Zonal summary of a surface — Task
Implementation: Aggregating a continuous raster to zones (Recipe - QGIS).
Zonal summary of a surface is the Task of describing a continuous raster — soil organic carbon, elevation, a vegetation index — inside each zone with a few numbers: mean, median, a high percentile. It is how a map of values becomes a comparison between landscapes.
Used in examples: Soil carbon and land use at Ledreborg (Example).
Task Schema (tool-agnostic)
- Know the surface (decision point) — what the value means, its unit, depth or date, and how it was made (measured, interpolated from samples, modelled from satellite data).
- Bring surface and zones to the same reference (Operator: reproject if needed) — same coordinate system; the study area plus a margin.
- Compute statistics per zone (Operator: zonal statistics) — count, sum, mean, median, a high percentile.
- Combine zones of the same type (decision point) — from sums and counts, not from the zones’ means.
- Optionally, split by a category (Operator: zonal statistics with a categorical raster as zones) — e.g. the mean value per land-use class.
Output: a table of statistics per zone or zone type, ready to set beside other zonal results (such as land-use composition).
Epistemic Choices to record
- Concentration or stock: soil organic carbon in percent is a concentration; tonnes of carbon per hectare would also need bulk density and depth. Say which the analysis is about.
- Mean or median: a few extreme cells (peat in a valley) pull the mean far above the median. Report both when they differ.
- What the surface covers: a soil model predicts values under lakes and buildings too; leave such cells out or say why they stay.
- Model uncertainty: a modelled surface is an estimate with its own error; differences between zones smaller than that error are not findings.