aridagri

Comprehensive Statistical Tools for Agricultural Research

aridagri is an R package providing 33 functions for statistical analysis in agricultural research, with a focus on experimental design analysis and agronomic calculations. It is written in base R with no hard dependencies.

Key Features

ANOVA suite covering all common field designs
Post-hoc tests: LSD, Duncan (DMRT), Tukey, SNK, Scheffe, Bonferroni, Dunnett
Stability analysis: 7 methods (Eberhart-Russell, AMMI, Finlay-Wilkinson, Shukla, Wricke, CV, Lin-Binns) with an integrated ranking
Thermal indices: GDD, HTU, PTU, Heat Use Efficiency
Crop growth analysis: CGR, RGR, NAR
Publication-ready output: formatted tables with SE, CD, CV

Installation

# From CRAN
install.packages("aridagri")

# Development version from GitHub
# install.packages("remotes")
remotes::install_github("lalitrolaniya/aridagri")

Function Overview

Experimental Design ANOVA (16 functions)

Function	Design
`anova_crd()`	Completely Randomized Design
`anova_rbd()`	Randomized Block Design
`anova_rbd_pooled()`	Pooled RBD (multi-environment)
`anova_latin()`	Latin Square Design
`anova_factorial()`	Two-Factor Factorial
`anova_factorial_3way()`	Three-Factor Factorial
`anova_spd()`	Split-Plot Design
`anova_spd_ab_main()`	SPD with (A x B) in main plot
`anova_spd_c_main_ab_sub()`	SPD with C main, (A x B) sub
`anova_spd_ab_cd()`	SPD with (A x B) main, (C x D) sub
`anova_spd_pooled()`	Pooled Split-Plot Design
`anova_sspd()`	Split-Split-Plot Design
`anova_sspd_pooled()`	Pooled SSPD
`anova_strip()`	Strip-Plot Design
`anova_augmented()`	Augmented Block Design
`anova_alpha_lattice()`	Alpha Lattice Design

Stability, agronomic, statistical and nutrient functions

Function	Analysis
`stability_analysis()`	7 stability methods with integrated ranking
`thermal_indices()`	GDD, HTU, PTU, HUE
`crop_growth_analysis()`	CGR, RGR, NAR
`harvest_index()`	Harvest index and partitioning
`yield_gap_analysis()`	Yield gap calculations
`economic_indices()`	B:C ratio, net returns
`correlation_analysis()`	Correlation matrix with significance
`pca_analysis()`	Principal Component Analysis
`path_analysis()`	Path coefficient analysis
`sem_analysis()`	Structural Equation Modeling
`nue_calculate()`	Nutrient Use Efficiency indices
`nutrient_response()`	Response curve and economic optimum
`economic_analysis()`	Gross/net return, B:C ratio

Post-hoc, diagnostics and utilities

Function	Purpose
`perform_posthoc()`	Post-hoc comparisons (7 methods)
`check_assumptions()`	Normality, homogeneity, independence, outliers
`arid_plot()`	Base-graphics plots for ANOVA, correlation and stability objects
`export_results()`	Export results to Excel or CSV

Usage

Every function is called with the data frame first, then the column names as character strings. Set verbose = FALSE to suppress the printed report and just capture the returned object.

Analysis of variance

library(aridagri)

## Completely Randomized Design
crd <- data.frame(
  treatment = factor(rep(c("T1", "T2", "T3", "T4"), each = 4)),
  yield     = c(20, 22, 19, 21,  25, 27, 24, 26,
                30, 32, 29, 31,  28, 30, 27, 29)
)

res <- anova_crd(crd, response = "yield", treatment = "treatment")

## the returned object holds the table, means and statistics
res$anova_table
res$treatment_means
res$cv

## Randomized Block Design
anova_rbd(data, response = "yield", treatment = "variety", block = "block")

## Two-factor factorial
anova_factorial(data, response = "yield", factor1 = "nitrogen", factor2 = "variety")

## Split-plot design
anova_spd(data, response = "yield",
          main_plot = "irrigation", sub_plot = "variety", replication = "rep")

Post-hoc comparisons

There are two equivalent ways to run a post-hoc test.

