https://doi.org/10.5061/dryad.hmgqnk9th

Description of the data and file structure

METHODOLOGICAL INFORMATION

Description of methods used for collection/generation of data:
The study was located in the Great Hungarian Plain in Eastern Hungary, Europe. We worked on 13 sites (11 kurgans and two reference flat grasslands) where Salvia nemorosa was present, spanning across 135 km (E - W) and 79 km (N - S). Kurgans were selected to represent different habitat area (basal area ranged between 528 - 4,321 m2) and connectivity (Hanski connectivity index ranged between 0 - 338.4). S. nemorosa grew in remnant grassland patches dominated by Festuca rupicola. 
We recorded site (GPS coordinates, date of visit, management, habitat type), and transect level (cardinal aspect and slope) information.
Data collection followed the protocol of Plantpopnet, a spatially distributed model system for population ecology (Buckley et al., 2019). Data was collected at peak flowering of S. nemorosa in June-July, over three consecutive years between 2021-2023. Eight sites were censused each year throughout the length of the study, further four sites were censused only twice as new sites were added to the study between 2022-2023, while one site was only censused once in 2021 due to subsequent census difficulties. We established up to four permanent transects of varying lengths (4-10 m), depending on the area of the habitat patch, where the number of individuals was representative of the habitat. On kurgans, S. nemorosa individuals were often limited to only one side of the kurgan (typically on South-Southwestern slopes or on the top), but on three kurgans a larger number of individuals allowed setting up transects on two contrasting slopes. Along each transect, we laid down permanent contiguous plots of 0.5 x 0.5 m, within which we permanently marked all S. nemorosa individuals with a numbered linoleum tag, including new individuals each consecutive year. For each plant, we recorded in-situ three vegetative traits: height of the tallest stem, number of stems, length and width of two biggest leaves forming a leaf pair, and two reproductive traits: inflorescence length and number of primary side inflorescence pairs. In this study, we narrowed our analyses only to mature plants, to capture maximum attainable growth responses that can be related to the reproductive effort of same individuals. We defined mature individuals as plants that had a flowering probability (i.e., the likelihood that a plant develops inflorescence) higher than 50%.

Methods for processing the data:
To calculate mean leaf area, we approximated the measured leaf shapes to a triangle, and we multiplied the leaf length and width, which we averaged across the two measured leaves.
To identify plants that had a flowering probability (i.e., the likelihood that a plant develops inflorescence) higher than 50%, we modelled the flowering state (yes or no) as a function of stem height, by fitting a Generalized Linear Mixed-effects Model (GLMM) with binomial error distribution in the lme4 package in R (Bates et al., 2015). Fixed effects were the tallest stem height in interaction with study year, while study site, transect, plot and plant were nested random effects. We tested the model assumptions using the DHARMa package (Hartig, 2022). We then used the generic function predict to obtain the flowering probability values for each individual, and individuals predicted a flowering probability lower than 50% were removed from further analyses.

We used the connectivity index by Hanski et al., (2000) to quantify the level of habitat isolation within a 300 m buffer around the focal habitat patch (a lower value indicating higher isolation. We set α (a species-specific parameter related to the species’ dispersal ability) and β (a parameter describing the scaling of immigration) to 5 based on the assumption of a relatively low dispersal of 200 m for S. nemorosa.
The area of the neighbouring grassland was calculated from digitised present day (2014-2016) habitat maps (source: Unified National Projection System, 1:10,000 topographic map of Hungary (Institute and Museum of Military History, Budapest), Satellite images (Google Maps) and field surveys, as in Deák et al., 2021).
We calculated the habitat area as the basal area of the kurgan or the area of the grassland patch using the same method as described above (Deák et al., 2021). To ease calculations due to the disproportionately large area of one of the reference grasslands compared to the area of the kurgans, we unified the Hanski connectivity and habitat area values for the two reference grasslands to the value of the grassland with the smaller area (232,061 m2).

To characterise the microclimate of the different transects we used the heat load index, a direct measure of incident solar radiation on a land surface that combines aspect (converted into degrees, slope, and latitude values (McCune & Keon, 2002), and which we calculated with a function by (Zelený & Lin, (2019) in R, using the first equation:
Heat load= e^((-1.467+1.582*cos(L)cos(S)-1.500cos(A)*sin(S)sin(L)-0.262sin(L)sin(S)+0.607sin(A)*sin(S))),
where L is the site latitude, A is the aspect where S. nemorosa populations where found, and S is the slope angle of the transect.

References:
Bates, D., Mächler, M., Bolker, B. M., & Walker, S. C. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1), 1–48. https://doi.org/10.18637/JSS.V067.I01
Hartig, F. (2022). DHARMa: Residual Diagnostics for Hierarchical (Multi-Level / Mixed) Regression Models.
Hanski, I., Alho, J., & Moilanen, A. (2000). Estimating the Parameters of Survival and Migration of Individuals in Metapopulations. Ecology, 81(1), 239 https://doi.org/10.2307/177147
Deák, B., Bede, Á., Rádai, Z., Tóthmérész, B., Török, P., Nagy D., D., Torma, A., Lőrinczi, G., Nagy, A., Mizser, S., Kelemen, A., & Valkó, O. (2021b). Different extinction debts among plants and arthropods after loss of grassland amount and connectivity. Biological Conservation, 264, 109372. https://doi.org/10.1016/j.biocon.2021.109372
McCune, B., & Keon, D. (2002). Equations for potential annual direct incident radiation and heat load. Journal of Vegetation Science, 13(4), 603–606. https://doi.org/10.1111/J.1654-1103.2002.TB02087.X
Zelený, D., & Lin, T.-Y. (2019). Heat load calculation: Analysis of community ecology data in R. https://www.davidzeleny.net/anadat-r/doku.php/en:customized_functions:heatload

Instrument- or software-specific information needed to interpret the data:
R Studio version: 2024.4.2.764
R Core Team (2024). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

Standards and calibration information, if appropriate: n/a

Environmental/experimental conditions: n/a

Files and variables

File: Salvia_2021_2024_final_dataset(Dryad).csv

Description: Trait records of marked Salvia nemorosa L. individuals censused annually between 2021-2023 in 13 localities of the Great Hungarian Plain

Variables

• c_year: census year;

• date_number: numbered survey dates in each year starting from the 1st of June;

• site_id: identity of the site

• microsite_id: identity of the microsite within a site

• trans_id: identity of the transect within a site

• plot_id: identity of the plot within a transect

• plant_id: identity of the plant within a transect

• stem_hei_noinf_allplants: height (mm) of the tallest stem of a plant, without inflorescence

• num_stems: number of stems in a plant

• mean_leafarea: mean leaf area of the fourth leaf pair counted from the stem node at soil level

• inf_len: length (mm) of the main inflorescence of the tallest stem

• inf_phe: phenological status of the plant (bud, flo: flowering, dev: seeds developing, mat: seeds matured, dis: seeds dispersed)

• Prim_side_inf_pairs: number of primary side inflorescence pairs

• heatload: heat load value calculated as described above

• habitat_area: habitat area calculated as described above

• Hanski: Hanski index calculated as described above

• log_heatload: log10 of heat load

• log_habitat_area: log10 of habitat area

  Missing data code: NA

Code/software

R Studio version: 2024.4.2.764

R Core Team (2024). _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

Microsoft Excel versions

R script