Type:	Package
Title:	Sampling Design and Estimation Methods for Natural Resource Management
Version:	0.2.2
Description:	Provides functions for probability and non-probability sampling design, sample selection, and population estimation tailored to natural resource management. Probability methods include simple random sampling, stratified sampling, systematic sampling, cluster sampling, and probability-proportional-to-size sampling. Non-probability methods include convenience, judgement-based, and quota sampling. Estimation functions cover means, totals, ratio estimators, regression estimators, and the unequal-probability estimator of Horvitz and Thompson (1952, <doi:10.2307/2280784>) for unequal-probability designs. Utilities support biomass, soil-loss, and carbon-stock estimation from field plots. Spatial extensions provide random, systematic, stratified, and raster-weighted sampling within geographic polygons using the 'sf' and 'terra' packages, with extraction of remote-sensing covariates at sample locations. Applications include forest inventory, soil erosion monitoring, watershed studies, and ecological field surveys.
License:	GPL (≥ 3)
Encoding:	UTF-8
Language:	en-US
Depends:	R (≥ 4.1.0)
Imports:	stats, utils
Suggests:	sf (≥ 1.0.0), terra (≥ 1.7.0), ggplot2 (≥ 3.4.0), dplyr (≥ 1.1.0), testthat (≥ 3.0.0), knitr (≥ 1.40), rmarkdown (≥ 2.18), covr, spelling
VignetteBuilder:	knitr
RoxygenNote:	7.3.3
Config/testthat/edition:	3
LazyData:	true
NeedsCompilation:	no
Packaged:	2026-04-22 04:35:17 UTC; acer
Author:	Sadikul Islam [aut, cre]
Maintainer:	Sadikul Islam <sadikul.islamiasri@gmail.com>
Repository:	CRAN
Date/Publication:	2026-04-22 13:10:14 UTC

NRMSampling: Sampling Design and Estimation for Natural Resource Management

Description

Provides functions for probability and non-probability sampling design, sample selection, and estimation tailored to natural resource management.

Probability Sampling

srs_sample, stratified_sample, systematic_sample, cluster_sample, pps_sample

Non-Probability Sampling

convenience_sample, purposive_sample, quota_sample

Estimation

estimate_mean, estimate_total, estimate_variance, estimate_se, estimate_ci, ratio_estimator, regression_estimator, ht_estimator, ht_variance, stratified_estimator

NRM Utilities

biomass_estimate, soil_loss_estimate, carbon_stock_estimate, plot_summary, sampling_efficiency

Spatial Sampling

to_sf_points, spatial_random_sample, spatial_systematic_sample, spatial_stratified_sample, spatial_cluster_sample, raster_stratified_sample, raster_pps_sample, extract_raster_values, spatial_biomass_estimate, plot_sampling, plot_sampling_gg

Datasets

sample_nrm, sample_spatial

Note

sample(), runif(), setNames(), and mean() are base:: functions and must NOT be listed in importFrom(stats, ...). They are always available without any import declaration.

Author(s)

Maintainer: Sadikul Islam sadikul.islamiasri@gmail.com (ORCID)

References

Cochran, W.G. (1977). Sampling Techniques, 3rd ed. Wiley. Lohr, S.L. (2022). Sampling: Design and Analysis, 3rd ed. CRC Press.

Biomass Estimation from Field Plots

Description

Estimates total standing biomass over a landscape or management unit from plot-level measurements.

Usage

biomass_estimate(df, biomass_var, area)

Arguments

Value

A named list with elements:

mean_biomass

Mean biomass density across sampled plots.

total_biomass

Estimated total biomass over the study area.

se

Standard error of the mean biomass density.

n

Number of plots used.

References

Avery, T.E. and Burkhart, H.E. (2002). Forest Measurements, 5th ed. McGraw-Hill, New York.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
biomass_estimate(srs, biomass_var = "biomass", area = 1000)

Carbon Stock Estimation

Description

Converts a biomass estimate to carbon stock using a biomass-to-carbon conversion factor.

