Using artfishr for Small-Scale Fisheries Estimation • artfishr

Introduction

This vignette demonstrates how to apply the ARTFISH methodology for estimating fisheries catch and effort, and related indicators using the artfishr R package. This methodology aims to extrapolate sample-based observations to produce total estimates for each stratum of the sample.

The vignette walks through each computation step and shows how to use the unified workflow function artfish_compute_report(), which automates the full estimation process.

Data preparation

The package includes example datasets stored under inst/extdata/samples/.

Let’s load them using system.file():

active_vessels <- readr::read_csv(
  system.file("extdata/samples", "active_vessels.csv", package = "artfishr")
)
#> Rows: 12 Columns: 6
#> ── Column specification ────────────────────────────────────────────────────────
#> Delimiter: ","
#> chr (1): fishing_unit
#> dbl (5): year, month, minor_stratum, landing_site, fleet_engagement_number
#> 
#> ℹ Use `spec()` to retrieve the full column specification for this data.
#> ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

effort <- readr::read_csv(
  system.file("extdata/samples", "effort_fisher_interview.csv", package = "artfishr")
)
#> Rows: 164 Columns: 9
#> ── Column specification ────────────────────────────────────────────────────────
#> Delimiter: ","
#> chr (2): fishing_unit, effort_fishing_duration_unit
#> dbl (7): year, month, day, minor_stratum, landing_site, effort_fishing_durat...
#> 
#> ℹ Use `spec()` to retrieve the full column specification for this data.
#> ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

active_days <- readr::read_csv(
  system.file("extdata/samples", "active_days.csv", package = "artfishr")
)
#> Rows: 12 Columns: 6
#> ── Column specification ────────────────────────────────────────────────────────
#> Delimiter: ","
#> chr (1): fishing_unit
#> dbl (5): year, month, minor_stratum, landing_site, effort_fishable_duration
#> 
#> ℹ Use `spec()` to retrieve the full column specification for this data.
#> ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

landings <- readr::read_csv(
  system.file("extdata/samples", "landings.csv", package = "artfishr")
)
#> Rows: 570 Columns: 15
#> ── Column specification ────────────────────────────────────────────────────────
#> Delimiter: ","
#> chr (6): fishing_trip, fishing_unit, effort_fishing_duration_unit, species, ...
#> dbl (9): year, month, day, minor_stratum, landing_site, effort_fishing_durat...
#> 
#> ℹ Use `spec()` to retrieve the full column specification for this data.
#> ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

The Following datasets are required to produce the estimates: - Active vessels: number of active vessels per stratum (or per landing site) - Active days: number of days in the month recording a fishing activity - Effort survey data: fisher interviews or boat counting - Landings data: catch by species for each observed fishing trip For more information on the data structure, refer to the example datasets and to the data validation requirements.

Identify data collection strategy

(to complete with different cases + schema)

Different cases are listed above, depending on the different data sources. The effort survey type (fisher interview or boat counting) must be indicated in the function to use the correct computation. In case of boat counting, the dataset active_days is mandatory.

Step-by-step ARTFISH workflow

Below is the detailed workflow showing each function used by artfishr to compute the various ARTFISH components.

Indicators are computed for each stratum of the sampling plan. Major and minor strata must be specified in the function to define the level of aggregation.

1. Effort Activity Coefficient

The activity coefficient represents the probability that a certain boat is out on a certain day. It is obtained with the effort survey using either fisher interviews or boat counts.

activity_coefficient <- artfishr::compute_effort_activity_coefficient(
  effort = effort,
  effort_source = "fisher_interview",
  minor_strata = "minor_stratum"
)

2. Effort Estimate

Effort estimate calculation is based on the three components active vessels, active days and activity coefficient.

