january-msemakweli/Mann-Kendall-Analysis-of-Climate-variables-in-Dar-es-Salaam-2014-to-2024-: Decadal Trends in Seasonal Climatic Variables in Dar es Salaam, Tanzania: A Non-Parametric Approach Using the Mann-Kendall Test

Decadal Trends in Seasonal Climatic Variables in Dar es Salaam, Tanzania: A Non-Parametric Approach Using the Mann-Kendall Test Overview This repository contains R code and documentation for conducting Mann-Kendall trend analysis on seasonal climatic variables in Dar es Salaam, Tanzania. The analysis uses monthly meteorological data from the Tanzania Meteorological Authority covering a 10-year period (2014–2024) to detect and quantify monotonic trends in rainfall, temperature, and relative humidity across different seasons. The Mann-Kendall test is a non-parametric statistical method widely used in climatology and hydrology for detecting trends in time series data. It does not require assumptions about data distribution, making it particularly suitable for environmental data analysis. This implementation includes Sen's slope estimator to quantify the magnitude and direction of detected trends. Author January G. Msemakweli Graduate Student | ScM in Epidemiology Bloomberg School of Public Health Johns Hopkins University https://orcid.org/0009-0007-6743-8479 Features Non-parametric trend detection: Mann-Kendall test implementation with Z-score calculation Trend quantification: Sen's slope estimator for magnitude assessment Seasonal analysis: Separate trend analysis for each season Statistical power analysis: Minimum Detectable Trend (MDT) calculations Autocorrelation diagnostics: Lag-1 ACF assessment with Trend-Free Pre-Whitening (TFPW) Change-point detection: Pettitt test for identifying abrupt shifts OLS sensitivity analysis: Parametric validation of non-parametric results Comprehensive visualization suite: Publication-ready plots Statistical tables: Formatted HTML tables with trend test results Requirements Software R (version ≥ 4.0.0 recommended) RStudio (optional, for R Markdown rendering) R Packages The following R packages are required: readr # Data import dplyr # Data manipulation Kendall # Mann-Kendall test trend # Sen's slope estimator and Pettitt test knitr # Dynamic report generation kableExtra # Enhanced table formatting ggplot2 # Data visualization tidyr # Data tidying broom # Model output tidying (for OLS analysis) Installation Install all required packages using: install.packages(c("readr", "dplyr", "Kendall", "trend", "knitr", "kableExtra", "ggplot2", "tidyr", "broom")) Repository Structure . ├── README.md # This file ├── MK_Test.Rmd # Main Mann-Kendall trend analysis ├── OLS_Sensitivity_Check.Rmd # OLS regression sensitivity analysis ├── CLIMATE DAR DATASET.csv # Input climate data (required) └── [generated outputs] # PNG plots and HTML reports Analysis Files MK_Test.Rmd — Main Trend Analysis The primary analysis document containing: Mann-Kendall Trend Test: Non-parametric trend detection for each season-variable combination with Z-scores, p-values, and Bonferroni correction Statistical Power Analysis: Calculates Minimum Detectable Trends (MDT) at 80% power to contextualize non-significant findings Lag-1 Autocorrelation Diagnostics: Assesses serial correlation and its implications for Type I error rates Trend-Free Pre-Whitening (TFPW): Applies pre-whitening to series with substantial autocorrelation (|ACF| ≥ 0.3) Pettitt Change-Point Test: Detects abrupt shifts in time series central tendency Visualizations: 8 publication-ready plots including time series, box plots, and trend magnitude charts OLS_Sensitivity_Check.Rmd — Parametric Validation A sensitivity analysis using Ordinary Least Squares (OLS) regression to validate the non-parametric Mann-Kendall results: OLS Regression: Fits linear models (Y ~ time) for each variable and season Method Comparison: Compares OLS slopes with Sen's slopes, and t-test significance with Mann-Kendall significance Agreement Statistics: Quantifies direction agreement and significance agreement between methods Regression Diagnostics: Shapiro-Wilk normality test and Durbin-Watson autocorrelation test Slope Comparison Plot: Scatter plot showing OLS vs Sen's slope with 1:1 reference line This sensitivity check strengthens confidence in findings when both parametric and non-parametric methods agree. Data