Usage

carbon_stock_estimate(df, biomass_var, area, carbon_fraction = 0.47)

Arguments

Value

A named list with elements total_biomass, total_carbon, carbon_fraction, and n.

References

IPCC (2006). IPCC Guidelines for National Greenhouse Gas Inventories, Volume 4: Agriculture, Forestry and Other Land Use. IGES, Hayama, Japan.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
carbon_stock_estimate(srs, biomass_var = "biomass", area = 1000)

Cluster Sampling

Description

Performs single-stage cluster sampling.

Usage

cluster_sample(data, cluster_var, n_clusters)

Arguments

Value

A data frame of sampled clusters.

Convenience Sampling

Description

Returns the first n rows of a data frame as a convenience sample. This is the simplest non-probability method; results are generally not representative of the population.

Usage

convenience_sample(data, n)

Arguments

Details

Convenience sampling is fast but prone to selection bias. It may be appropriate for pilot studies or logistical constraints, but population inference requires strong assumptions. See Lohr (2022) for discussion.

Value

A data frame containing the first n rows.

References

Lohr, S.L. (2022). Sampling: Design and Analysis, 3rd ed. CRC Press, Boca Raton, FL.

Examples

data(sample_nrm)
cs <- convenience_sample(sample_nrm, n = 10)
nrow(cs)

Confidence Interval for the Population Mean

Description

Computes a confidence interval for the population mean based on a simple random sample, using the t-distribution.

Usage

estimate_ci(y, N = NULL, conf_level = 0.95)

Arguments

Value

A named numeric vector with elements mean, lower, and upper.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
estimate_ci(srs$biomass, N = nrow(sample_nrm))

Sample Mean

Description

Computes the sample mean of a numeric vector, ignoring missing values.

Usage

estimate_mean(y)

Arguments

Value

Numeric scalar. The sample mean.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
estimate_mean(srs$biomass)

Standard Error of the Sample Mean

Description

Computes the estimated standard error of the sample mean under simple random sampling without replacement (SRSWOR).

Usage

estimate_se(y, N = NULL)

Arguments

Details

With fpc:

SE(\bar{y}) = \sqrt{\frac{s^2}{n}\left(1 - \frac{n}{N}\right)}

Without fpc:

SE(\bar{y}) = \sqrt{\frac{s^2}{n}}

Value

Numeric scalar. Estimated standard error.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
estimate_se(srs$biomass, N = nrow(sample_nrm))

Population Total Estimator

Description

Estimates the population total by expanding the sample mean to the full population size.

Usage

estimate_total(y, N)

Arguments

Details

\hat{Y} = N \bar{y}

Value

Numeric scalar. Estimated population total.

References

Cochran, W.G. (1977). Sampling Techniques, 3rd ed. John Wiley & Sons, New York.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
estimate_total(srs$biomass, N = nrow(sample_nrm))

Sample Variance

Description

Computes the unbiased sample variance of a numeric vector.

Usage

estimate_variance(y)

Arguments

Value

Numeric scalar. The unbiased sample variance s^2.

Examples

data(sample_nrm)
estimate_variance(sample_nrm$biomass)

Extract Raster Values at Sample Points

Description

Extracts cell values from a SpatRaster at the locations of sf point features, returning the results as a data frame.

Usage

extract_raster_values(raster, points_sf)

Arguments

Details

Requires both the terra and sf packages. The CRS of points_sf is reprojected to match raster if needed.

Value

A data frame with one row per point and one column per raster layer.

Examples

if (requireNamespace("terra", quietly = TRUE) &&
    requireNamespace("sf", quietly = TRUE)) {
  r    <- terra::rast(nrows=20, ncols=20, vals=runif(400))
  bbox <- sf::st_as_sfc(sf::st_bbox(terra::ext(r)))
  pts  <- spatial_random_sample(bbox, n = 10)
  sf::st_crs(pts) <- sf::st_crs(terra::crs(r))
  extract_raster_values(r, pts)
}

Horvitz-Thompson (HT) Estimator

Description

Provides a design-unbiased estimate of the population total for unequal-probability sampling designs.

