AI for Climate & Environment

Course outline and learning expectations

This is a tutor-led course. The outline shows what your tutor will cover; teaching materials and examinations are provided directly to enrolled students.

What you will learn

• Identify and assess climate, rainfall, satellite and environmental data sources relevant to Uganda, East Africa and Africa
• Prepare, clean, visualise and combine time-series and geospatial climate datasets
• Build and evaluate baseline rainfall and drought prediction models while avoiding time-series data leakage
• Calculate and interpret rainfall anomalies and drought indicators for monitoring and early warning
• Use satellite imagery and machine-learning methods to map land cover and monitor possible deforestation or vegetation change
• Develop baseline models for seasonal water availability and borehole-yield screening using climate, terrain and hydrogeological features
• Design flood and drought early-warning workflows with thresholds, lead times, uncertainty and impact-based messaging
• Estimate basic carbon and environmental impacts using transparent assumptions, activity data and emission factors
• Create accessible community-facing alerts for web, mobile, SMS, USSD, WhatsApp or radio workflows
• Build, document and present a Gomba District rainfall and drought early-warning dashboard prototype

Modules

1. 1

   Climate data sources for Uganda, East Africa and Africa

   Find, assess and prepare climate and Earth observation data for responsible analysis.

   Uganda Department of Meteorological Services and UNMA productsICPAC climate services, East Africa Drought Watch and Hazards WatchCHIRPS rainfall data and gridded precipitationCopernicus Sentinel, Landsat and Google Earth Engine cataloguesFAO WaPOR water-productivity dataSpatial and temporal resolution, metadata and licensingMissing data, bias, ground observations and local validation
2. 2

   Rainfall and drought prediction models

   Develop reproducible baseline and machine-learning models for rainfall and drought analysis.

   Define prediction targets, lead times and decision questionsRainfall time-series exploration and seasonal patternsFeature engineering with lags, rolling statistics and climate indicatorsPersistence, climatology and statistical baselinesRegression and classification models for rainfall and drought riskRainfall anomalies, SPI concepts and vegetation indicatorsTime-aware validation, data leakage, model metrics and uncertainty
3. 3

   Deforestation monitoring using satellite imagery

   Use Earth observation and machine learning to detect land-cover and vegetation change.

   Remote-sensing fundamentals for African landscapesSentinel-2 and Landsat image selectionCloud masking, mosaics and image compositesVegetation, moisture and burn indices including NDVI and NBRSupervised land-cover classificationChange detection and possible forest-loss alertsAccuracy assessment, reference data and limitations
4. 4

   Water-resource prediction and seasonal availability

   Apply climate, terrain, satellite and hydrogeological data to water-resource screening.

   Water-resource questions and data requirementsRainfall, evapotranspiration, soil moisture and vegetation indicatorsFAO WaPOR and other open water-related datasetsTerrain, geology, land cover and borehole-record featuresBaseline borehole-yield screening modelsSeasonal water-availability analysisValidation, uncertainty and responsible interpretation
5. 5

   Early-warning systems for floods and drought

   Design practical warning workflows that connect hazard information to early action.

   Hazard, exposure, vulnerability and riskMonitoring indicators, thresholds and alert levelsFlood and drought lead times and forecast horizonsImpact-based forecasting and anticipatory actionFalse alarms, missed events and model uncertaintyData pipelines, update schedules and dashboard architectureGovernance, escalation and communication protocols
6. 6

   Carbon and environmental impact estimation basics

   Learn transparent approaches for preliminary carbon and environmental impact analysis.

   Purpose and boundaries of an environmental assessmentActivity data, emission factors and calculation logicLand-use change, vegetation and carbon-stock conceptsBaseline and project scenariosData quality, assumptions and uncertainty rangesAvoiding double counting and false precisionClear reporting for decision-makers and communities
7. 7

   Community-facing environmental alert tools

   Turn technical risk information into understandable and usable community alerts.

   User research with farmers, local leaders and extension workersAlert content, timing and calls to actionLocal language, visual symbols and accessibilityLow-bandwidth web, SMS, USSD, WhatsApp and radio workflowsLocation-specific alerts and consent-based messagingFeedback, correction and trust-building mechanismsData privacy, responsible AI and inclusive design
8. 8

   Capstone — a climate-risk prototype

   Build and present a rainfall and drought early-warning dashboard prototype for Gomba District farmers.

   Define the farmer decision problem and dashboard usersSelect and document rainfall, drought and vegetation datasetsPrepare a reproducible data pipelineTrain and evaluate a baseline risk modelCreate maps, charts, alert levels and action-oriented messagesTest usability and review uncertainty with local contextDocument limitations, deployment needs and next stepsPresent the final prototype