Effects of Soil Amendments on Carbon-Based Indicators of Ecosystem Recovery in Revegetated Mine Soils: Integrating Labile Carbon, Soil Organic Carbon, and CO₂ Flux
Description
This dataset contains the soil carbon and respiration measurements underlying the study of carbon-based indicators of ecosystem recovery in revegetated mine soils at Sweetwater Mine, a lead–zinc tailings site in Reynolds County, Missouri, USA. Five soil-amendment treatments were evaluated: Untreated control; Lime + Manure (L+M); Lime + Manure + Woodchips (L+M+W); Lime + Manure + Biochar (L+M+B); and Lime + Manure + Biochar + Arbuscular Mycorrhizal Fungi (L+M+B+AMF), arranged across 15 revegetation plots. Three response variables are provided: (1) in-situ soil CO₂ flux, measured with a LI-8100A soil gas flux system across 45 collars over a 17-week monitoring period (14 successful measurement events; 630 observations); (2) permanganate-oxidizable carbon (POXC), a labile carbon fraction determined by the Weil method; and (3) soil organic carbon (SOC). POXC and SOC were measured on 45 end-of-season composite soil samples (three subsamples per plot). Accompanying covariates recorded during flux measurement include soil temperature, soil moisture (sensor voltage), atmospheric pressure, and week of measurement. These data support the estimated marginal means, percent-change comparisons, correlation analyses, and AR(1) linear mixed-effects models reported in the associated manuscript.
Files
Steps to reproduce
Study site and design Data were collected at Sweetwater Mine, a lead–zinc tailings site in Reynolds County, Missouri, USA. Five soil-amendment treatments were established across 15 revegetation plots (three plots per treatment): Untreated control; Lime + Manure (L+M); Lime + Manure + Woodchips (L+M+W); Lime + Manure + Biochar (L+M+B); and Lime + Manure + Biochar + Arbuscular Mycorrhizal Fungi (L+M+B+AMF). Soil CO₂ flux In-situ soil CO₂ flux was measured with a LI-COR LI-8100A automated soil gas flux system fitted to soil collars installed in each plot (45 collars total). Measurements were taken weekly over a 17-week monitoring period, yielding 14 successful measurement events (630 observations) after excluding events lost to equipment and weather conditions. Soil temperature, soil moisture (reported as sensor voltage), atmospheric pressure, and week of measurement were recorded concurrently as covariates. Soil sampling and preparation Soil samples were collected from each of the 45 collar locations (three collars per plot) to a depth of 0–15 cm using a stainless-steel T-handle soil probe (1.9 cm core diameter; Seedburo Equipment Co., Des Plaines, IL, USA). At each collar, multiple adjacent cores were combined into a single composite sample, yielding 45 composite samples (nine per treatment, three per plot). Samples were placed in labeled polyethylene bags, transported to the laboratory in a cooler, then air-dried at room temperature, gently ground, and passed through a 2 mm sieve prior to analysis. Soil organic carbon (SOC) SOC was determined by the Walkley–Black wet oxidation method (Nelson and Sommers, 1996; Walkley and Black, 1934). Soil organic carbon was oxidized by adding potassium dichromate and concentrated H₂SO₄; the excess dichromate was then back-titrated with ferrous sulfate using ferroin as an indicator. Results were reported as g kg⁻¹. Permanganate-oxidizable carbon (POXC) POXC, a labile soil carbon fraction, was determined by the potassium permanganate oxidation method (Culman et al., 2012; Weil et al., 2003). Air-dried soil was reacted with 0.02 M KMnO₄ in 0.1 M CaCl₂, and the absorbance of the diluted supernatant was measured at 550 nm. POXC was quantified from the KMnO₄ consumed against a calibration curve (R² ≥ 0.995), with reagent blanks and 10% duplicate samples (CV < 10%) included for quality control. Results were reported as mg kg⁻¹. Data analysis Data were analyzed in R (version 4.5.2). Soil CO₂ flux was modeled using linear mixed-effects models with a first-order autoregressive [AR(1)] covariance structure to account for temporal autocorrelation among repeated weekly measurements, with random intercepts for plot and collar. POXC and SOC were analyzed with linear mixed-effects models including a plot-level random intercept. Estimated marginal means, Dunnett- and Tukey-adjusted treatment comparisons, and Pearson correlations among indicators were derived from the fitted models.
Institutions
- Missouri University of Science and TechnologyMissouri, Rolla