Usage

ht_estimator(y, pi)

Arguments

Details

\hat{Y}_{HT} = \sum_{i \in s} \frac{y_i}{\pi_i}

Value

Numeric scalar. HT estimate of the population total.

References

Horvitz, D.G. and Thompson, D.J. (1952). A generalization of sampling without replacement from a finite universe. Journal of the American Statistical Association, 47(260), 663–685.

Examples

data(sample_nrm)
pps <- pps_sample(sample_nrm, size_var = "size", n = 20)
pi_i <- pps$.inclusion_prob
ht_estimator(pps$biomass, pi = pi_i)

Variance of the Horvitz-Thompson Estimator

Description

Estimates the variance of the Horvitz-Thompson total estimator using the Sen-Yates-Grundy approximation for with-replacement PPS designs.

Usage

ht_variance(y, pi)

Arguments

Value

Numeric scalar. Estimated variance of the HT total.

References

Yates, F. and Grundy, P.M. (1953). Selection without replacement from within strata with probability proportional to size. Journal of the Royal Statistical Society B, 15, 253–261.

Examples

data(sample_nrm)
pps <- pps_sample(sample_nrm, size_var = "size", n = 20)
ht_variance(pps$biomass, pps$.inclusion_prob)

Plot Sample Points (Base Graphics)

Description

Produces a simple plot of sample point locations using base graphics.

Usage

plot_sampling(
  sf_points,
  col = "steelblue",
  pch = 16,
  main = "Sample Locations"
)

Arguments

Details

Requires the sf package.

Value

NULL invisibly. Called for its side effect (a plot).

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  bbox <- sf::st_as_sfc(sf::st_bbox(c(xmin=77, xmax=78,
                                       ymin=30, ymax=31),
                         crs = sf::st_crs(4326)))
  pts <- spatial_random_sample(bbox, n = 20)
  plot_sampling(pts)
}

Plot Sample Points with ggplot2

Description

Produces a publication-quality map of sample point locations using ggplot2.

Usage

plot_sampling_gg(
  sf_points,
  colour = "tomato",
  size = 2,
  title = "Spatial Sample Locations"
)

Arguments

Details

Requires the sf and ggplot2 packages.

Value

A ggplot object.

Examples

if (requireNamespace("sf", quietly = TRUE) &&
    requireNamespace("ggplot2", quietly = TRUE)) {
  bbox <- sf::st_as_sfc(sf::st_bbox(c(xmin=77, xmax=78,
                                       ymin=30, ymax=31),
                         crs = sf::st_crs(4326)))
  pts <- spatial_random_sample(bbox, n = 25)
  plot_sampling_gg(pts)
}

Summary Statistics for a Sampled NRM Plot Dataset

Description

Returns a concise summary table of key variables from a sample, including means, standard deviations, and sample sizes.

Usage

plot_summary(df, vars = NULL)

Arguments

Value

A data frame with columns variable, n, mean, sd, min, and max.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
plot_summary(srs, vars = c("biomass", "soil_loss"))

PPS Sampling

Description

Performs probability proportional to size sampling.

Usage

pps_sample(data, size_var, n)

Arguments

Value

A data frame with inclusion probabilities.

Purposive (Judgement) Sampling

Description

Selects units from a data frame that satisfy a logical condition supplied as a character string. This is a non-probability method in which units are selected based on expert judgement or predetermined criteria.

Usage

purposive_sample(data, condition)

Arguments

Details

The condition is parsed and evaluated in the context of data using with(). Only columns present in data may be referenced.

Value

A data frame of rows satisfying condition.

