What counts as the WUI boundary
At first glance, the Wildland–Urban Interface seems like the sort of boundary that should be easy to name. One imagines houses on one side, vegetation on the other, and a line that can be drawn where the two meet. But the apparent simplicity of that picture dissolves as soon as one asks what, precisely, is being bounded. Before any perimeter can be measured, a boundary object must first be brought into being. In this project, that prior act of definition is represented by the term (d): the delineation choice that determines what the WUI boundary is taken to be before its length is ever computed.
This matters because the WUI is not a naturally given curve waiting passively in the landscape to be recovered by better instruments. It is a spatial construction assembled from decisions about representation, threshold, adjacency, and scale. A settlement may be represented as a point, a polygon, or a gridded proxy. Vegetation may be classified according to stricter or looser thresholds. Neighbor relations may be defined through immediate contact, through a radius of influence, or through a more expansive notion of connectivity. Each of these choices may be scientifically defensible. None is neutral. Together they determine the geometry of the object whose perimeter we later report.
Seen in that light, the WUI boundary is not a single immutable feature but a family of possible boundaries generated by alternative, and often reasonable, delineation rules. Changing those rules does not merely perturb a measurement; it changes the object being measured. This is the first source of variation that the project seeks to keep in view. If one delineation produces a smoother and more aggregated boundary while another produces a more intricate and fragmented one, the resulting difference is not yet a story about measurement error or computational nuance. It is a story about what the boundary has been taken to mean.
For that reason, the quantity of interest is best written not simply as a length, but as (L_d(\varepsilon)): measured boundary length as a function of both delineation choice and measurement scale. The symbol (d) reminds us that perimeter begins with an act of definition. The symbol (\varepsilon) reminds us that even once a boundary object has been fixed, its reported length still depends on the scale at which it is traced. Holding (d) constant while varying (\varepsilon) reveals measurement-scale sensitivity. Holding (\varepsilon) constant while varying (d) reveals object-definition sensitivity. These are related but fundamentally distinct sources of variation, and the scientific argument depends on not collapsing them into one another.
That separation is more than a matter of tidy notation. Without it, disagreements in reported perimeter can easily be mistaken for ecological differences in the world when they are in fact consequences of method. One study may appear to describe a more irregular or extensive interface than another not because the landscapes differ in any deep biophysical sense, but because the boundary objects were constructed differently at the outset. Another pair of studies may hold their delineation logic constant yet arrive at different lengths simply because they measured that same object at different effective ruler lengths. If these sources of variation are not explicitly disentangled, interpretation begins to drift. Methodological choices acquire the appearance of ecological truth.
The purpose of this page, then, is to slow the reader down before measurement begins. The central question of the project is often framed as a question of length: how long is the WUI boundary? But that question contains a prior and more subtle one: what, exactly, counts as the WUI boundary in the first place? Only by answering that question carefully can the later analysis of scaling be understood on its proper terms. The line whose length we report is never independent of the conceptual and computational decisions that first called it into existence.