Dataset DOI: 10.5061/dryad.bnzs7h4qv

Description of the data and file structure

Trait and demographic (size, survival) data were collected for 15 temperate tree species, from published, online data sources accessed on May 2015 (demographic data) and January 2021 (trait data).

• FIA_dat.csv. This file contains a geographical (northeastern United states) and species subset (15 temperate tree species) of the original FIA dataset (see Methods), in which only relevant columns were retained with information on tree species, plot locations, state, tree identity code, census year, DBH and survival records. Each row (n = 736,087) represents a single census measurement of DBH and survival status (i.e., dead or alive) of a given tree, from a given species, in a given FIA plot, censused approximately every five years between 2000 and 2014 during which each tree was sampled 1 – 4 times.
• FIA_spp_traits.csv. This file contains the species mean trait data obtained from existing, online dataset (i.e., TRY; FRED; see Methods in main article), the leaf habit and mycorrhizal type of each species, as well as species’ mean growth (in favorable environments, i.e., AGR₉₅) and survival (in unfavorable environments, SURV₂₅) rates. Each row (n = 15) represents a single species.
• FIA_demo.csv. In this file, we cleaned the FIA data, and merged it with the species-level trait datasets. We calculated the absolute growth rate (AGR) per tree, and added spatial variables in order to pool nearby FIA census plots in a single 50 x 50 grid cell (see Methods in main article). Each row (n = 735,683) represents a single census measurement of DBH, AGR and survival status (i.e., dead or alive) of a given tree, from a given species, in a given FIA plot and 50 x 50 km grid cell. This dataset was used for our further statistical analyses.

Files and variables

File: FIA_spp_traits.csv

Description: Trait values per species obtained from previous datasets (TRY, FRED, Chave et al. 2009 – see references in Methods of main article), and species-averaged growth and survival (i.e., AGR95, SURV25; see below). Each row represents a single species.

Variables

• Species: Species (Latin) name
• Leaf_habit: Leaf habit per tree with two categories: angio, angiosperm; gymno, gymnosperm.
• MF: Mycorrhizal fungal association per tree with two categories: AM, arbuscular mycorrhizal; EcM, ectomycorrhizal.
• RD: Root diameter (mm)
• RN: Root nitrogen (mg g-1)
• RTD: Root tissue density (g cm-3)
• SRL: Specific root length (m g-1)
• LN: Leaf nitrogen (mg g-1)
• SLA: Specific leaf area (mm2 mg-1)
• WD: Wood density (g cm-3)
• AGR: Absolute growth rate (cm y-1)
• SURV: Survival status (0 = dead; 1 = alive)
• AGR95: Absolute growth rate of the % fastest-growing trees per species (cm y-1)
• SURV25: Survival probability of the 25% slowest-growing trees per species

File: FIA_dat.csv

Description: Census data from the FIA program measured per tree, per census. Each row represents a single census measurement. Trees measured at multiple censuses are thus represented in multiple rows (and can be identified by the variable ‘ID’, see below).

Variables

·         Species: Species (Latin) name

·         PLOT: Unique identifier of the FIA plot in which a tree was censused

·         SUBP: Unique identifier of the FIA sub-plot in which a tree was censused

·         LAT: Latitude degrees of the FIA plot center

·         LON: Longitude degrees of the FIA plot center

·         STATE: Unique 2-digit code for the state in which a plot was located

·         STATEAB: Abbreviated state name

·         TREE: Tree number assigned by FIA

·         ID: Unique identifier of each censused tree

·         YEAR: Year in which a tree was censused

·         STATUS: Tree survival status at census, with 2 categories: 0, dead; 1, alive

·         DBH: Diameter at breast height of tree at census (cm)

File: FIA_tree_demo.csv

Description: This dataset contains the data from FIA_dat.csv, to which new variables were added (listed below) related to species traits, demographic rates, spatial grid system. Trees that had shrunk more 75% in their DBH between censuses were removed. Each row represents a single census measurement. Trees measured at multiple censuses are thus represented in multiple rows (and can be identified by the variable ‘ID’, see below).

