This project investigates how extreme weather, specifically drought, dzud and extreme rainfall, creates climate risk in semi‑arid and arid regions. Using satellite derived precipitation data, reanalysis data and geospatial analysis in Python, the project applies percentile based and persistence based frameworks to evaluate climate extremes relative to local climatology.
The completed part of the analysis focuses on Chad during 2011, a year embedded within a multi year regional drought that led to severe food insecurity across the Sahel. Although national annual rainfall appeared near normal, this project demonstrates that seasonal timing, spatial heterogeneity, and persistence of dry conditions explain the severity of impacts. The continuation of the analysis will involve the analysis of Mongolia during 2009-2010, a 10 month time frame in which serious drought and low rainfall in the summer of 2009 set up for poor conditions going into the winter dzud that lasted well into 2010, resulting in serious food insecurity and mass migration to the urban centres by the rural subsistence farmers of Mongolia.
This repository is intended as a portfolio‑quality demonstration of applied climate data science and geospatial analysis.
Note: Visual themes for all visualisations and figures was selected off of a single accent colour. This colour was chosen for personal consistency and as a reminder to of explaining concepts out loud to someone who may not have knowledge in the area.
The climatological analysis of Chad in 2011 has been completed as of 29/12/2025
This project combines temporal, seasonal, and spatial analyses:
Historical Baseline
- 29 year climatology (1981–2009) to represent pre drought conditions and provide context to the 2011 drought
- Annual and wet season rainfall distributions
- Anomalies, standardized z scores, and percentile thresholds
Event Detection
- Drought (local scale): 7 day rolling mean 10th percentile thresholds of rainfall in N'Djamena
- Drought (country scale): ≥7 consecutive days with rainfall < 1 mm per grid cell Persistence based definition suitable for semi arid climates
- Extreme rainfall: Daily precipitation above the 90th percentile per grid cell
Spatial Analysis
- Grid cell specific thresholds to account for strong climatic gradients
- Percentile normalization to enable comparison across desert, Sahelian, and tropical zones
- Representative drought and extreme rainfall days mapped across Chad
Annual Rainfall Is Misleading
National annual rainfall in 2011 was near the historical mean. This masks severe climate stress experienced on the ground.
Wet‑Season Failure Explains the Crisis
June–September rainfall in the Sahel fell below the 10th percentile. 2011 ranks among the driest wet seasons in the historical record.
Drought and Flood Risk Co‑exist
Large portions of Chad experienced persistent drought. Localized regions simultaneously experienced extreme rainfall events.Highlights compound climate risk in semi arid regions.
Spatial Perspective Is Essential
National averages hide regional extremes. Percentile spatial mapping reveals coherent drought patterns and localized rainfall extremes.
Python, xarray, Pandas & GeoPandas, rioxarray, NumPy, Matplotlib, Seaborn
This notebook is still being worked upon.
This notebook is still being worked upon.
- Comparative case study in Mongolia and the Dzud of 2009-2010
- Cross regional comparison of drought and extreme rainfall patterns
- Summary statistics comparing climate risk profiles across regions