Skip to contents

Linking Herbarium Specimens to Inventory Individuals

CafriplotsR Package

2026-01-22

specimen_linking_workflow.Rmd

Introduction

The Challenge: Disconnected Taxonomic Knowledge

Forest inventory data and herbarium specimen data often exist in separate silos. Field botanists collect specimens from individual trees during plot censuses, but the connection between the voucher specimen and the measured tree is typically recorded only in field notebooks. When taxonomists later revise the specimen’s identification in the herbarium, this improved knowledge doesn’t automatically flow back to the ecological dataset.

This creates a fundamental problem: Your inventory data can become taxonomically outdated even though the correct identification exists somewhere in a herbarium database.

The CafriplotsR package solves this by creating persistent, database-level connections between:

  • Individual trees in your forest plots (with DBH, height, coordinates, etc.)
  • Herbarium specimens collected from those exact individuals (with expert-verified taxonomy)

Once linked, taxonomic updates to specimens automatically propagate to your inventory data—forever.

How It Works: Step-by-Step Workflow

Prerequisites

Before linking specimens, ensure you have:

  1. Database access with appropriate credentials
  2. Herbarium reference data in your individuals table (see columns below)
  3. Specimens already entered in the herbarium database

Understanding the two types of herbarium columns:

The system uses two columns to track different types of specimen-individual relationships:

  • herbarium_nbe_type: The ACTUAL tree where the herbarium specimen was physically collected
    • Direct evidence: “This specimen was collected FROM this tree”
    • High confidence link
    • Link type: type_individual
  • herbarium_nbe_char: Other trees that were field-identified as the SAME SPECIES as the collected specimen
    • Indirect evidence: “We believe this tree is the same species as the specimen tree”
    • Lower confidence (based on field identification assumption)
    • Link type: referenced_individual

Rationale: Extending specimen utility

Since it’s impractical to collect herbarium specimens from every single tree in a forest inventory, this system maximizes the value of collected specimens:

  1. Collect a specimen from one representative individual (recorded in herbarium_nbe_type)
  2. In the field, identify other trees as the same species (recorded in herbarium_nbe_char)
  3. Link all these trees to the one specimen
  4. When taxonomists revise the specimen → all linked trees inherit the updated identification

The trade-off: Links via herbarium_nbe_char are less reliable because they depend on the accuracy of field identification. However, they greatly extend specimen utility by providing taxonomic updates to many more trees than would otherwise be possible.

Step 1: Launch the Linking Application 

library(CafriplotsR)

# Launch the interactive specimen linking app
launch_individual_specimen_linking_app()

The app will prompt you for database credentials if not already connected.

Step 2: Select Plot(s) and Filter Individuals

The app starts by loading all individuals from your accessible plots that have herbarium reference information.

What happens: - Reads the herbarium_nbe_char column from data_individuals table - This column typically contains collector name and specimen number (e.g., “Dauby 123”, “Smith G. 456”) - Displays a summary table showing which individuals have specimen references

Your actions: - Review the individuals with herbarium information - Optionally filter by plot, collector, or taxonomic group - The table highlights individuals with specimen references (green background)

Step 3: Parse Herbarium References

The app automatically extracts collector names and specimen numbers from the text references.

What happens: - Parses herbarium_nbe_char using pattern matching - Handles various formats: “Dauby 123”, “G. Dauby 123”, “Dauby, G. 123” - Matches collector names to the table_colnam database (people lookup table) - Extracts specimen numbers as integers

Result: - A table showing parsed collector IDs and specimen numbers - Any unparseable references are flagged for review

Step 4: Retrieve Matching Specimens

The app queries the herbarium database to find specimens matching the parsed references.

What happens: - Groups by collector and gets min/max specimen number range - Queries the specimens table using batch retrieval (very fast!) - For example: Instead of 200 individual queries, makes ~3 queries (one per collector)

Result: - A table of potential specimen matches - Shows specimen ID, collector, number, and current taxonomic identification

Performance note: The batch query optimization means even large linking operations (200+ individuals) complete in seconds.

Step 5: Validate Taxonomic Concordance

Before creating links, the app validates that specimen and individual identifications are compatible.

What happens: - Retrieves taxonomy for both individuals and specimens from the taxa database - Constructs full taxonomic names (genus + species + infraspecific) - Compares at multiple levels: - Exact match: Same species (idtax_n matches) - Same genus: Different species but same genus - Same family: Different genus but same family - Different family: Potential misidentification

Visual indicators: - ✓ Green = Exact match (auto-approve recommended) - Yellow = Same genus (review recommended) - Orange = Same family (review required) - Red = Different family (⚠ likely error, manual verification needed)

Your actions: - Review the validation table - Click “Approve All Exact Matches” for confident links - Manually review and approve/reject uncertain links - Filter by match type to focus on problematic cases

Once you’ve approved the links, the app writes them to the database.