Examples

data(sample_nrm)
# Select high-biomass forest plots
ps <- purposive_sample(sample_nrm,
                       condition = "biomass > 30 & strata == 'forest'")
nrow(ps)

Quota Sampling

Description

Selects a fixed number of units from the top of each stratum. This non-probability method resembles stratified sampling but does not use random selection within strata.

Usage

quota_sample(data, strata_var, quota)

Arguments

Value

A data frame containing the selected rows from each stratum.

References

Lohr, S.L. (2022). Sampling: Design and Analysis, 3rd ed. CRC Press, Boca Raton, FL.

Examples

data(sample_nrm)
qs <- quota_sample(sample_nrm, strata_var = "strata", quota = 5)
table(qs$strata)

# Variable quotas
q <- c(forest = 6, agriculture = 3, grassland = 4)
qs2 <- quota_sample(sample_nrm, "strata", quota = q)

Raster-Weighted PPS Spatial Sampling

Description

Selects sample points from a SpatRaster with probability proportional to cell values (e.g., vegetation density, erosion risk).

Usage

raster_pps_sample(raster, n)

Arguments

Details

Requires both the terra and sf packages. Identical in behaviour to raster_stratified_sample; exposed as a separate function to match PPS nomenclature.

Value

An sf POINT object.

Examples

if (requireNamespace("terra", quietly = TRUE) &&
    requireNamespace("sf", quietly = TRUE)) {
  r <- terra::rast(nrows = 20, ncols = 20, vals = runif(400, 0, 1))
  pts <- raster_pps_sample(r, n = 10)
}

Raster-Stratified Spatial Sampling

Description

Samples a specified number of cells from a SpatRaster with probability proportional to cell values, and returns their coordinates as an sf point object.

Usage

raster_stratified_sample(raster, n)

Arguments

Details

Requires both the terra and sf packages. Cells with NA values are excluded from sampling.

Value

An sf POINT object with CRS taken from raster.

Examples

if (requireNamespace("terra", quietly = TRUE) &&
    requireNamespace("sf", quietly = TRUE)) {
  r <- terra::rast(nrows = 20, ncols = 20, vals = runif(400, 1, 100))
  pts <- raster_stratified_sample(r, n = 15)
  plot(sf::st_geometry(pts))
}

Ratio Estimator

Description

Estimates the population total or mean of a study variable y using a correlated auxiliary variable x with a known population total X.

Usage

ratio_estimator(y, x, X_total)

Arguments

Details

\hat{R} = \frac{\sum y_i}{\sum x_i}, \qquad \hat{Y}_R = \hat{R} \cdot X

Value

Numeric scalar. Ratio estimate of the population total of y.

References

Cochran, W.G. (1977). Sampling Techniques, 3rd ed. John Wiley & Sons, New York.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
X_total <- sum(sample_nrm$size)
ratio_estimator(y = srs$biomass, x = srs$size, X_total = X_total)

Regression Estimator

Description

Provides a model-assisted estimate of the population mean of y using a known population mean of the auxiliary variable x.

Usage

regression_estimator(y, x, X_mean)

Arguments

Details

\hat{\bar{Y}}_{\text{reg}} = \bar{y} + \hat{\beta}(\bar{X} - \bar{x})

where \hat{\beta} is the ordinary least-squares slope from regressing y on x.

Value

Numeric scalar. Regression estimate of the population mean.

References

Sarndal, C.E., Swensson, B., and Wretman, J. (2003). Model Assisted Survey Sampling. Springer, New York.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 30)
X_mean <- mean(sample_nrm$size)
regression_estimator(y = srs$biomass, x = srs$size, X_mean = X_mean)

Simulated NRM Plot Dataset

Description

A synthetic dataset of 100 field plots representing a heterogeneous natural resource management landscape with three strata and ten spatial clusters.

