From Data Deluge to Discovery: Modern Archiving and Sharing Strategies for Biologging Data

Abstract

This article provides a comprehensive guide for researchers and scientists on managing the entire lifecycle of biologging data. It covers foundational principles and the critical importance of data preservation, explores modern archiving methods and platforms like the Biologging intelligent Platform (BiP) and Movebank, addresses common challenges in data standardization and machine learning validation to prevent overfitting, and presents best practices for ensuring data quality and interoperability. The content synthesizes the latest 2025 research and community standards to equip professionals with the strategies needed to maximize the value of biologging data for ecological discovery, conservation, and biomedical applications.

Why Biologging Data Preservation is a Cornerstone of Modern Science

Defining Biologging Data Archiving and Its Strategic Value

Biologging, the use of animal-borne electronic tags to document movements, behaviour, physiology, and environments of wildlife, has experienced rapid growth [1]. This growth has led to an unprecedented accumulation of complex data, creating a critical need for systematic archiving. Biologging data archiving is not merely the storage of data files; it is the process of preserving sensor-derived data and its associated metadata in a standardized, accessible, and reusable format for long-term use. This practice ensures that these unique datasets, which capture animal life on Earth, are safeguarded against loss and remain available for future scientific discovery, policy-making, and conservation efforts [2] [1].

The value of these data extends far beyond their initial research purpose. They form dynamic digital archives of natural history, capturing vital information about animals in the context of their changing environments [1]. As the field expands, a strategic approach to archiving becomes essential to mitigate biodiversity threats and maximize the return on investment from costly and logistically challenging biologging studies.

The Strategic Value of Archiving Biologging Data

Effective archiving transforms biologging data from a static record into a powerful, reusable resource with multi-faceted strategic importance.

Enhancing Scientific Research and Collaboration

• Facilitating Meta-Analyses and Large-Scale Studies: Standardized, archived data enables researchers to integrate results from multiple studies and species to uncover broad ecological patterns. This has already empowered international collaborations and studies involving dozens of co-authors [2].
• Promoting Open Science and Reproducibility: Publicly accessible archives under licenses like CC BY 4.0 allow for the verification of results and foster a culture of transparency and collaboration [2].
• Enabling Cross-Disciplinary Innovation: Biologging data are increasingly valuable to fields beyond biology, such as oceanography and meteorology [2] [3]. For instance, data from marine animals provides critical oceanographic data in regions inaccessible to traditional buoys or satellites, improving the predictive accuracy of environmental models [2] [3].

Supporting Conservation and Policy

Archived biologging data provides objective, evidence-based insights crucial for conservation.

• Informing Protected Area Design: Movement data can identify critical habitats and migration corridors, enabling the design of effective Marine Protected Areas (MPAs) [3].
• Understanding Anthropogenic Threats: Data on animal interactions with fishing gear, marine debris, and shipping lanes helps quantify threats like bycatch and informs mitigation policies [3].
• Monitoring Biodiversity and Ecosystem Health: Long-term archives serve as a baseline to track species' responses to climate change and other environmental pressures [1] [4].

Ensuring Long-Term Data Preservation and Legacy

Biologging studies represent a significant investment of resources and often capture irreplaceable information about animal life in a specific time and place. Archiving ensures this "digital natural history" is preserved for future generations, protecting it from data loss due to hardware failure, format obsolescence, or simply the passage of time [2] [1]. This aligns with the view of biodiversity itself as a data system, where each species holds irreplaceable information refined over millennia [4].

Maximizing Return on Investment

By making data Findable, Accessible, Interoperable, and Reusable (FAIR), archiving platforms maximize the scientific and societal return on the substantial financial and human investment required for biologging studies [2] [1]. It prevents redundant research and allows secondary users to extract new value from existing data without the cost of new field deployments.

Table 1: Strategic Value of Biologging Data Archiving

Strategic Area	Key Benefits	Example
Scientific Research	Enables large-scale meta-analyses; Fosters cross-disciplinary discovery; Improves model accuracy.	Using seal-collected data to map Antarctic Circumpolar Current fronts [2].
Conservation & Policy	Informs design of protected areas; Quantifies anthropogenic threats; Tracks climate change impacts.	Mitigating fisheries bycatch using data on animal movement in shipping lanes [3].
Data Stewardship	Preserves irreplaceable data for future generations; Prevents data loss; Ensures research legacy.	Creating a "dynamic archive of animal life on Earth" for long-term use [1].
Economic Efficiency	Maximizes ROI from expensive deployments; Prevents redundant studies; Enables secondary data use.	Shared datasets allowing new research questions without new field work [2].

Key Components of a Biologging Data Archive

A robust biologging archive is built on interconnected pillars of data, metadata, and community standards.