Variables

• Species: Species (Latin) name
• Leaf_habit: Leaf habit per tree with two categories: angio, angiosperm; gymno, gymnosperm.
• MF: Mycorrhizal fungal association per tree with two categories: AM, arbuscular mycorrhizal; EcM, ectomycorrhizal.
• RD: Root diameter (mm)
• RN: Root nitrogen (mg g-1)
• RTD: Root tissue density (g cm-3)
• SRL: Specific root length (m g-1)
• LN: Leaf nitrogen (mg g-1)
• SLA: Specific leaf area (mm2 mg-1)
• WD: Wood density (g cm-3)
• PLOT: Unique identifier of the FIA plot in which a tree was censused
• SUBP: Unique identifier of the FIA sub-plot in which a tree was censused
• LAT: Latitude degrees of the FIA plot center
• LON: Longitude degrees of the FIA plot center
• STATE: Unique 2-digit code for the state in which a plot was located
• STATEAB: Abbreviated state name
• TREE: Tree number assigned by FIA
• ID: Unique identifier of each censused tree
• YEAR: Year in which a tree was censused
• STATUS: Tree survival status at census (0, dead; 1, alive)
• DBH: Diameter at breast height of tree at census (cm)
• location: unique identified or a spatial grid cell to which PLOTs were assigned (see Methods in main article)
• long2: X coordinate of the grid cell (‘location’)
• lat2: Y coordinate of the grid cell (‘location’)
• grain: grain of grid system (here: 50 x 50 km)
• PYEAR: Previous census year
• PDBH: Tree DBH at previous census (cm)
• AGR: Absolute growth rate of a tree based on difference in DBH between current and previous census (cm y-1)
• SURV: Survival status (= STATUS)
• PDBHQ: 0.75 quantile of PDBH used to filter trees that had shrunk more than 75% in DBH between censuses

File: FIA_sub_fast.csv

Description: Subset of FIA_tree_demo.csv in which only the 5%-fastest growing trees per species were selected. We added new variables (listed below) that were used to run our Trait-Growth models. Each row represents a single census measurement. Trees measured at multiple censuses are thus represented in multiple rows (and can be identified by the variable ‘ID’, see below).

Variables

• Species: Species (Latin) name
• Leaf_habit: Leaf habit per tree with two categories: angio, angiosperm; gymno, gymnosperm.
• MF: Mycorrhizal fungal association per tree with two categories: AM, arbuscular mycorrhizal; EcM, ectomycorrhizal.
• RD: Species root diameter (mm)
• RN: Species root nitrogen concentration (mg g-1)
• RTD: Species root tissue density (g cm-3)
• SRL: Species specific root length (m g-1)
• LN: Species leaf nitrogen concentration (mg g-1)
• SLA: Species specific leaf area (mm2 mg-1)
• WD: Species wood density (g cm-3)
• PLOT: Unique identifier of the FIA plot in which a tree was censused
• SUBP: Unique identifier of the FIA sub-plot in which a tree was censused
• LAT: Latitude degrees of the FIA plot center
• LON: Longitude degrees of the FIA plot center
• STATE: Unique 2-digit code for the state in which a plot was located
• STATEAB: Abbreviated state name
• TREE: Tree number assigned by FIA
• ID: Unique identifier of each censused tree
• YEAR: Year in which a tree was censused
• STATUS: Tree survival status at census (0, dead; 1, alive)
• DBH: Diameter at breast height of tree at census (cm)
• location: unique identified or a spatial grid cell to which PLOTs were assigned
• long2: X coordinate of the grid cell (‘location’)
• lat2: Y coordinate of the grid cell (‘location’)
• grain: grain of grid system
• PYEAR: Previous year at which a tree was censuses
• PDBH: Tree DBH at previous census (cm)
• AGR95: Absolute growth rate of a tree between current and previous census for only the 5%-fastest growing trees per species (cm y-1).
• SURV: Survival status (= STATUS)
• PDBHQ: 0.75 quantile of PDBH used to filter trees that had shrunk more than 75% in DBH between censuses (cm)
• logAGR95: Log-transformed AGR (see above) per tree.
• logAGR95.sc: Scaled and log-transformed AGR (see above) per tree.
• logDBH.sc: Scaled and log-transformed DBH (see above) per tree.
• RD.sc: Scaled species root diameter
• RN.sc: Scaled species root nitrogen concentration
• SRL.sc: Scaled species specific root length
• LN.sc: Scaled species leaf nitrogen concentration
• SLA.sc: Scaled species specific leaf area
• WD.sc: Scaled species wood density
• RTD.sc: Scaled species root tissue density

File: FIA_sub_slow.csv

Description: Subset of FIA_tree_demo.csv in which only the 25%-slowest growing trees per species were selected. We added new variables (listed below) that were used to run our Trait-Survival models. Each row represents a single census measurement. Trees measured at multiple censuses are thus represented in multiple rows (and can be identified by the variable ‘ID’, see below).

