The 2nd and 3rd National Survey on the Natural Environment: Vegetation Survey (common species)

Geographic coordinates were generalized from original locality to representative coordinates in consideration of protecting sensitive species. The center of either secondary mesh (10x10km), the prefectural capitals or the capital of Japan was used as the representative coordinates. The coordinates were obtained from National Land Numerical Data (MLIT 2022a, MLIT 2022b). The closest terrestrial point from the center was chosen as an alternative point instead of the point in the sea for secondary mesh level coordinates. The maximum distance from the coordinates to the polygon edge was calculated as the accuracy of coordinates. R programs and ArcGIS were used for calculation (竹中 2014, Hijmans 2021, Karney 2013, Pebesma 2018).

Recorded Japanese common names were cleaned since some obvious typos scattered throughout the data. Data deletion was avoided as much as possible during data cleaning, however 161 occurrence records with incomplete species names such as blanks, "?" and "? sp" were deleted. Ambiguous search using R package: stringdist（van der Loo 2014) was conducted against the checklist of Japanese plant names (Yamanouchi et al. 2019), which is available on the JBIF website, to get candidates during data cleaning. The checklist was also used to get scientific names based on Japanese common names. Scientific names in Green List (Ebihara et al. 2016, Ito et al. 2016) and YList (based on Yonekura and Kajita 2003–) were selected in priority order, referring to other sources (Ebihara 2016a, 2016b, 大橋ほか（編）2015、2016a、2016b、2017a、2017b) in the checklist as appropriate. Taxon of fungi and algae was checked against literatures (広瀬・山岸 1977, 吉田ほか2015, National Museum of Nature and Science 2018, NIES 2022, Ohmura and Kashiwadani 2018, Suzuki 2016). YList (Yonekura and Kajita 2003–) was also referred to identify names for moss plants. Family names were mostly extracted from the checklist (Yamanouchi et al. 2019) and higher taxon was extracted from GBIF Backbone Taxonomy using GBIF Species API (GBIF 2023, GBIF Secretariat 2023) in Python programs.

Spelling variants in event dates, prefectures, quadrat areas and vegetation layers were fixed if obvious. Prefecture names were checked against the Digital Agency Registry Catalog (Digital Agency 2023). Vegetation layers were cleaned and normalized based on the outline of the 2nd National Survey on the Natural Environment (MoE 1979). All data including spatial and taxonomic information were organized into occurrence data using a MySQL database. Total of 413 occurrence records were deleted from original data, because either species names were invalid or relations between occurrences and events were incomplete.