R project with data and code to reproduce the analysis on

This R project contains everything to reproduce the results found in the paper titled "Tackling spatial sampling bias in biodiversity records “blindly”: A simple solution to improve species distribution models and conservation decisions". Data contained is publicly available from diverse sources, all of which are detailed both in the main text of the manuscript and within the code documentation.The paper's abstract is reproduced below: Abstract AimSpatial sampling bias (SSB) arises when spatial variation in reporting rate—the probability that a species present at a site is observed and reported—is not accounted for in species distribution models (SDMs). Existing bias-correction methods implicitly assume a homogeneous observation process across presence-only records, an assumption often violated when opportunistic observations are pooled with systematically collected survey data in biodiversity aggregator databases. We aim to demonstrate how this mixing weakens the detectable SSB signal. We propose a simple metadata solution that would improve SDM predictions and conservation decisions. LocationAustralia. MethodsUsing > 2.3 million georeferenced marsupial records from the Atlas of Living Australia, we applied extensive metadata filtering to discriminate records likely derived from systematic surveys from opportunistic observations. We compared spatial patterns of record density before and after filtering, and quantified the strength of associations between record density and known drivers of SSB (human population density and accessibility) using generalized linear models. ResultsApproximately 75% of records were classified as likely systematic, although this required labour-intensive and subjective filtering due to inconsistent metadata. Filtering out systematic records strengthened the association between presence-only records and known drivers of sampling bias, improving the accountability of the reporting-rate signal. Main conclusionsPooling records with fundamentally different observation processes undermines attempts to diagnose and correct SSB when using presence-only data. We propose adding a simple, standardised “incidental observation” metadata flag at data intake AND retrospectively to allow bias-correction methods to be applied where most appropriate and therefore strengthen SDMs used for conservation planning. If made available across presence-only biodiversity aggregator databases, this information would unlock substantial gains from existing and emerging analytical tools. Keywords Citizen science, data integration, observation process, observers bias, sampling bias, sampling effort, species distribution models, reporting rate.