MerQuR+Statistics Glossary

Statistics Glossary → Ripley's K/L and the G, F, J functions

Ripley's K/L and the G, F, J functionsconcept

Summary functions that examine whether a point pattern is clustered, random or regular at different distance scales by comparing it with complete spatial randomness (CSR).

Ripley's K function is the expected number of other points within radius r of a typical point, divided by the point intensity (λ). Under CSR, K(r) = πr²; larger values indicate clustering at that scale and smaller values regularity. The transformation L(r) = √(K(r)/π) gives a straight line under CSR (L(r) = r) and is easier to read; the pair correlation function g(r) gives the relative frequency of pairs exactly at distance r (above 1 means clustering).

G(r) is the distribution of nearest-neighbour distances and F(r) the distribution of the distance from a random location to the nearest point (the empty-space function). J(r) = (1 − G(r)) / (1 − F(r)) equals 1 under CSR; values below 1 indicate clustering and above 1 regularity. All these functions are affected by incomplete neighbourhoods at the edge of the study area and need edge correction. Departures of the observed curve from CSR are usually judged with simulation envelopes; reading the envelopes separately at each r raises a multiple-testing problem.

An everyday example

The locations of 300 oak trees in a forest are examined. The L(r) − r curve stays above the CSR envelope between 5 and 30 metres: trees cluster at that scale, probably because acorns fall near the parent tree. Below 2 metres, however, the curve lies under the envelope; trees cannot stand very close together, because competition between crowns and roots creates regularity at small scales.

In MerQur

The 'Summary functions: K, L, g, G, F, J' checkboxes and the 'Upper limit of r (K, L, g)' and 'Upper limit of r (G, F, J)' parameters (0 = automatic) in Point Pattern Analysis.

Analyses where this concept appears

Point Pattern Analysis (Ripley K/L, G/F/J, CSR)

Related concepts

Complete spatial randomness (CSR)Monte Carlo (simulation) envelope