Skip to contents

aridagri

aridagri analyses the field experiments that agronomists and plant breeders run every season: CRD, RCBD, factorials, the whole split-plot family, strip plots, combined analysis over years and locations, augmented and alpha lattice trials, repeated measurements, and stability analysis. Each analysis gives the correct error term for every source, SEm, SEd, CD and CV, and letter groupings, in the form journals and theses ask for.

Website: https://lalitrolaniya.github.io/aridagri/

Installation 

install.packages("aridagri")

# development version
# install.packages("remotes")
remotes::install_github("lalitrolaniya/aridagri")

A table for your paper in one call 

library(aridagri)

tab <- anova_traits(mothbean_rbd,
                    traits = c("plant_height", "pods_per_plant",
                               "seed_yield", "harvest_index"),
                    design = "rbd", treatment = "treatment",
                    block = "block", posthoc = "duncan")
export_results(tab, "Table_2.xlsx")
 treatment            plant_height pods_per_plant seed_yield harvest_index
 T1 Control           30.25 c      19.05 e        516.25 d   29.77
 T2 RDF               33.33 ab     24.70 cd       604.75 c   30.30
 ...
 T5 RDF+Rhizobium+PSB 35.00 a      29.10 a        705.00 a   32.33
 SEm ±                0.63         0.75           15.60      0.77
 CD (p = 0.05)        1.86         2.21           45.88      NS
 CV (%)               3.77         5.93           4.97       4.91

What it covers

Area Functions
Single-factor designs anova_crd(), anova_rbd() (missing plots handled), anova_latin()
Factorials anova_factorial(), anova_factorial_3way()
Split-plot family anova_spd(), anova_spd_ab_main(), anova_spd_c_main_ab_sub(), anova_spd_ab_cd(), anova_sspd(), anova_strip()
Over environments anova_rbd_pooled() (fixed or random environments), anova_spd_pooled(), anova_sspd_pooled()
Breeding trials anova_augmented() (adjusted means, genetic parameters), anova_alpha_lattice() (REML or intra-block)
Repeated measures anova_repeated() (sphericity corrections, CS/AR1/UN models)
Mean comparisons perform_posthoc(), compare_means(): LSD, DMRT, Tukey, SNK, Scheffe, Bonferroni, Dunnett
G x E stability_analysis() (Eberhart-Russell, AMMI, Finlay-Wilkinson, Shukla, Wricke, CV, Lin-Binns), gge_biplot()
Agronomic indices thermal_indices(), day_length(), crop_growth_analysis(), harvest_index(), yield_gap_analysis(), nue_calculate(), nutrient_response(), economic_analysis()
Multivariate correlation_analysis(), pca_analysis(), path_analysis(), sem_analysis()
Figures and export plot() methods, plot_means(), plot_interaction(), plot_stability(), plot_response(), arid_plot(), export_results()

Examples and figures

Most analyses have a plot() method, and there are dedicated functions for interaction, stability and response plots. Every figure is a ggplot2 object, so you can change labels or themes with + and save it with ggsave(). The example datasets are simulated.

Treatment means with letter groups 

res <- anova_rbd(mothbean_rbd, "seed_yield", "treatment", "block",
                 posthoc = "duncan")
plot(res, ylab = "Seed yield (kg/ha)",
     title = "Moth bean seed yield (RCBD, DMRT)")

Bar chart of moth bean seed yield for eight treatments with standard error bars and Duncan letter groups

Split plot: Error(a), Error(b) and the four CDs 

sp <- anova_spd(clusterbean_spd, "seed_yield", main_plot = "irrigation",
                sub_plot = "nitrogen", replication = "rep")
plot_interaction(sp, ylab = "Seed yield (kg/ha)",
                 title = "Cluster bean: irrigation x nitrogen")

Interaction plot of cluster bean seed yield over three irrigation levels for four nitrogen treatments