Sensor Data and Standardization

The primary data from biologging devices can include location, depth, acceleration, water temperature, and more. A major challenge is the heterogeneity of data formats across sensor types and manufacturers. Strategic archiving requires data standardization, which involves converting data into consistent formats using common column names, standardized date-time formats (e.g., ISO 8601), and uniform file structures [2]. This step is critical for enabling data integration and automated analysis across datasets.

Metadata: The Backbone of Reusability

Sensor data alone are meaningless without context. Metadata—data about the data—provides this essential context and includes three key categories:

• Animal Metadata: Species, sex, body size, and life history.
• Instrument Metadata: Device type, manufacturer, sensor specifications, and calibration data.
• Deployment Metadata: Who deployed the device, when and where the deployment occurred, and the retrieval method [2].

To ensure interoperability, metadata should conform to international standards and vocabularies, such as the Integrated Taxonomic Information System (ITIS) and Climate and Forecast (CF) Metadata Conventions [2].

Platform Infrastructure and Services

An archiving platform is more than a storage drive. It should provide:

• User Management: Registration systems for data upload and access control.
• Visualization Tools: Basic maps and profiles to explore data before download.
• Analytical Processing (OLAP): Integrated tools to calculate derived parameters, such as estimating ocean currents from animal movement or classifying behaviours from acceleration data [2].
• Persistent Identifiers: Assignment of Digital Object Identifiers (DOIs) to link datasets directly to scientific publications [2].

Diagram 1: Biologging data archiving workflow

Protocols for Effective Biologging Data Archiving

The following protocols provide a roadmap for researchers to prepare and submit data to an archive, and for archives to manage that data effectively.

Protocol 1: Data Preparation and Submission for Researchers

This protocol outlines the steps a researcher should take from data retrieval to successful archive submission.

1. Pre-Deployment Planning:

• Action: Before deploying devices, consult the target archive's metadata requirements.
• Rationale: Proactive planning ensures all necessary contextual data (e.g., animal measurements, precise deployment coordinates) are collected at the time of deployment, preventing information loss.

2. Data Retrieval and Initial Backup:

• Action: Download raw data from devices and create multiple secure, backed-up copies in original format.
• Rationale: Preserves the primary data before any transformation or potential corruption during processing.

3. Data Standardization and Quality Control:

• Action: Convert raw data to a standardized format. Adhere to the archive's specified column names, units, and date-time format (preferably ISO 8601). Perform basic quality checks for sensor errors or impossible values.
• Rationale: This is the most critical step for enabling data interoperability and reuse. Consistent formatting allows other researchers and analytical tools to parse the data correctly.

4. Metadata Compilation:

• Action: Systematically complete all three categories of metadata (Animal, Instrument, Deployment) using the archive's templates. Utilize pull-down menus for standardized terms (e.g., species names) where available to minimize typos [2].
• Rationale: Comprehensive and accurate metadata is what transforms a data file into a meaningful and reusable scientific resource.

5. Submission and Access Level Setting:

• Action: Upload the standardized data and metadata files through the archive's portal. Choose an appropriate sharing license (e.g., CC BY 4.0) and set the dataset to "Open" or "Private" based on embargo needs.
• Rationale: Formal submission deposits the data into the custodianship of the archive. Setting the correct access level balances the need for immediate open access with situations requiring temporary privacy, such as until publication.

Protocol 2: Data Curation and Preservation for Archives

This protocol describes the responsibilities of the data archive in managing and preserving submitted data.

1. Data Ingestion and Validation:

• Action: Upon submission, run automated checks to verify file integrity and confirm that data and metadata conform to the platform's schema and standards.
• Rationale: Ensures the quality and consistency of all data entering the archive, preventing the inclusion of corrupt or non-compliant datasets.

2. Curation and Annotation:

• Action: Curators may enhance the dataset by linking it to broader ontologies or cross-referencing it with other environmental datasets. They may also assist in anonymizing sensitive location data for threatened species.
• Rationale: Adds value to the submission, increasing its interoperability and reusability (the "I" and "R" in FAIR).

3. Secure Archiving and Preservation:

• Action: Store data in a secure, resilient, and geographically redundant system. Implement a digital preservation strategy that includes regular integrity checks and plans for format migration as technology evolves.
• Rationale: Fulfills the long-term preservation mission of the archive, safeguarding data against loss, corruption, or technological obsolescence.

4. Distribution and Access Management:

• Action: Provide user-friendly tools for searching, visualizing, and downloading data. Manage access controls for private datasets and facilitate communication between data owners and potential users.
• Rationale: Makes the data accessible to the community, fulfilling the core purpose of sharing and reuse.

5. Integration and Sustainability:

• Action: Enable data exchange with other relevant repositories (e.g., Movebank, GenBank) to increase reach and utility. Secure sustainable funding models to ensure the archive's long-term operation.
• Rationale: Promotes a connected data ecosystem and ensures the archive itself remains a permanent resource.