What happens: - Writes to link_individual_specimen table with columns: - id_n: Individual ID (foreign key to data_individuals) - id_specimen: Specimen ID (foreign key to specimens) - id_linktype: Link type ID (e.g., “referenced_individual”, “type_individual”) - created_by: Your username (audit trail) - created_at: Timestamp (audit trail)

Validation: - Checks that all foreign key references exist (batch validation) - Prevents duplicate links - Ensures data integrity

Result: - Confirmation message showing number of links created - Links are now permanent and queryable

The system supports multiple link types to capture the nature of the relationship:

type_individual

Direct evidence: The herbarium specimen was physically collected from this specific tree during plot census. This provides the highest confidence link between the measured individual and its taxonomic identification.

referenced_individual

Indirect evidence: This tree was field-identified as the same species as another tree from which a specimen was collected. The tree is linked to that specimen, but the specimen itself was not collected from this individual. This extends the taxonomic value of collected specimens to additional trees, though with lower confidence since it relies on field identification accuracy.

How it’s determined: - The app checks if herbarium_nbe_type column matches herbarium_nbe_char - If yes → type_individual (specimen collected from this tree) - If no → referenced_individual (tree identified as same species)

Querying Linked Specimens

Once links are established, you can query them:

library(CafriplotsR)

# Connect to database
mydb <- call.mydb()

# Get all specimen links for a plot
links <- query_all_specimen_links(plot_ids = 1, con = mydb)

# Get individuals with their linked specimens
individuals <- query_plots(
  id_plot = 1,
  extract_individuals = TRUE,
  con = mydb
)

# Check which individuals have specimen links
linked_individuals <- individuals %>%
  filter(!is.na(id_specimen))

Benefits in Practice

Scenario 1: Taxonomic Revision

Initial state: - Individual #12345 identified as Guarea thompsonii in field notes - Specimen DAU-1234 collected from this individual - Link created between individual and specimen

Years later: - Taxonomist revises specimen DAU-1234 → Guarea cedrata - Update made in herbarium database

Result: - Next query of individual #12345 automatically returns Guarea cedrata - No manual updates needed in inventory database - Historical record preserved: you can still see the original field ID

Scenario 2: Data Quality Metrics

With specimen links, you can calculate:

# Proportion of individuals with specimen backing
specimen_coverage <- sum(!is.na(individuals$id_specimen)) / nrow(individuals)

# Taxonomic accuracy: how many field IDs match specimen IDs?
accuracy <- sum(individuals$field_idtax == individuals$specimen_idtax, na.rm = TRUE) /
            sum(!is.na(individuals$id_specimen))

Scenario 3: Citation and Traceability

When publishing, you can now cite the herbarium specimens directly:

“Species identifications were verified by herbarium specimens (Dauby 123, 145, 167; deposited at LBV) collected from the same individuals during plot census.”

Database Schema

The linking system uses these key tables:

data_individuals
├─ id_n (PK)
├─ herbarium_nbe_char (specimen reference text)
├─ herbarium_nbe_type (reference for specimen collected from this tree)
└─ idtax_n (field identification)

specimens
├─ id_specimen (PK)
├─ id_colnam (collector ID → table_colnam)
├─ colnbr (specimen number)
└─ idtax_specimen (specimen identification)

link_individual_specimen
├─ id_link (PK)
├─ id_n (FK → data_individuals)
├─ id_specimen (FK → specimens)
├─ id_linktype (FK → linktypelist)
├─ created_by (audit)
└─ created_at (audit)

Troubleshooting

“No specimens found to retrieve”

Cause: The herbarium_nbe_char column is empty or unparseable.

Solution: - Check that your data has specimen references in the correct format - Use format: “Collector Name ####” (e.g., “Dauby 123”)

“Validation failed: Missing specimen IDs”

Cause: The specimens don’t exist in the herbarium database yet.

Solution: - Ensure specimens are entered in the specimens table first - Check collector name spelling matches table_colnam

“Taxonomic match: Different family”

Cause: Field identification and specimen identification are very different.

Solution: - This is often a real discrepancy worth investigating - Check field notes and specimen label - Consult with taxonomists before creating the link

Best Practices

  1. Link specimens as soon as they’re databased: Don’t wait years to establish connections.

  2. Review taxonomic discrepancies carefully: Different families often indicate either:

    • Field misidentification (common)
    • Wrong specimen linked (data entry error)
    • Mislabeled specimen (rare)

  3. Document link types accurately: Distinguish between direct collection (type_individual) and field-identified references (referenced_individual).

  4. Maintain audit trails: Never delete links; add new ones if corrections are needed.

  5. Regular updates: Periodically query for new specimen identifications to ensure your inventory data stays current.

Summary

The specimen linking system transforms static inventory data into a living dataset that improves over time as taxonomic knowledge advances. By creating formal database-level connections between ecological measurements and herbarium vouchers, you ensure that your research benefits from ongoing taxonomic revision—automatically, permanently, and transparently.

This is not just data management; it’s a paradigm shift toward cumulative, self-improving ecological data infrastructure.