Data for: Historical floras reflect broad shifts in flowering phenology in response to a warming climate A.G. Auffret 2021 Contact for further information and collaboration: Alistair Auffret alistair.auffret@slu.se This is the data set used for the publication "Historical floras reflect broad shifts in flowering phenology in response to a warming climate" published in Ecosphere. The dataset is based on the months of flowering given in 11 guides to the Swedish flora over time. In this file, I briefly describe the components of the data set. ANALYSED DATA phenology_data.csv This is one of the tables analysed for the results published in the paper. Each row is an 'observation', i.e. the months of flowering given for a particular species in a particular flora. Columns describe information relating to the flora, as well as other variables used in the analysis. Flora: A code representing each flora, made up of three letters of the main author's surname and the year of publication. Auth: Three letter code describing the main author's surname. Year: Year of publication. Species: Binomial scientific name of the species. Taxonomy according to the Swedish Taxonomic Database Dyntaxa. Group: Grouping of the species, can be one of 199 randomly-selected widespread species ('Random') or one of 42 'Phenology' species with species names describing flowering time, or that the species is included in the Swedish phenology project Naturens kalender (http://www.naturenskalender.se). Start: The month number (1-12) of flowering start given in the flora. Stop: The month numner (1-12) of flowering cessation given in the flora. If only one month is given, Start and Stop have the same number, unless the flora states 'Flowering starts in ...', in which case Stop is NA. Length: The length of flowering, calculated as Stop - Start + 1 Season: Ordinal scale 1-3. 1 = early: flowering starts in May or earlier, 64 species; 2 = mid: flowering starts June or July, 170 species; 3 = late: flowering starts in August or later, 7 species). Information for this variable is taken from the first flora in which each species is listed. Temp: Tthe annual temperature anomaly from CRUTEM4 (https://crudata.uea.ac.uk/cru/data/crutem/ge/) across the nine 5-degree grid squares covering Sweden for the 30 years up to and including the year of the flora’s publication. Only available for floras published from 1901 and onwards. HerbCum: The cumulative number of herbarium records in Sweden at the year of publication of the flora, including all species. Data are from Sweden's Virtual Herbarium http://herbarium.emg.umu.se/. HerbCumSpe: The cumulative number of herbarium records in Sweden at the year of publication of the flora, for the particular species. phenology_species_models_time.csv This shows the outputs of the species-level models describing phenological change over time. These outputs were later used to analyse the effect of flowering seasonality on shifts in flowering phenology and are therefore provided here. Species: Binomial scientific name of the species. Taxonomy according to the Swedish Taxonomic Database Dyntaxa. Group: Grouping of the species, can be one of 199 randomly-selected widespread species ('Random') or one of 42 'Phenology' species with species names describing flowering time, or that the species is included in the Swedish phenology project Naturens kalender (http://www.naturenskalender.se). Season: Ordinal scale 1-3. 1 = early: flowering starts in May or earlier, 556 species; 2 = mid: flowering starts June or July, 1502 species; 3 = late: flowering starts in August or later, 63 species). Information for this variable is taken from the first flora in which each species is listed. Remaining columns describe outputs for linear models describing the effect of Year on flowering phenology (Start, Stop or Length), with the residual effect of increasing herbarium records per species, i.e. the residuals of a linear model describing the increase of herbarium records HerbCumSpe over time. Columns are thus named as follows, with Start as an example. Start.Lin.Est: Parameter estimate of the effect of Year on flowering start. Start.Lin.P: P-value of above parameter estimate. Start.Lin.Est.Samp: Parameter estimate of residual effect of increasing herbarium records on flowering start. Start.Lin.P.Samp: P-value of above parameter estimate. Start.Lin.mod.R2: R-squared value of model. Start.Lin.mod.P: P-value of model. NA values in this table can occur for species that are mentioned in fewer than five floras (NAs across entire row), or where parameter estimates are zero because of no change in flowering phenology over time (NAs in affected parameter estimate cells). phenology_species_models_temp.csv This shows the outputs of the species-level models describing phenological change with climate change. These outputs were later used to analyse the effect of flowering seasonality on shifts in flowering phenology and are therefore provided here. Species: Binomial scientific name of the species. Taxonomy according to the Swedish Taxonomic Database Dyntaxa. Group: Grouping of the species, can be one of 199 randomly-selected widespread species ('Random') or one of 42 'Phenology' species with species names describing flowering time, or that the species is included in the Swedish phenology project Naturens kalender (http://www.naturenskalender.se). Season: Ordinal scale 1-3. 1 = early: flowering starts in May or earlier, 556 species; 2 = mid: flowering starts June or July, 1502 species; 3 = late: flowering starts in August or later, 63 species). Information for this variable is taken from the first flora in which each species is listed. Remaining columns describe outputs for linear models describing the effect of Temp on flowering phenology (Start, Stop or Length), with the residual effect of increasing herbarium records per species, i.e. the residuals of a linear model describing the increase of herbarium records HerbCumSpe over time. Columns are thus named as follows, with Start as an example. Start.Lin.Est: Parameter estimate of the effect of Temp on flowering start. Start.Lin.P: P-value of above parameter estimate. Start.Lin.Est.Samp: Parameter estimate of residual effect of increasing herbarium records on flowering start. Start.Lin.P.Samp: P-value of above parameter estimate. Start.Lin.mod.R2: R-squared value of model. Start.Lin.mod.P: P-value of model. FULL REFERENCES OF FLORAS USED LIL1798: Liljeblad, S. 1798. Utkast til en svensk flora : eller, Afhandling om svenska växternas väsendteliga kännetecken och nytta. Second edition. J. F. Edman, Uppsala. HAR1820: Hartman, C. J. 1820. Handbok i Skandinaviens flora: innefattande sveriges och norges växter, till och med mossorna. Z. Haeggström, Stockholm. NYM1867: Nyman, K. F. 1867. Utkast til svenska växternas naturhistoria. N. M. Lindh, Örebro. NEU1901: Neuman, L. M., and F. E. Ahlfvengren. 1901. Sveriges flora (fanerogamerna). C. W. K. Gleerup, Lund. LIN1918: Lindman, C. A. M. 1918. Svensk fanerogamflora. Norstedts, Stockholm. URS1944: Ursing, B. 1944. Svenska växter i text och bild. Nordisk rotogravyr, Stockholm. BOL1950: Bolin, L. W., and L. O. A. von Post. 1950. Floran i färg. H. Geber, Stockholm. HUL1958: Hultén, E., and H. Anthon. 1958. Vår svenska flora i färg. Svensk litteratur, Stockholm. MOS1992: Mossberg, B., L. Stenberg, and S. Ericsson. 1992. Den nordiska floran. Wahlström & Widstrand, Stockholm. AND1997: Den virtuella floran (website), http://linnaeus.nrm.se/flora/ MOS2018: Mossberg, B., and L. Stenberg. 2018. Nordens flora. Bonnier Fakta. SUPPLEMENTARY APPENDIX I also include Appendix.pdf, which is an appendix to the published paper.