Table 2: Repository Types for Biologging Data

Repository Type	Focus	Example Platforms	Key Characteristics
Domain-Specific	Biologging and animal movement data.	Movebank, Biologging intelligent Platform (BiP)	High level of community engagement; data curation specific to biologging; supports complex sensor data types [2] [5].
Generalist	Data from any scientific discipline.	Zenodo, Dryad	Accepts data regardless of type or discipline; less specialized curation but broad visibility [5].
Institutional	Data output from a specific institution.	University Data Repositories	Serves the institution's staff; preservation may be tied to the institution's lifespan [5].
Project-Specific	Data from a specific large-scale project or collaboration.	AniBOS (Animal Borne Ocean Sensors)	Focused on a project's goals; enables data sharing within a defined scope and for reuse [2] [5].

The Scientist's Toolkit: Research Reagent Solutions

Successful biologging research and data archiving rely on a suite of "research reagents"—both physical and digital.

Table 3: Essential Materials and Tools for Biologging Research

Category	Item	Function
Hardware & Field Equipment	Satellite Relay Data Loggers (SRDLs)	Transmits compressed data (e.g., dive profiles, temperature) via satellite, eliminating need for recapture [2].
	GPS & Argos Transmitters	Provides animal location data.
	Bio-Logging Tags (Accelerometers, Depth Sensors)	Records fine-scale behaviour (e.g., flipper strokes, dive profiles) and environmental data [2] [3].
Software & Digital Tools	R / Python with Movement Ecology Packages	For processing, visualizing, and analyzing complex biologging data.
	Online Analytical Processing (OLAP) Tools	Integrated in platforms like BiP to calculate environmental (e.g., surface currents) or behavioural parameters from raw data [2].
Data Standards & Vocabularies	Integrated Taxonomic Information System (ITIS)	Provides standardized taxonomic names, ensuring consistency in species identification across datasets [2].
	Climate and Forecast (CF) Metadata Conventions	A standard for encoding metadata for earth science data, promoting interoperability [2].
Data Platforms & Infrastructure	Biologging intelligent Platform (BiP)	An integrated platform for sharing, visualizing, and analyzing standardized biologging data [2].
	Movebank	A global data repository for animal tracking data, supporting a wide range of taxa and sensor types [2].

Diagram 2: Biologging data archiving ecosystem

Biologging data archiving is a strategic imperative, not an administrative afterthought. It is the foundation for a sustainable and impactful biologging research ecosystem. By defining and implementing robust archiving practices—centered on standardization, rich metadata, and dedicated platform infrastructure—the scientific community can fully leverage the immense value of these unique datasets. This approach ensures that biologging data continues to drive scientific discovery, inform effective conservation policies, and preserve a digital legacy of life on Earth for generations to come. The vision of a globally integrated network, such as the Internet of Animals (IoA), where data flows seamlessly from animals to researchers to decision-makers, depends entirely on the foundations of a strong, collaborative archiving culture [3] [1].

The Critical Role of Data Sharing in Collaborative Research and Conservation

The expanding field of biologging, which involves attaching data recorders to animals to monitor their behavior, physiology, and environment, is generating unprecedented amounts of critical data [2]. This data is invaluable not only for understanding animal ecology but also for secondary applications in oceanography, meteorology, and environmental science [3]. However, the full potential of this data can only be realized through systematic archiving and sharing approaches that ensure data accessibility, standardization, and interoperability. Effective data sharing practices are fundamental to collaborative research and biological conservation, enabling the mapping of animal distributions and movements essential for informed conservation strategies [2]. This application note outlines standardized protocols and platforms for biologging data management, facilitating broader collaboration and data reuse across scientific disciplines.

The following tables summarize key quantitative aspects and metadata requirements for effective biologging data sharing, enabling easy comparison of platform capabilities and data structuring principles.

Table 1: Key Platform Capabilities for Biologging Data Management

Platform Name	Primary Function	Data Types Supported	Metadata Standards	Access Protocol
Biologging intelligent Platform (BiP)	Data storage, standardization, analysis, and sharing [2]	Sensor data (location, depth, temperature, acceleration), animal traits, deployment info [2]	ITIS, CF, ACDD, ISO [2]	CC BY 4.0 for open data; request required for private data [2]
protocols.io	Protocol sharing and peer review [6]	Step-by-step experimental and analytical methods [6]	Integrated submission systems (e.g., Nature journals) [6]	Free-to-read, open access CC-BY; private during peer review [6]
Movebank	Biologging data management [2]	Primarily location data (7.5 billion points as of 2025) [2]	Not specified in search results	Not specified in search results

