means a change in wind speed or direction with height. Those variables can act like fuel and friction for a storm. Warm water can add energy, while strong shear can tear a storm apart. SHAP is a method that helps you see which variables your model leaned on most when it made a prediction.

Why This Is a Good Topic

This topic works well because the question is measurable, the data are public, and the answer is not obvious. You can test whether rapid intensification frequency changes by decade, then ask which environment features line up with those changes. That connects to forecasting, disaster planning, and climate risk. You can also learn real research skills, like cleaning time series data, building a classification model, and interpreting feature importance without a wet lab.

Research Questions

• How does the frequency of rapid intensification events change across decades in the Atlantic basin?
• What is the effect of sea surface temperature on the odds that a storm rapidly intensifies?
• Does vertical wind shear explain rapid intensification better in one decade than in another?
• To what extent do sea surface temperature and wind shear together improve a model of rapid intensification compared with either variable alone?
• Which environmental variable has the highest SHAP importance for rapid intensification prediction in each decade?
• How does storm intensity at the start of a six-hour period affect the chance of rapid intensification?
• Which basin, Atlantic or East Pacific, shows the strongest decade-by-decade shift in rapid intensification frequency?

Basic Materials

• Laptop with enough storage for gridded climate data
• Internet access for downloading IBTrACS and ERA5 data
• Spreadsheet software for basic sorting and checks
• Python installed with pandas, numpy, matplotlib, scikit-learn, and shap
• External drive or cloud storage for backups
• Notebook for tracking variable definitions, filters, and model choices.

Advanced Materials

• High-performance laptop or university workstation
• Python with xarray, netCDF4, pandas, scikit-learn, shap, and cartopy
• Access to large ERA5 netCDF files or a climate data server
• GIS or mapping software for basin filtering and track visualization
• Version control system such as Git for reproducible analysis
• Optional cluster access for repeated model runs and decade-specific resampling.

Software & Tools

• Python: Cleans the storm records, joins environmental features, and runs the predictive model.
• Pandas: Organizes IBTrACS tables and prepares decade-based subsets.
• Xarray: Reads ERA5 netCDF files and extracts gridded climate variables.
• Scikit-learn: Builds classification models that predict rapid intensification.
• SHAP: Estimates which variables most influenced model predictions in each decade.

Experiment Steps

1. Define your rapid intensification label and choose the exact storm archive years you will analyze.
2. Build a data table that matches each storm time step with the closest environmental conditions from ERA5.
3. Decide which predictors you will test first, such as sea surface temperature, wind shear, and starting intensity.
4. Split the data by decade and plan a model for each time period so you can compare changes over time.
5. Choose one evaluation metric and one explanation method, then predefine how you will compare variable importance across decades.
6. Set rules for missing data, outliers, and basin selection before you run the full analysis.

Common Pitfalls

• Using inconsistent rapid intensification thresholds, which changes the label and makes decade comparisons meaningless.
• Mixing basin records with different storm naming and tracking rules, which can create false trends.
• Matching ERA5 values to storm points without checking time lag, which can pair the wrong environment with the wrong storm state.
• Treating all storm observations as independent, which inflates confidence because the same storm appears many times.
• Reading SHAP values as direct cause, when they only explain how the model used the variables you gave it.

What Makes This Competitive

A strong project does more than count storms. It tests whether the trend holds after careful filtering, basin-by-basin checks, and decade-specific models. You can raise the level by comparing multiple feature sets, using resampling or bootstrapping, and asking whether the same variable stays dominant across decades or flips over time. A clean, reproducible pipeline matters as much as the final chart.

Project Variations

• Focus on one basin, such as the Atlantic, and compare decade-level rapid intensification trends within that basin only.
• Replace SHAP with permutation importance and see whether the top environmental driver stays the same across models.
• Add ocean heat content or relative humidity to test whether a multivariable model beats sea surface temperature and wind shear alone.

Learn More

• IBTrACS, NOAA, and the National Hurricane Center: Search for the International Best Track Archive for Climate Stewardship to download storm track and intensity data.
• ERA5, Copernicus Climate Data Store: Search the climate data store for reanalysis fields such as sea surface temperature and wind shear inputs.
• NOAA Climate.gov: Read explainer articles on hurricane intensification, ocean heat, and wind shear.
• NASA Earth Observatory: Find background articles on tropical cyclones, ocean temperatures, and satellite-based storm monitoring.
• MIT OpenCourseWare: Search for courses on statistics, machine learning, and climate science to strengthen your analysis workflow.