Bayesian spatial record linkage for multi-temporal LiDAR (Plantae)

Knowledge graph centered on Bayesian spatial record linkage for multi-temporal LiDAR (Plantae) with 35 nodes and 122 connections. Top connected: Rocky Mountain Biological Laboratory, Gothic, Crested Butte, Engelmann Spruce, Populus tremuloides.

Method synopsis

A two-stage Bayesian hierarchical approach that identifies unique individuals across LiDAR scans by modeling observed locations as noisy transformations of latent true locations, then propagates linkage uncertainty into downstream analysis through linkage-averaging.

Synthesized from method descriptions across 1 paper using this protocol.

Procedure as described in the canonical source

Steps below were extracted from the paper that introduces this protocol — A Bayesian record linkage approach to applications in tree demography using overlapping LiDAR scans (2025), Annals of Applied Statistics. Implementations in other papers (listed below) may differ.

1. Bayesian record linkage model fitting
   Spatial record linkage model implemented using Markov chain Monte Carlo (MCMC) Gibbs sampling with spatial bounding box constraints. Observed locations modeled as noisy transformations of latent true locations with rotation, translation, and Gaussian measurement error. Linkage structure sampled by considering only latent individuals within maximum distance bounding boxes around observed records, with iterative expansion of box sizes until minimum candidate requirements met. Model written in R and C++ using Rcpp for computational optimization.
   Quantities: Bounding box margins varied from 2.5m to 200m, minimum 2 candidate matches required per bounding box, maximum latent population size N = q × max(ni) where q = 1.25Duration: MCMC sampling with multiple chains, convergence assessed using Gelman-Rubin statisticsConditions: Computational sampling under spatial constraints to improve efficiency
   Equipment: R statistical software, Rcpp package, C++ implementation, RcppArmadillo package

Typical Equipment

• R statistical software
• Rcpp package
• C++ implementation
• RcppArmadillo package

Output Measurements

• linkage probabilities
• latent individual locations
• linkage structure samples

Papers Using This Protocol (1)

Species Studied With This Protocol

Places Where Used