Overview

aelab is an R package for aquatic ecology data analysis. It provides tools for:

1. GHG flux calculation — process raw data from LI-COR and LGR gas analyzers, calculate chamber flux, convert units, compute Minimum Detectable Flux (MDF), and aggregate to CO₂-equivalents.
2. Net ecosystem metabolism (NEM) — calculate gross primary production (GPP), ecosystem respiration (ER), and net ecosystem production (NEP) from high-frequency dissolved oxygen data.
3. Statistical analysis — descriptive statistics, outlier detection, normality testing, data transformation, one-way ANOVA (Tukey HSD), and Kruskal-Wallis (Dunn post-hoc) with compact letter display.
4. Visualization — ggplot2-based plot wrappers (point, line, box, bar) with a consistent theme and custom colour palettes.
5. Gantt-style progress tables — gantt_table() renders a coloured flextable (PDF/DOCX/PPTX-compatible) where columns are sampling campaigns and rows are measurement categories × systems; gantt_text() renders the same data as Unicode text for terminal/Neovim viewing.

GHG Flux — What's New (v1.1.5)

calculate_regression() has been substantially extended:

• fit_type = "auto" — automatic model selection per measurement window: tries linear → quadratic → exp_tz in order, accepting a more complex model only when its AICc is lower (Kutzbach et al. 2007), a quadratic only when it levels off rather than accelerates (Levy et al. 2011), and the exponential fit only when its curvature κ does not exceed kappamax. The model chosen is recorded in the new fit_used output column.
• window_type = "fixed" + start_offset_s — regresses a single fixed window starting at a user-specified dead-bar offset (e.g. 30 s) after the reference time. Ensures all gases are regressed over the same time period. The default "sliding" window remains available for backward compatibility.
• Per-row window overrides — reference_df may include optional duration_minutes and/or start_offset_s columns. When a row has a non-NA value, it overrides the global parameter for that measurement. Useful when individual closures have different deployment lengths or early-saturation artefacts.
• kappamax — caps exponential curvature at a physically plausible maximum. When fit_type = "auto", kappamax is computed automatically per window from the built-in analyzer_precision dataset (κmax = |linear slope| ÷ instrument 1σ precision; Hüppi et al. 2018).
• analyzer_precision dataset — built-in lookup table of 1-second precision values (1σ, in each analyzer's reporting unit: ppb for LI-COR CH₄/N₂O, otherwise ppm) for LI-7810 (CH₄, CO₂), LI-7820 (N₂O), and LGR UGGA (CH₄, CO₂). Used automatically by fit_type = "auto".
• fit_used column — new output column recording which model was actually selected ("linear", "quadratic", or "exp_tz"). Always matches fit_type unless fit_type = "auto".

make_reference() builds a reference_df for calculate_regression() from a compact multiline string of field notes — one closure per line: time site analyzer [dur=N] [off=N]. Optionally writes the result to .xlsx.

window_type = "pareto" independently searches deadband (start skip) and cutoff (window end) times per closure, trading off model fit against propagated slope uncertainty via a Pareto front — adapted from Kobayashi et al. (2026), with the existing kappamax gate and zero/discard fallback always applied so no candidate can be a numerically degenerate fit and no empty or failing candidate pool is reported as a confident number. New output columns: deadband, cutoff, slope_uncertainty_lo, slope_uncertainty_hi, n_candidates.

References

• Yong, Z.-J., Lin, W.-J., Lin, C.-W., & Lin, H.-J. (2024). Tidal influence on carbon dioxide and methane fluxes from tree stems and soils in mangrove forests. Biogeosciences, 21(22), 5247–5260. https://doi.org/10.5194/bg-21-5247-2024
• Staehr, P. A., Bade, D., Van de Bogert, M. C., Koch, G. R., Williamson, C., Hanson, P., Cole, J. J., & Kratz, T. (2010). Lake metabolism and the diel oxygen technique: State of the science. Limnology and Oceanography: Methods, 8(11), 628–644. https://doi.org/10.4319/lom.2010.8.0628
• Hutchinson, G. L., & Mosier, A. R. (1981). Improved soil cover method for field measurement of nitrous oxide fluxes. Soil Science Society of America Journal, 45(2), 311–316. https://doi.org/10.2136/sssaj1981.03615995004500020017x
• Hüppi, R., Felber, R., Krauss, M., Six, J., Leifeld, J., & Fuß, R. (2018). Restricting the nonlinearity parameter in soil greenhouse gas flux calculations for more reliable flux estimates. PLOS ONE, 13(7), e0200876. https://doi.org/10.1371/journal.pone.0200876
• Gaudard, J., Telford, R. J., Chacon-Labella, J., Dawson, H. R., Enquist, B. J., Töpper, J. P., Trepel, J., Vandvik, V., Baumane, M., Birkeli, K., Holle, M. J. M., Hupp, J. R., Santos-Andrade, P. E., Satriawan, T. W., & Halbritter, A. H. (2025). fluxible: An R package to process ecosystem gas fluxes from closed-loop chambers in an automated and reproducible way. Methods in Ecology and Evolution, 16(11), 2560–2568. https://doi.org/10.1111/2041-210X.70161
• Kobayashi, A. N. A., Roques, C., Hunkeler, D., & Brunner, P. (2026). Uncertainty quantification in soil gas fluxes: Reducing the user's arbitrariness. European Journal of Soil Science, 77, e70415. https://doi.org/10.1111/ejss.70415
• Jeffrey, S. W., & Humphrey, G. F. (1975). New spectrophotometric equations for determining chlorophylls a, b, c1 and c2 in higher plants, algae and natural phytoplankton. Biochemie und Physiologie der Pflanzen, 167(2), 191–194. https://doi.org/10.1016/S0015-3796(17)30778-3

Installation

The aelab package can be installed from:

1. GitHub (development version):

# install.packages("devtools")
# devtools::install_github("Zhao-Jun-Yong/aelab")
install.packages("aelab", repos = c('https://zhao-jun-yong.r-universe.dev', 'https://cloud.r-project.org'))

2. CRAN (stable version):

install.packages("aelab")

Features

For a full walkthrough with worked examples and code output, see the package vignette:

vignette("aelab")