Variables

• Species: Species (Latin) name
• Leaf_habit: Leaf habit per tree with two categories: angio, angiosperm; gymno, gymnosperm.
• MF: Mycorrhizal fungal association per tree with two categories: AM, arbuscular mycorrhizal; EcM, ectomycorrhizal.
• RD: Species root diameter
• RN: Species root nitrogen concentration
• RTD: Species root tissue density
• SRL: Species specific root length
• LN: Species leaf nitrogen concentration
• SLA: Species specific leaf area
• WD: Species wood density
• PLOT: Unique identifier of the FIA plot in which a tree was censused
• SUBP: Unique identifier of the FIA sub-plot in which a tree was censused
• LAT: Latitude degrees of the FIA plot center
• LON: Longitude degrees of the FIA plot center
• STATE: Unique 2-digit code for the state in which a plot was located
• STATEAB: Abbreviated state name
• TREE: Tree number assigned by FIA
• ID: Unique identifier of each censused tree
• census1: Previous census year (= YEAR in FIA_tree_demo.csv)
• STATUS: Tree survival status at census (0, dead; 1, alive)
• DBH1: Diameter at breast height of tree at previous census (i.e. ‘census1’)
• location: unique identified or a spatial grid cell to which PLOTs were assigned
• long2: X coordinate of the grid cell (‘location’)
• lat2: Y coordinate of the grid cell (‘location’)
• grain: grain of grid system
• PYEAR: Previous year at which a tree was censuses
• PDBH: Tree DBH at previous census (cm)
• AGR: Absolute growth rate of a tree (cm y-1)
• PDBHQ: 0.75 quantile of PDBH used to filter trees that had shrunk more than 75% in DBH between censuses
• RD.sc: Scaled species root diameter
• RN.sc: Scaled species root nitrogen concentration
• SRL.sc: Scaled species specific root length
• LN.sc: Scaled species leaf nitrogen concentration
• SLA.sc: Scaled species specific leaf area
• WD.sc: Scaled species wood density
• RTD.sc: Scaled species root tissue density
• census2: Current census year
• DBH2: DBH (see above) at ‘census2’
• surv: Survival status at current census (‘census2’)
• time: Number of years between censuses (‘census1’ and ‘census2’)
• d.DBH: Difference between current and previous DBH (‘DBH1’ and ‘DBH2’)
• d.surv: Difference in survival status
• logDBH1.sc: Scaled and log-transformed ‘DBH1’
• logDBH2.sc: Scaled and log-transformed ‘DBH2’
• DBH1.sc: Scaled ‘DBH1’

Code/software

Open software used: Data were cleaned, calculated and analysed ysing R Statistical Software (R Core Team. (2024). R: A Language and Environment for Statistical Computing; version 4.4.2). The Directed Acyclic Diagram (DAG) was constructed in DAGitty open software (Textor et al. 2016).

Pcakages used: tidyverse, brms, tidybayes, loo, Hmisc, ggridges, ggplot2, gridExtra, modelr, rstan, ggpubr, dplyr, dagitty, lavaan, data.table, lubridate, sf, rFIA, rnaturalearth, rnaturalearthdata

Workflow:

• FIA_dat.csv is the original but cleaned data from the FIA program, obtained at: XXX
• FIA_tree_demo.csv is a cleaned subset of FIA_dat.csv with added calculated variables, incorporates FIA_spp_traits.csv for each individual tree measurement.
• FIA_spp_traits.csv contains trait data from aforementioned existing datasets, and (demographic) variables calculated from FIA_tree_demo.csv, i.e., AGR and SURV averaged per species.
• FIA_sub_fast.csv and FIA_sub_slow.csv are subsets of FIA_tree_demo.csv with the 95% fastest-growing trees per species, and the 25% slowest-growing trees per species, respectively, and additional variables for running our statistical models.

Code:

Weemstra_et_al_RCode.R includes code for:

• Data cleaning, calculations, subsetting and merging of species- and individual level trait and demographic datasets.
• Defining functions
• Testing DAG-data consistency
• Running statistical tests, i.e.:
  • pairwise and multivariate correlations
  • linear regression models
  • t tests
  • Bayesian demographic models

• Produce the figures in the main article.

Access information

Other publicly accessible locations of the data:

• N/A

Data was derived from the following sources:

• Tree demographic data (DBH, survival) from the USDA Forest Inventory and Analysis (FIA) program: U.S. Department of Agriculture. Forest Service, “Forest Inventory Analysis”, https://research.fs.usda.gov/programs/fia

  Leaf trait data from the TRY Plant Trait Database: https://www.try-db.org/TryWeb/Home.php. Kattge J, Bönisch G, Díaz S, et al. TRY plant trait database – enhanced coverage and open access. Glob Change Biol. 2020; 26: 119–188. https://doi.org/10.1111/gcb.14904 

• Root trait data from the Fine-root Ecological Database (FRED): https://roots.ornl.gov/. Iversen CM, et al. 2021. Fine-Root Ecology Database (FRED): A Global Collection of Root Trait Data with Coincident Site, Vegetation, Edaphic, and Climatic Data, Version 3. Oak Ridge National Laboratory, TES SFA, U.S. Department of Energy, Oak Ridge, Tennessee, U.S.A. Access on-line at: https://doi.org/10.25581/ornlsfa.014/1459186.

• Wood trait data from: Chave, J., Coomes, D., Jansen, S., Lewis, S.L., Swenson, N.G. & Zanne, A.E. (2009). Towards a worldwide wood economics spectrum. Ecology Letters, 12, 351–366. https://onlinelibrary.wiley.com/doi/10.1111/j.1461-0248.2009.01285.x