Format

A data frame with 100 rows and 7 variables:

plot_id

Integer. Unique plot identifier (1–100).

biomass

Numeric. Aboveground biomass density (Mg/ha), drawn from Uniform(5, 50).

soil_loss

Numeric. Annual soil loss (Mg/ha/yr), drawn from Uniform(0.1, 10).

strata

Character. Land-use stratum: one of "forest", "agriculture", or "grassland".

cluster

Integer. Spatial cluster identifier (1–10).

size

Numeric. Plot size measure used for PPS sampling (e.g., stand basal area), drawn from Uniform(1, 100).

carbon

Numeric. Estimated carbon stock (Mg C/ha), derived as 0.47 * biomass.

Source

Synthetic data generated in data-raw/generate_datasets.R.

Examples

data(sample_nrm)
head(sample_nrm)
table(sample_nrm$strata)
summary(sample_nrm[, c("biomass", "soil_loss")])

Simulated Spatial NRM Dataset

Description

A synthetic dataset of 100 geo-referenced field observations within a one-degree tile (77–78 degrees E, 30–31 degrees N) representing a Himalayan watershed zone.

Format

A data frame with 100 rows and 8 variables:

id

Integer. Observation identifier.

lon

Numeric. Longitude (decimal degrees, WGS 84).

lat

Numeric. Latitude (decimal degrees, WGS 84).

biomass

Numeric. Aboveground biomass (Mg/ha).

soil_loss

Numeric. Annual soil loss (Mg/ha/yr).

strata

Character. Land-use class: "forest" or "agriculture".

cluster

Integer. Spatial cluster (1–5).

ndvi

Numeric. Synthetic NDVI value (0–1), correlated with biomass.

Source

Synthetic data generated in data-raw/generate_datasets.R.

Examples

data(sample_spatial)
head(sample_spatial)

if (requireNamespace("sf", quietly = TRUE)) {
  pts <- to_sf_points(sample_spatial, lon = "lon", lat = "lat")
  plot(sf::st_geometry(pts))
}

Sampling Efficiency Comparison

Description

Computes the relative efficiency (RE) of two sampling designs by comparing their estimated variances of the mean. RE > 1 indicates Design 2 is more efficient than Design 1.

Usage

sampling_efficiency(y1, y2, N = NULL)

Arguments

Value

A named numeric vector with var_design1, var_design2, and relative_efficiency (var1 / var2).

Examples

data(sample_nrm)
srs1 <- srs_sample(sample_nrm, n = 20)
srs2 <- srs_sample(sample_nrm, n = 30)
sampling_efficiency(srs1$biomass, srs2$biomass, N = 100)

Soil Loss Estimation from Sample Plots

Description

Estimates mean and total soil loss from a set of erosion measurement plots, with a standard error.

Usage

soil_loss_estimate(df, loss_var, area)

Arguments

Value

A named list with elements mean_loss, total_loss, se, and n.

References

Wischmeier, W.H. and Smith, D.D. (1978). Predicting Rainfall Erosion Losses. USDA Agriculture Handbook 537.

Examples

data(sample_nrm)
srs <- srs_sample(sample_nrm, n = 25)
soil_loss_estimate(srs, loss_var = "soil_loss", area = 500)

Spatial Biomass Estimation from sf Points

Description

Estimates total biomass over a study area from field measurements stored in an sf point object.

Usage

spatial_biomass_estimate(sf_data, biomass_var, area)

Arguments

Details

Requires the sf package.

Value

A named list with mean_biomass, total_biomass, se, and n.

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  data(sample_spatial)
  pts_sf <- to_sf_points(sample_spatial, lon = "lon", lat = "lat")
  spatial_biomass_estimate(pts_sf, biomass_var = "biomass", area = 1000)
}

Spatial Cluster Sampling

Description

Randomly selects a number of spatial clusters (e.g., sub-watersheds, administrative units) and returns all features within selected clusters.

Usage

spatial_cluster_sample(sf_data, cluster_var, n_clusters)

Arguments

Details

Requires the sf package.

Value

An sf object containing features from the selected clusters.

