Scientific evidence

The repository now has source-bound results for Nordic observed-pollen prediction, selected author method correspondence, and an English LRA reconstruction with forward-pollen predictions. Each supports a different claim. A calculation matching published author code says how faithfully a method was implemented; a withheld pollen score says how it predicts pollen on its tested sites; neither measures vegetation accuracy.

Start here	Question answered
Scientific concepts	What do counts, age bins, LRA components, error metrics and evidence roles mean?
Papers and evidence roles	Which original finding, calculation and validation claim does each paper support?
Reports and presentations	Download the nine canonical paper PDFs and three audience briefings for offline reading.
Country evidence status	What is scored, refused or still missing in each Nordic country?
Sources and scientific scope	Which four-country cohorts are counted, and why are some LRA inputs still refused?
Results and model use	Which numerical results agree or differ, how did direct and LRA pollen predictions score, and which parameters are supported?

The public pages summarize governed tables without distributing private raw sources. Exact rows, producer commands and limitations are linked from each page. The repository's claim-to-evidence index connects headline findings to numeric tables, manifests and producers. Open source questions remain visible in the repository's reports/needed register.

Use the application

Follow the atlas walkthrough to inspect one published modeled-context cell, trace its source and reproduce its bounded country/window summary. The route separates map behavior from scientific acceptance.

For installation and command usage, follow installation and setup and the supported interfaces. Read the input prerequisites before selecting an exploratory reconstruction; a working map or command does not supply missing scientific inputs.