Stability over six environments 

st <- stability_analysis(pearlmillet_met, "genotype", "environment",
                         "rep", "grain_yield")
plot_stability(st, "eberhart",
               title = "Pearl millet: mean yield vs regression slope")

Eberhart-Russell plot of mean grain yield against regression slope for ten pearl millet genotypes

AMMI and GGE biplots come from the same data:

plot_stability(st, "ammi2")
gge <- gge_biplot(pearlmillet_met, "genotype", "environment", "grain_yield")
plot(gge, type = "which_won_where")

Germplasm in an augmented design 

au <- anova_augmented(chickpea_augmented, "seed_yield", "entry", "block",
                      check_names = c("C1", "C2", "C3", "C4"))
plot(au, top = 15, ylab = "Adjusted seed yield (kg/ha)",
     title = "Chickpea germplasm: top 15 accessions and checks")

Dot plot of adjusted seed yield for the top 15 chickpea accessions and four checks

Repeated measurements on the same plots 

rm <- anova_repeated(mungbean_repeated, "spad", "treatment", "das",
                     block = "block")
plot(rm, ylab = "SPAD reading", title = "Mung bean chlorophyll over time") +
  ggplot2::labs(x = "Days after sowing")

Line plot of mean SPAD readings at 25, 35, 45 and 55 days after sowing for four foliar spray treatments

Nutrient response and economic optimum dose 

dose  <- c(0, 20, 40, 60, 80, 100)
yield <- c(1100, 1350, 1520, 1610, 1650, 1660)
nr <- nutrient_response(dose, yield, model = "all",
                        price_output = 25, price_nutrient = 40)
plot_response(nr, dose, yield, title = "Nitrogen response: four models")

Yield response to nitrogen fitted with quadratic, linear plateau, quadratic plateau and Mitscherlich models, with economic optimum doses marked

Letters from a published table 

compare_means(c(T1 = 1180, T2 = 1265, T3 = 1342, T4 = 1410),
              mse = 3600, df_error = 12, reps = 4, method = "duncan")

What changed in 2.1.0

Version 2.1.0 corrects several calculations (relative efficiency in RCBD, the pooled RCBD error terms and silent mode, the strip-plot F-tests, the split-plot table returned to the user, Shukla’s variance, the AMMI stability value, Eberhart-Russell S2di, the variance test in check_assumptions(), and the utilization efficiency in nue_calculate()). If you used these functions in 2.0.x, please rerun the analyses. See NEWS for details.

All letter groupings are now computed in the package and agree with ‘agricolae’ on balanced data; unequal replication and missing plots are handled exactly. Results are tested against ‘agricolae’, ‘emmeans’, aov() error strata and base R’s sphericity tests (607 test expectations).

Citation 

citation("aridagri")

Rolaniya, L.K., Jat, R.L., Punia, M., Choudhary, R.R. and Poonia, H. (2026). aridagri: An R package for comprehensive statistical analysis of agricultural field experiments. Software Impacts, 29, 100851. https://doi.org/10.1016/j.simpa.2026.100851

Authors

Author Affiliation
Lalit Kumar Rolaniya (maintainer) Scientist (Agronomy), ICAR-Indian Institute of Pulses Research, Regional Centre, Bikaner
Ram Lal Jat Senior Scientist (Agronomy), ICAR-Indian Institute of Pulses Research, Regional Centre, Bikaner
Monika Punia (deceased) Scientist (Genetics and Plant Breeding), ICAR-Indian Institute of Pulses Research, Regional Centre, Bikaner
Raja Ram Choudhary Scientist (Agronomy), ICAR-Indian Institute of Groundnut Research, Regional Research Station, Bikaner
Hemant Poonia (deceased) Assistant Professor, Department of Mathematics and Statistics, CCS Haryana Agricultural University, Hisar

Version 2.1.0 is dedicated to the memory of our co-authors Dr. Monika Punia and Dr. Hemant Poonia.

License

GPL-3