Overlay: Clip, Intersect, Union

Think back to the layer cake from Module 0. Overlay is what happens when you stop merely stacking layers and start cutting them against each other. Every overlay operation is really a set operation - the same AND, OR and NOT logic you know from Venn diagrams - applied to geometry and attributes at once.

There are a handful you will use constantly:

- Clip - use one layer as a cookie-cutter to trim another to a boundary. Keeps geometry from the first, discards everything outside the second. (Attributes of the cutter are not carried over.) - Intersect - keep only the area where two layers overlap, and carry the attributes of both. This is the workhorse. - Union - keep everything from both layers, splitting them where they cross, so every resulting piece knows which inputs it came from. - Difference (erase) - keep the part of one layer that is not covered by the other.

The trick is not running the tool - it is choosing the right verb for the question.

Clip and intersect produce the same shape - the overlapping area - which is exactly why they get confused. The difference is the attributes.

Clip is spatial trimming. You have a roads layer covering a whole district and a study-area polygon; clip the roads to the study area and you get just the roads inside it, still carrying only the roads' own columns. The study-area polygon was a stencil, nothing more.

Intersect is a spatial join by geometry. Overlay flood zones with plots using intersect and each resulting sliver carries both the plot's columns and the flood zone's columns. Now you can filter for flood_risk = high AND land_use = residential in one table. If your next step needs data from both layers, you wanted intersect. If you only needed to cut one layer to a boundary, clip is cleaner and faster.

Union is the generous one: it keeps every part of both inputs and splits them wherever they cross, tagging each piece with the attributes of whichever inputs cover it (and nulls where one does not). Use it when you are building a composite in which no area may be lost - a full land budget where every square metre must be accounted for.

Difference (called Erase in ArcGIS) removes the overlap: give it plots and a lakes layer and it hands back the dry land only. It is how you subtract the un-buildable from the buildable.

Dissolve is the tidy-up you will reach for after almost every overlay. Overlay tends to shatter your data into many small slivers; dissolve merges adjacent features that share a chosen attribute value back into single polygons - all the 'residential' fragments into one residential region. It is aggregation by attribute, and it turns a messy result into a clean, mappable one.

A practical warning about overlay outputs: watch for sliver polygons - hair-thin scraps produced where the edges of two layers nearly, but not exactly, coincide. They are the classic symptom of overlaying data digitised at different scales or from different sources (a Survey of India sheet against an OSM extract, say). Slivers inflate your feature count, distort area totals and clutter the map. The cure is partly upstream - clean topology and a shared CRS before you overlay, from Module 3 - and partly downstream: dissolve, or filter out features below a minimum area threshold. A good analyst inspects the attribute table and the smallest features after every overlay rather than trusting the output blind.

None of this began with software. In 1969, the landscape architect Ian McHarg published Design with Nature and described drawing each site constraint - slope, drainage, soils, ecology, scenic value - on a separate sheet of transparent paper, shading the unsuitable areas darkest. Stack the sheets on a light table and the darkest patches were where the most constraints piled up: the places to avoid. The clear patches were where to build. This 'sieve mapping' is manual overlay, and it is the direct ancestor of every suitability analysis a GIS runs today.

What the computer added was not the idea but the bookkeeping: it can overlay dozens of layers exactly, weight them, and recompute in seconds when a number changes. Jerome Dobson and later Jerome Malczewski formalised how you combine and weight the layers into land-use suitability. When you run an intersect-and-score workflow in Module 5's AHP suitability lesson, you are drawing McHarg's tracing paper - just faster, and honest about your weights.