Requirements The analysis requires a CSV file named CLIMATE DAR DATASET.csv with the following columns: Year: Year of observation (e.g., 2014, 2015, ...) Month: Month abbreviation (e.g., Jan, Feb, Mar, ...) Season: Season classification (e.g., Dry, Wet, etc.) rainfall_mm: Rainfall in millimeters daytime_temperature_celcius: Daytime temperature in °C nighttime_temperature_celcius: Nighttime temperature in °C relative_humidity_%: Relative humidity in percentage Usage Open the desired .Rmd file in RStudio Ensure CLIMATE DAR DATASET.csv is in the working directory Click Knit to generate an HTML report with embedded results and visualizations # Knit from the console rmarkdown::render("MK_Test.Rmd") # Main analysis rmarkdown::render("OLS_Sensitivity_Check.Rmd") # Sensitivity analysis Expected Outputs Running the analysis will generate: High-resolution PNG plots (300 DPI) HTML tables with statistical results Console output showing progress and key findings Methodology Statistical Methods Mann-Kendall Test: Non-parametric test for monotonic trends Null hypothesis (H₀): No monotonic trend exists Alternative hypothesis (H₁): A monotonic trend exists Test statistic: Z-score (standardized Mann-Kendall S statistic) Significance level: α = 0.05 Sen's Slope Estimator: Robust non-parametric slope estimator Calculates the median of all pairwise slopes Units: change per time period (e.g., mm/year, °C/year) Analysis Workflow Data Preparation: Load and format the climatic dataset with proper date handling Variable Selection: Define climatic variables of interest and their labels Seasonal Stratification: Separate analysis for each season Trend Detection: Apply Mann-Kendall test to each variable-season combination Trend Quantification: Calculate Sen's slope for detected trends Visualization: Generate comprehensive plots showing trends and patterns Results Export: Save statistical tables and publication-ready figures Outputs Statistical Results Table A formatted HTML table containing: Season Variable name Z-score (test statistic) p-value (statistical significance) Sen's slope (trend magnitude) Interpretation Guidelines Statistical Significance p < 0.05: Statistically significant trend at the 95% confidence level |Z| > 1.96: Corresponds to p < 0.05 for a two-tailed test Trend Direction Positive Sen's slope: Increasing trend over time Negative Sen's slope: Decreasing trend over time While statistical significance indicates that a trend is unlikely due to chance, consider the magnitude (Sen's slope) for practical implications. Small but statistically significant trends may have limited practical importance for long-term climate planning. Citation If you use this code or methodology in your research, please cite: @software{msemakweli2024mannkendall, author = {Msemakweli, January G.}, title = {{Decadal Trends in Seasonal Climatic Variables in Dar es Salaam, Tanzania: A Non-Parametric Approach Using the Mann-Kendall Test}}, year = {2024}, publisher = {Zenodo}, doi = {10.5281/zenodo.17720494}, url = {https://doi.org/10.5281/zenodo.17720494} } Data Source Climate data sourced from: Tanzania Meteorological Authority (TMA) License This project is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). You are free to: Share: Copy and redistribute the material in any medium or format Adapt: Remix, transform, and build upon the material for any purpose, even commercially Under the following terms: Attribution: You must give appropriate credit, provide a link to the license, and indicate if changes were made Contributing Contributions, issues, and feature requests are welcome. For major changes, please open an issue first to discuss what you would like to change. Acknowledgments Tanzania Meteorological Authority for providing the climate data The R community for developing and maintaining the statistical packages used in this analysis Contact For questions, suggestions, or collaboration opportunities, please contact: January G. Msemakweli Graduate Student | ScM in Epidemiology Bloomberg School of Public Health Johns Hopkins University 615 N Wolfe St ORCID: 0009-0007-6743-8479 Keywords: Mann-Kendall test, Sen's slope, climate trend analysis, time series analysis, Dar es Salaam, Tanzania, rainfall trends, temperature trends, seasonal analysis, non-parametric statistics Last Updated: January 2026