$Effort = ActiveVessels \times ActiveDays \times ActivityCoefficient$

effort_estimate <- artfishr::compute_effort_estimate(
  active_vessels = active_vessels,
  active_vessels_strategy = "latest",
  landings = landings,
  effort = effort,
  effort_source = "fisher_interview",
  active_days = active_days,
  minor_strata = "minor_stratum"
)
#> [1] "2025-01-01"

3. Catch per Unit of Effort (CPUE)

CPUE is calculated with the landing survey data. For each stratum, the overall average CPUE is calculated.

cpue <- artfishr::compute_cpue(
  landings,
  minor_strata = "minor_stratum"
)

4. Catch Estimate

Catch estimate calculation is based on the formula:

$Catch = CPUE \times Effort$

catch_estimate <- artfishr::compute_catch_estimate(
  effort_estimate,
  landings = landings,
  minor_strata = "minor_stratum"
)

5. Catch Estimate by Species

Catch estimates by species are obtained by distributing the estimated total catch according to the observed species composition within each stratum.

catch_estimate_by_species <- artfishr::compute_catch_estimates_by_species(
  landings,
  catch_estimate,
  minor_strata = "minor_stratum"
)

Using the unified workflow

The individual steps above are integrated into a single convenience function:
artfish_compute_report().
This function executes the complete workflow, returning a structured report that includes all intermediate and final results.

report <- artfishr::compute_report(
  active_vessels = active_vessels,
  effort = effort,
  effort_source = "fisher_interview",
  active_days = active_days,
  active_vessels_strategy = "closest",
  landings = landings,
  minor_strata = "minor_stratum"
)
#> [1] "2025-01-01"

# Inspect report structure
str(report, max.level = 1)
#> tibble [22 × 36] (S3: tbl_df/tbl/data.frame)

Interpreting results

Each component of the output can be inspected individually:

head(effort_estimate)
#> # A tibble: 6 × 11
#>    year month fishing_unit            minor_stratum effort_sample_size
#>   <dbl> <dbl> <chr>                           <dbl>              <int>
#> 1  2025     1 Artisanal 0-6m trap                 1                 20
#> 2  2025     1 Artisanal 0-6m trap                 2                 19
#> 3  2025     1 Artisanal 6-12m gillnet             1                 12
#> 4  2025     1 Artisanal 6-12m gillnet             2                 49
#> 5  2025     1 Artisanal 6-12m trawl               1                 25
#> 6  2025     1 Artisanal 6-12m trawl               2                 39
#> # ℹ 6 more variables: effort_coefficient_variation <dbl>,
#> #   effort_total_fishing_duration <dbl>, effort_activity_coefficient <dbl>,
#> #   fleet_engagement_number <dbl>, effort_fishable_duration <int>,
#> #   effort_nominal <dbl>
head(catch_estimate_by_species)
#> # A tibble: 6 × 18
#>    year month fishing_unit minor_stratum species catch_number effort_sample_size
#>   <dbl> <dbl> <chr>                <dbl> <chr>          <dbl>              <int>
#> 1  2025     1 Artisanal 0…             1 Crab             248                 20
#> 2  2025     1 Artisanal 0…             1 Grouper          485                 20
#> 3  2025     1 Artisanal 0…             1 Lobster          630                 20
#> 4  2025     1 Artisanal 0…             1 Parrot…          902                 20
#> 5  2025     1 Artisanal 0…             1 Snapper          534                 20
#> 6  2025     1 Artisanal 0…             2 Crab             268                 19
#> # ℹ 11 more variables: effort_coefficient_variation <dbl>,
#> #   effort_total_fishing_duration <dbl>, catch_sample_size <int>,
#> #   catch_coefficient_variation <dbl>, catch_cpue <dbl>,
#> #   catch_number_species <int>, catch_species_ratio <dbl>,
#> #   catch_nominal_landed <dbl>, trade_price <dbl>, trade_value <dbl>,
#> #   catch_fish_average_weight <dbl>

Summary

This vignette illustrated:

The data requirements for ARTFISH (active vessels, effort, active days, landings)
The sequential workflow used to compute all indicators manually
The integrated workflow provided by artfish_compute_report() for convenience

For production use, users should adapt the workflow to their own datasets, ensuring that data formats comply with the specifications in inst/extdata/format_specs/.