Environmental Remote Sensing Training Course

Participants can select sensors and platforms appropriate to monitoring objectives and constraints.

• Passive optical, thermal and active radar sensing principles compared
• Landsat, Sentinel-2 and MODIS characteristics for environmental monitoring
• Spatial, spectral, radiometric and temporal resolution trade-offs evaluated
• Drone, aircraft and satellite platforms matched to monitoring scale
• Cloud cover, acquisition timing and seasonal phenology considered
• Sensor selection matrix built from question, budget and evidence needs

Participants can inspect raster properties and interpret common image-display combinations.

• Pixels, bands, digital numbers and reflectance values distinguished
• Raster metadata checked for projection, date, resolution and processing level
• True-colour, false-colour and infrared composites interpreted systematically
• Histogram stretching methods used to reveal relevant image features
• Mixed pixels, shadows and saturation identified during visual interpretation
• Raster alignment checked against vector reference and field observations

Participants can document spatial data fitness for a stated environmental analysis.

• Coordinate reference systems selected for area, distance and overlay accuracy
• ISO 19115 metadata elements recorded for imagery and derived products
• Positional, temporal and thematic accuracy assessed before analysis
• Data provenance tracked from source acquisition through final outputs
• Licensing, access restrictions and data-sharing obligations reviewed
• Quality assurance checklist applied to environmental spatial datasets

Participants can prepare imagery for analysis using documented preprocessing steps.

• Surface reflectance products distinguished from top-of-atmosphere reflectance products
• Cloud, cirrus and shadow masks applied using quality-assurance bands
• Atmospheric correction purpose and residual limitations explained clearly
• Study-area boundaries used to clip imagery and reduce processing
• Band stacks assembled with consistent extent, resolution and projection
• Mosaic and compositing methods used for multi-scene coverage

Participants can calculate and interpret spectral indices for environmental indicators.

• Spectral signatures compared for vegetation, soil, water and built surfaces
• NDVI calculated to assess vegetation greenness and seasonal condition
• NDWI and MNDWI applied to surface-water detection questions
• NBR used to identify burn severity and vegetation disturbance
• Bare Soil Index interpreted alongside land-cover and moisture context
• Index thresholds tested against reference imagery and field evidence

Participants can design vegetation-condition monitoring using seasonal image observations.

• Phenology curves distinguished from disturbance and persistent vegetation decline
• NDVI anomalies calculated against multi-year seasonal reference values
• Crop calendars aligned with image acquisition and interpretation windows
• Vegetation condition maps classified using documented threshold rules
• Drought indicators combined with rainfall, soil and field observations
• Management zones identified from spatial patterns and uncertainty evidence

Participants can derive spatial evidence for surface water and wetland condition.

• Water occurrence mapped using spectral indices and quality masking
• Turbidity proxies interpreted with calibration and atmospheric limitations
• Wetland extent delineated from seasonal imagery and elevation data
• Riparian vegetation condition assessed using buffer-based raster summaries
• Catchment land-cover proportions calculated for water-risk screening
• Water maps reviewed against field records and high-resolution imagery

Participants can use terrain derivatives to support environmental interpretation and risk mapping.

• Digital elevation model sources compared for resolution and vertical accuracy
• Slope, aspect and hillshade derived from elevation surfaces
• Flow direction and flow accumulation modelled for drainage patterns
• Topographic wetness index interpreted for runoff and saturation screening
• Terrain correction needs identified for mountainous optical imagery
• Elevation derivatives combined with land-cover and hydrological evidence

Participants can design reference data suitable for classification and validation.

• Reference classes defined using observable and mutually exclusive categories
• Stratified random sampling applied across land-cover classes and regions
• Training and validation samples separated to prevent biased accuracy results
• Field forms designed for location, class, condition and photo evidence
• High-resolution imagery used cautiously as reference-data support
• Sample representativeness reviewed across season, terrain and class variability

Participants can produce a supervised land-cover classification from prepared imagery and training data.

• Classification schema aligned with monitoring objectives and reporting categories
• Training polygons digitised from verified reference observations and imagery
• Maximum likelihood classification applied to multispectral image bands
• Random forest classification compared with parametric classification assumptions
• Class confusion diagnosed through spectral overlap and sample review
• Post-classification filtering applied without obscuring meaningful small features

Participants can quantify classification performance and report limitations honestly.

• Confusion matrix constructed from independent validation reference samples
• Overall accuracy, producer's accuracy and user's accuracy calculated
• Omission and commission errors interpreted for decision consequences
• Area-adjusted estimates derived from stratified classification validation results
• Confidence intervals reported alongside mapped environmental class estimates
• Uncertainty statements written without implying unsupported environmental change

Participants can select and apply a change detection approach for environmental monitoring.

• Image differencing applied to matched seasonal reflectance or index images
• Post-classification comparison used to quantify land-cover transitions
• Change vector analysis introduced for multi-band directional change assessment
• Temporal consistency checks separate real change from acquisition artefacts
• Minimum detectable change defined from resolution, accuracy and class variability
• Transition matrices summarised for reporting land-cover conversion patterns

Participants can structure time-series monitoring and run reusable cloud-based image analyses.

• Image collections filtered by location, date, sensor and cloud cover
• Median composites created to reduce cloud and outlier effects
• Seasonal time series charted for vegetation and water indicators
• Google Earth Engine scripts structured with variables, functions and exports
• Trend lines distinguished from seasonal cycles and isolated anomalies
• Batch outputs named consistently for traceable monitoring periods

Participants can communicate remote sensing findings through clear maps and defensible reporting.

• Map purpose, audience and decision context defined before layout design
• Symbology selected for categorical classes, indices and change magnitude
• Scale bars, north arrows and coordinate grids applied appropriately
• Methods sections document data sources, processing and validation results
• Interpretation separates observed patterns from plausible causal explanations
• Limitations, assumptions and recommended field checks stated explicitly

Participants can build, review and present a complete environmental remote sensing monitoring project.

• Monitoring question scoped with measurable indicators and decision users
• Suitable imagery, reference data and processing methods justified
• Prepared datasets analysed through indices, classification or change detection
• Accuracy evidence and uncertainty statements reviewed against project claims
• Decision-ready maps, charts and technical notes assembled coherently
• Peer review addresses reproducibility, interpretation and action recommendations