Wood trait–decay relationships vary with topography and rainfall seasonality in a subtropical forest in China
Data files
Jan 30, 2025 version files 646.96 KB
-
decomp.period.csv
66.02 KB
-
decomp.tre
5.91 KB
-
decomp1.csv
66.06 KB
-
decomp2.csv
61.01 KB
-
dem.tif
31.60 KB
-
density_spatial_variations.csv
275.96 KB
-
README.md
5.10 KB
-
sr.tif
67.60 KB
-
twi.tif
67.70 KB
Abstract
Plant traits affect wood decomposition. Far less acknowledged, wood trait-decay relationships may vary with environmental conditions that alter the spatiotemporal distribution of decomposer activities. Unfavorable environments can suppress decomposer activities and weaken the wood trait-decay relationships. We hypothesize that waterlogged soils in the valley during the rainy season may decrease termite activities and decomposition rates, especially of palatable deadwood. By comparison, wood trait-decay relationships could be less affected at the ridge and hilltop during the rainy season. We conducted wood decay experiments of 137 tree species in three sites (valley/ridge/hilltop) differing in microclimate in a 50-ha plot in a subtropical forest. Eight physical and chemical traits were measured to determine the most influential wood properties for termite feeding activities and wood decomposition rates. Samples were harvested after the rainy or after a whole year (rainy-plus-dry season). Mass loss and volume loss (i.e. termite feeding activities) of the first harvest were attributed to the rainy season, while the difference between the two harvests was attributed to the dry season. We found that termite feeding activities and wood mass loss during the rainy season were decreased by soil waterlogging in the valley but not on the ridge or hilltop. Wood density predominantly controlled termite feeding and wood mass loss, but its negative effects appeared at the ridge and hilltop only during the rainy season, and in the valley only during the dry season. Notably, warmer soil temperatures in the valley accelerated termite activities during the dry season, causing similar final wood mass loss as higher-elevation sites.
Synthesis. Our study reveals that environmental effects on wood decomposition are not only mediated by trait changes directly but also via the spatiotemporal variations of wood trait-decay relationships indirectly. Moreover, wood decomposition does not strictly follow the elevational temperature gradients due to the negative effects of rainfall at lower elevations. Notably, future droughts may reduce soil waterlogging and accelerate termite activities and the decomposition of lower-density wood, particularly in valleys. We are concerned that deadwood carbon stocks once waterlogged by rainfall could become a C source under future droughts, especially in regions where termites are dominant decomposers.
README: Higher-order environmental effects on wood trait-decay relationships
https://doi.org/10.5061/dryad.jdfn2z3hq
Description of the data and file structure
Eight wood traits, wood mass loss, the absence/presence of termite feeding (0/1), and proportional termite feeding per each tree species (n = 137) are given for each topographic site (valley/ridge/hilltop) and each season (rainy vs. dry). Three data files are related. 'decomp.period.csv' presents the period change of wood mass loss and proportional termite feeding across two seasons. 'decomp1.csv' presents the wood mass loss and the absence/presence of termite feeding after the first rainy season. 'decomp2.csv' presents the wood mass loss and the absence/presence of termite feeding after the rainy and subsequent dry seasons. Note that only 124 species are successfully retrieved after the dry season, and thus we only present the data for 124 species in 'decomp.period.csv' and 'decomp2.csv'. To verify if topography strongly changes the distribution of wood density, we compute the functional dispersion and average values of wood density per 10 × 10 m grid tree communities ('density spatial variations.csv'). To extract the three metrics related to topography, we provide the raw rasters for elevation ('dem.tif'), annual cumulative solar radiation ('sr.tif') and topographic wetness index ('twi.tif'). The phylogenetic tree of 137 species is also provided ('decomp.tre').
Code
All statistical analyses were carried out in R 4.2.3 (R Core Team 2021). First, we conduct phylogenetic generalized linear mixed models to test how the absence/presence of termite feeding and wood mass loss respond to eight wood traits to determine the most influential wood trait(s). Second, we use generalized linear models to test if wood trait effects on the period change of wood mass loss and proportion termite feeding vary with topographic position and seasons. Third, we use simple linear models to test if topography strongly changes the spatial distribution of wood traits that significantly alter termite feeding and wood mass loss. All code with step--by-step instructions can be found in the code file ('Code.R').
Description of the variables
decomp1.csv & decomp2.csv
- tag: ID per wood sample
- species: scientific name per wood sample
- site: topographic position (valley/ridge/hilltop)
- elevation: elevation above sea level (m)
- diameter: mean diameter per wood sample (cm)
- bark.ratio: ratio of bark radial width per species (unitless)
- wood.density: dry mass per volume per species (g/cm3)
- dry.matter: ratio of dry mass to fresh mass per species (unitless)
- c: carbon content per species (%)
- n.c: nitrogen-to-carbon ratio per species (unitless)
- lignin: lignin content per species (%)
- c.l: cellulose-to-lignin ratio per species (unitless)
- abundance: total number of individuals per species in the 50-ha plot
- proportional.termite.feeding: ratio of total lost volume to initial volume per wood sample
- termite01: presence (1) or absence (0) of termite feeding per wood sample
- termitenum: observed number of termite individuals per wood sample
- mass.loss: one minus the ratio of final dry mass to initial dry mass per wood sample
decomp.period.csv
- species: scientific name per wood sample
- site: topographic position (valley/ridge/hilltop)
elevation: elevation above sea level (m)
bark.ratio: ratio of bark radial width per species (unitless)
wood.density: dry mass per volume per species (g/cm3)
dry.matter: ratio of dry mass to fresh mass per species (unitless)
c: carbon content per species (%)
n.c: nitrogen-to-carbon ratio per species (unitless)
lignin: lignin content per species (%)
c.l: cellulose-to-lignin ratio per species (unitless)
abundance: total number of individuals per species in the 50-ha plot
termite.feeding.rainy: proportional wood volume loss due to termite feeding during the rainy season (unitless)
termite.feeding.dry: proportional wood volume loss due to termite feeding during the subsequent dry season (unitless)
termite.feeding.all: proportional wood volume loss due to termite feeding during the whole period (rainy-plus-dry)
mass.loss.rainy: proportional wood dry mass loss due to termite feeding during the rainy season (unitless)
mass.loss.dry: proportional wood dry mass loss due to termite feeding during the subsequent dry season (unitless)
mass.loss.all: proportional wood dry mass loss due to termite feeding during the whole period (rainy-plus-dry)
density spatial variations.csv
- plotid_10×10: ID of subplot (10 m × 10 m)
- latitude: latitude of subplot centroid
- longitude: longitude of subplot centroid
- density_mean: mean value of wood density across the subplot weighted by the total basal area per tree species (g/cm3)
- density_fdis: functional dispersion of wood density across the subplot weighted by the total basal area per tree species (g/cm3)