## Option 1: inline, as part of the ANOVA call
anova_crd(crd, "yield", "treatment", posthoc = "tukey")

## Option 2: standalone, on a fitted aov() model
model <- aov(yield ~ treatment, data = crd)
perform_posthoc(model, crd, response = "yield", treatment = "treatment",
                posthoc = "tukey")

Valid method names are "lsd", "duncan", "tukey", "snk", "scheffe", "dunnett", "bonferroni", or "all". Dunnett compares every level against the first factor level as the control; there is no separate control argument, so order the treatment factor with the control first if needed.

Assumption checks

model <- aov(yield ~ treatment, data = crd)
check_assumptions(model)   # Shapiro-Wilk, Bartlett, Durbin-Watson, outliers

Split-Split-Plot Design

data <- expand.grid(
  rep        = 1:3,
  irrigation = c("I1", "I2", "I3"),
  variety    = c("V1", "V2"),
  nitrogen   = c("N0", "N40", "N80")
)
data$yield <- rnorm(nrow(data), 1200, 150)

anova_sspd(data,
           response     = "yield",
           main_plot    = "irrigation",
           sub_plot     = "variety",
           sub_sub_plot = "nitrogen",
           replication  = "rep")

Stability analysis (7 methods)

data <- expand.grid(
  variety  = paste0("V", 1:10),
  location = paste0("L", 1:5),
  rep      = 1:3
)
data$yield <- rnorm(nrow(data), 1200, 200)

stability_analysis(data,
                   genotype    = "variety",
                   environment = "location",
                   replication = "rep",
                   trait       = "yield",
                   method      = "all")

Correlation, PCA and visualization

df <- data.frame(
  yield   = rnorm(30, 1200, 150),
  pods    = rnorm(30, 50, 6),
  biomass = rnorm(30, 3500, 400)
)

cor_res <- correlation_analysis(df, method = "pearson")
pca_res <- pca_analysis(df, scale = TRUE)

## arid_plot() draws factor/treatment means for any ANOVA design,
## a heatmap for a correlation object, and a ranking for a stability object
arid_plot(res)
arid_plot(cor_res)

Exporting results

export_results(res, "results.xlsx")                  # Excel (default)
export_results(res, "results.csv", format = "csv")   # CSV

Unique Features

All split-plot design variations in one package
SE and CD reported for each comparison type
Seven stability analysis methods in a single function
Thermal indices (GDD, HTU, PTU, HUE) and crop growth analysis (CGR, RGR, NAR)
Base-R implementation with no hard dependencies

Citation

Rolaniya, L.K., Jat, R.L., Punia, M., and Choudhary, R.R. (2026). aridagri:
Comprehensive Statistical Tools for Agricultural Research.
R package version 2.0.4. https://github.com/lalitrolaniya/aridagri

Authors

Lalit Kumar Rolaniya (Maintainer) Scientist (Agronomy) ICAR-Indian Institute of Pulses Research, Regional Centre, Bikaner, Rajasthan-334006, India ORCID: 0000-0001-8908-1211

Ram Lal Jat Senior Scientist (Agronomy) ICAR-Indian Institute of Pulses Research, Regional Centre, Bikaner, Rajasthan-334006, India ORCID: 0009-0003-4339-0555

Monika Punia Scientist (Genetics & Plant Breeding) ICAR-Indian Institute of Pulses Research, Regional Centre, Bikaner, Rajasthan-334006, India ORCID: 0009-0002-0294-6767

Raja Ram Choudhary Scientist (Agronomy) ICAR-Indian Institute of Groundnut Research, Regional Research Station, Bikaner, Rajasthan-334006, India

License

GPL-3

Contributing

Contributions are welcome. Please submit issues or pull requests on GitHub.