Extreme Value Analysis (EVA) of Waves and Wind

Description

This Python script performs a Comprehensive Extreme Value Analysis (EVA) on environmental data. Specifically, it includes:

• Reading and rounding CSV data.
• Computing descriptive statistics and joint distributions.
• Fitting Generalized Extreme Value (GEV) distributions for overall data and by 30° sectors.
• Generating windrose plots to visualize the distribution of data against directional variables.
• Creating a landscape PDF report containing formatted tables and plots.

Input Format

The input CSV file should contain the following columns:

• datetime: Timestamps for the data points.
• swh: Significant wave height.
• mwd: Mean wave direction.
• pp1d: 1-day precipitation.
• wind: Wind speed.
• dwi: Wind direction.

The script expects these columns to be present for successful analysis.

Functionality Overview

1. Data Scaling: Rounds numerical values in specified columns.
2. Descriptive Statistics: Calculates basic statistical measures including skewness and kurtosis.
3. GEV Fitting: Fits GEV distributions to annual maxima of significant wave height (swh) and wind data.
4. Joint Distributions: Computes joint distributions for pairs of variables.
5. Sub-sector Analysis: Analyzes GEV distributions within 30° directional sectors.
6. Visualization: Generates windrose plots and empirical vs. GEV CDF plots.
7. Reporting: Outputs results in a detailed CSV report and compiles findings into a PDF report.

Dependencies

To run this script, ensure the following packages are installed:

• numpy
• pandas
• matplotlib
• scipy
• fpdf
• windrose

You can install the required packages using pip:

pip install numpy pandas matplotlib scipy fpdf windrose

Usage

To execute the script, provide the filename of the input CSV file as an argument:

python stats_era5_(swh_mwd_pp1d_wind_dwi).py <filename.csv>

The script will generate two output files:

• A detailed CSV report (.rpt.csv);
• A PDF report (.pdf) with visualizations and statistics.

Utility Functions

• round_variables(df): Rounds specific columns of a DataFrame to defined decimal places;
• format_interval(interval_str): Converts an interval string to a compact format;
• make_joint_distribution(df, var1, var2, bins1, bins2): Computes a 2D joint distribution between two variables in percentage;
• add_sums_and_highlight(df): Enhances a DataFrame for highlighting maximum values for PDF output;
• gev_fit(annual_max_series): Fits a GEV distribution to an annual maximum series;
• plot_gev_with_return_lines(...): Generates a plot comparing empirical CDF with a fitted GEV CDF;
• plot_windrose(df, var, dir_col, ...): Creates a windrose plot for a given variable against its directional data;
• pdf_print_table(pdf, df, ...): Renders a DataFrame as a formatted table onto a PDF page.

Main Analysis Workflow

extreme_value_analysis(input_csv)

This function encapsulates the entire analysis process. It handles:

• Reading of the input CSV;
• Data rounding and descriptive statistics computation;
• GEV fitting and analysis (both overall and by sector);
• Joint distribution calculations;
• Visualization creation including CDF plots and windrose plots;
• PDF report generation containing all results.