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  data(sample_spatial)
  pts_sf <- to_sf_points(sample_spatial, lon = "lon", lat = "lat")
  cl_sp  <- spatial_cluster_sample(pts_sf, "cluster", n_clusters = 3)
  length(unique(cl_sp$cluster))
}

Spatial Random Sampling within a Polygon

Description

Draws a random sample of points uniformly distributed within an sf polygon or multipolygon geometry.

Usage

spatial_random_sample(polygon, n)

Arguments

Details

Requires the sf package. Points are drawn using sf::st_sample(..., type = "random").

Value

An sf object of POINT geometries within polygon.

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  # Create a simple rectangular polygon for illustration
  bbox <- sf::st_as_sfc(sf::st_bbox(c(xmin=77, xmax=78,
                                       ymin=30, ymax=31),
                         crs = sf::st_crs(4326)))
  pts <- spatial_random_sample(bbox, n = 20)
  plot(sf::st_geometry(pts))
}

Stratified Spatial Sampling

Description

Selects a random sample of n_per_stratum features from each stratum of an sf object.

Usage

spatial_stratified_sample(sf_data, strata_var, n_per_stratum)

Arguments

Details

Requires the sf package.

Value

An sf object with the selected features.

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  data(sample_spatial)
  pts_sf <- to_sf_points(sample_spatial, lon = "lon", lat = "lat")
  st_sp  <- spatial_stratified_sample(pts_sf, "strata", n_per_stratum = 5)
  table(st_sp$strata)
}

Systematic Grid Sampling within a Polygon

Description

Generates a systematic grid of points at a specified spacing within an sf polygon, retaining only points that fall inside the boundary.

Usage

spatial_systematic_sample(polygon, spacing)

Arguments

Details

Requires the sf package.

Value

An sf object of POINT geometries inside polygon.

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  bbox <- sf::st_as_sfc(sf::st_bbox(c(xmin=77, xmax=78,
                                       ymin=30, ymax=31),
                         crs = sf::st_crs(4326)))
  grid_pts <- spatial_systematic_sample(bbox, spacing = 0.1)
  plot(sf::st_geometry(grid_pts))
}

Simple Random Sampling

Description

Draws a simple random sample from a data frame.

Usage

srs_sample(data, n, replace = FALSE)

Arguments

Value

A data frame with sampled rows and a ".sample_id" column.

Stratified Mean Estimator

Description

Estimates the population mean from a stratified sample using stratum weights (proportional to stratum size).

Usage

stratified_estimator(y, strata, N_h)

Arguments

Details

\hat{\bar{Y}}_{st} = \sum_h W_h \bar{y}_h, \qquad W_h = N_h / N

Value

Numeric scalar. Estimated population mean.

References

Cochran, W.G. (1977). Sampling Techniques, 3rd ed. John Wiley & Sons, New York.

Examples

data(sample_nrm)
st <- stratified_sample(sample_nrm, "strata", n_per_stratum = 5)
N_h <- table(sample_nrm$strata)
stratified_estimator(st$biomass, st$strata, N_h)

Stratified Sampling

Description

Performs stratified random sampling.

Usage

stratified_sample(data, strata_var, n_per_stratum, replace = FALSE)

Arguments

Value

A data frame with sampled rows.

Systematic Sampling

Description

Performs systematic sampling using interval k.

Usage

systematic_sample(data, k)

Arguments

Value

A data frame of sampled rows.

Convert a Data Frame to an sf Point Object

Description

Creates an sf simple-features point object from longitude and latitude columns in a data frame.

Usage

to_sf_points(data, lon, lat, crs = 4326)

Arguments

Details

Requires the sf package.

Value

An sf object with a POINT geometry column.

Examples

if (requireNamespace("sf", quietly = TRUE)) {
  data(sample_spatial)
  pts <- to_sf_points(sample_spatial, lon = "lon", lat = "lat")
  print(pts)
}