Published September 4, 2026 | Version 1.0.0

Foliar and wood nutrient stoichiometry of sugar maple (Acer saccharum) after repeated summer throughfall exclusion

  • 1. ROR icon Université du Québec en Abitibi-Témiscamingue
  • 2. ROR icon Université de Montpellier
  • 3. ROR icon Université du Québec en Outaouais

Description

This dataset holds the nutrient composition of the foliage and the wood of 36 sugar maple (Acer saccharum Marshall) individuals sampled in a summer throughfall exclusion experiment, together with the plot-level soil ion supply rates and soil pH measured in the same plots. It supports the analysis of drought legacy effects on nutrient stoichiometry, expressed as isometric log-ratio (ilr) balances and as Compositional Nutrient Diagnosis (CND) indices. Raw concentrations and supply rates are provided alongside every derived coordinate, so all compositional transformations can be recomputed from the source measurements.

DESIGN

Randomised complete block design with three blocks (Site 5, 6 and 7) and two plots per block, one under summer throughfall exclusion and one control. Within each plot, three mature trees and two to five saplings were sampled, for 18 mature trees and 18 saplings in total. Each individual yields one foliar sample and one wood sample. Soil variables were measured once per plot and are repeated across all individuals of that plot.

FILE

One Excel file, one sheet, 36 rows, 56 columns. Missing values are empty cells and are read as NA in R. Only DBH has any, four in total. No value has been imputed.

COLUMNS: IDENTIFIERS AND DESIGN

Tree_name: unique individual code, built as site + plot number + A (mature tree) or G (sapling) + Num. Example: 52A735 is tree 735 in plot 5_2. This is the only unique key in the table.

Site: experimental block, 5, 6 or 7.

Plot: plot identifier (Site_Treatment), six levels, the unit to which the treatment was applied.

Treatment: 1 = throughfall exclusion, 2 = control.

Stage: Mature_tree or sapling.

Num: field tag number of the individual, unique within a plot and stage but not across the dataset.

DBH: diameter at breast height (1.3 m), cm. Mature trees 18.0 to 39.0, saplings 4.3 to 10.0.

N_Arbres, N_Gaules: number of mature trees (3) and of saplings (2 to 5) sampled in the plot.

COLUMNS: TISSUE CONCENTRATIONS

Expressed in mg per g of oven-dry mass.

Foliage: n_tot_F (total N), P_F, K_F, Ca_F, Mg_F, Mn_F, Al_F.

Wood: n_tot_B (total N), P_B, K_B, Ca_B, Mg_B, Mn_B, Al_B.

Aluminium was measured but is not part of the compositions analysed. It is retained here for reuse.

COLUMNS: SOIL, PLOT LEVEL

Each of the following holds one value per plot, repeated across the individuals of that plot.

pH_Organic, pH_Mineral: pH of the organic and of the mineral horizon.

Burial_Date: date of burial of the ion exchange membranes, 2025-07-04 in all plots.

NO3-N_soil, NH4-N_soil, P_soil, K_soil, Ca_soil, Mg_soil, Mn_soil, Al: ion supply rates measured with PRS anion and cation exchange membranes (Western Ag Innovations, Saskatoon, Canada), in micrograms per 10 square centimetres per burial period. These are supply rates, that is the quantity of each ion adsorbed by the membrane over the burial period. They are neither concentrations nor extractable pools and cannot be converted into either. They integrate concentration, diffusion and mass flow, and therefore depend on soil moisture and temperature during the burial period.

N_soil: total inorganic N, the sum of NO3-N_soil and NH4-N_soil. This is the nitrogen used in the soil composition.

COLUMNS: ILR BALANCES

Fifteen columns, suffix _F for the foliage, _B for the wood, _S for the soil. Each set is a complete orthonormal basis of a six-part composition (N, P, K, Ca, Mg, Mn) and therefore holds five coordinates, with no loss of information.

The three sets are built on two different kinds of measurement. The foliar and wood coordinates are computed from tissue concentrations (n_tot, P, K, Ca, Mg and Mn of the corresponding compartment, in mg per g of oven-dry mass). The soil coordinates are computed from the PRS supply rates (N_soil, P_soil, K_soil, Ca_soil, Mg_soil and Mn_soil, in micrograms per 10 square centimetres per burial period). The ilr transformation is scale invariant, so both are valid compositions, but a soil coordinate describes the relative supply of two groups of ions to an exchange membrane, not their relative concentration in the soil. The two kinds of coordinate are comparable in structure but not in nature, and a difference in level between a soil coordinate and a tissue coordinate has no interpretation.

Sequential binary partition: [NPK|CaMgMn], [NP|K], [N|P], [CaMg|Mn], [Ca|Mg].

Each coordinate is sqrt(r*s/(r+s)) * ln(g(numerator group) / g(denominator group)), where r and s are the numbers of parts in each group and g is the geometric mean of the corresponding parts. A positive value means the numerator group is enriched relative to the denominator group.

COLUMNS: CND INDICES

Five columns computed on the foliage only, each the corresponding foliar ilr coordinate standardised against a reference subpopulation (Parent and Dafir 1992): CND_i = (ilr_F,i - mu_i) / sigma_i. A value of zero places the individual at the reference centroid; a value far from zero, in either direction, indicates an imbalance.

Reference values, to four decimals:

CND_NPK_vs_CaMgMn, from ilr_NPK_CaMgMn_F: mu = 0.7711, sigma = 0.3030.
CND_NP_vs_K, from ilr_NP_K_F: mu = -0.3127, sigma = 0.1476.
CND_N_vs_P, from ilr_N_P_F: mu = 1.7027, sigma = 0.1004.
CND_CaMg_vs_Mn, from ilr_CaMg_Mn_F: mu = 1.4409, sigma = 0.4998.
CND_Ca_vs_Mg, from ilr_Ca_Mg_F: mu = 1.4780, sigma = 0.1290.

CND_G is the global imbalance index, the Euclidean norm of the five CND indices.

CAVEATS FOR REUSE

The soil variables are plot level. Sixteen columns (pH_Organic, pH_Mineral, the nine ion supply columns, and the five soil ilr coordinates) hold a single value per plot, repeated across the individuals of that plot. Treating them as individual-level observations inflates the effective sample size.

The treatment was applied at the plot level, with three blocks and one plot per treatment per block. Any test of a treatment effect that uses the 36 individuals as replicates is pseudoreplicated.

The compositions are closed. ilr coordinates and CND indices describe relative, not absolute, nutrition. A change in a balance can arise from a change in either group of elements, and should be interpreted in the ratio rather than in a single element.

Soil water content, relative extractable water and the weather records are not part of this file.

Files

Files (28.1 kB)

Name Size
md5:944c2ca78e91d371df04fd8a0c0f4d4f
28.1 kB Download

Additional details

Funding

Ministère des Ressources naturelles et des Forêts du Québec
#3323-2225-REX-2206-8
Natural Sciences and Engineering Research Council of Canada
ALLRP 588587-23