Table 2: Essential Metadata for Biologging Data Reusability

Metadata Category	Specific Elements	Standardization Importance
Animal Traits	Sex, body size, breeding history [2]	Enables analysis of individual differences on movement and behavior [2]
Instrument Details	Device type, manufacturer, sensor specifications [2]	Ensures proper interpretation of data quality and limitations [2]
Deployment Information	Who, when, where, and how deployed [2]	Provides critical context for data interpretation and reuse [2]
Data Collection Parameters	Sampling rates, calibration information, data formats [2]	Facilitates data integration and comparative analyses [2]

Experimental Protocols

Protocol for Standardized Data Submission to Biologging intelligent Platform (BiP)

Background: Inconsistent data formats—such as different column names for the same sensor data, variations in date-time formats, and differing file types—create significant barriers to collaborative research and secondary use of biologging data [2]. The BiP platform addresses this challenge through a standardized submission process.

Materials:

• Biologging datasets from animal-borne devices
• Associated metadata on animal subjects, instruments, and deployment conditions
• Computer with internet access
• BiP user account (registered at https://www.bip-earth.com)

Procedure:

• User Registration:
  • Visit the BiP website (https://www.bip-earth.com)
  • Complete the user registration form with institutional details and research interests

• Metadata Preparation:

  • Compile animal trait information using standardized vocabularies (e.g., species names from ITIS)
  • Document instrument specifications including manufacturer, model, and sensor types
  • Record deployment details (date, location, method)
  • Utilize BiP's pull-down menus to ensure consistency and minimize entry errors [2]

• Data Standardization:

  • Format sensor data columns according to BiP conventions
  • Ensure date-time values use ISO8601 format (YYYY-MM-DD)
  • Verify consistent precision in numerical values
  • Use standardized column headers (e.g., "lat" for latitude)

• Data Upload:

  • Select "Upload New Dataset" in user dashboard
  • Upload sensor data file (CSV format recommended)
  • Input required metadata through structured forms
  • Select sharing preference (open or private)

• Quality Control:

  • Review automated data validation reports
  • Correct any formatting errors identified by the system
  • Confirm metadata completeness

• Publication:

  • For open data: Set license to CC BY 4.0
  • For private data: Configure access permissions
  • Finalize submission to make data discoverable

Troubleshooting:

• If date formats are rejected, convert to ISO8601 format (YYYY-MM-DD)
• If species names are not recognized, use ITIS taxonomic identifiers
• For large file upload failures, check internet stability and file size limits

Protocol for Collaborative Protocol Sharing via protocols.io

Background: Clear, detailed methodologies are essential for research reproducibility. Integrating protocol sharing into the publication process enhances transparency and enables method improvement [6].

Materials:

• Detailed experimental or analytical method
• Figures, videos, or computational code (if applicable)
• protocols.io account (free)

Procedure:

• Protocol Drafting:
  • Create a new protocol on protocols.io
  • Enter step-by-step instructions with necessary granularity
  • Incorporate images, videos, or computational methods where helpful
  • Use the platform's collaborative features for co-author editing

• Platform Integration:

  • For Nature Cell Biology submissions: Select protocol deposition during manuscript submission [6]
  • Upload protocol to protocols.io and link directly to submission
  • The protocol remains private during peer review, maintaining referee anonymity [6]

• Version Management:

  • Utilize version history to track changes
  • Enable concurrent editing by several users for collaboration
  • Note that protocols cannot be edited while manuscript is under review [6]

• Publication and Linking:

  • Upon article acceptance, protocol becomes publicly available
  • A 'peer reviewed' badge is added if protocol was reviewed with the paper [6]
  • Protocol receives a DOI and is permanently linked to the published paper

Troubleshooting:

• If manuscript is not accepted, unlink the protocol, edit it, and include the reserved DOI in submissions to other journals [6]
• For technical issues, consult the protocols.io help center

Mandatory Visualization

Biologging Data Sharing Workflow

Biologging Data Reuse Applications

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Tools for Biologging Data Management and Analysis

Tool/Platform	Primary Function	Key Features	Application Context
Biologging intelligent Platform (BiP)	Integrated data storage, standardization, and analysis [2]	OLAP tools for environmental parameter calculation, standardized metadata, CC BY 4.0 licensing [2]	Cross-disciplinary research, environmental monitoring, collaborative studies [2]
protocols.io	Protocol development, sharing, and peer review [6]	Version history, collaborative editing, integration with journal submission systems [6]	Method documentation, research reproducibility, protocol peer review [6]
Movebank	Biologging data management [2]	Large-scale data storage (7.5+ billion location points), multi-taxa support [2]	Animal movement studies, migration analysis, distribution mapping [2]
AniBOS (Animal Borne Ocean Sensors)	Global ocean observation using animal-borne sensors [2]	Physical environmental data collection, integration with Argo float data [2]	Oceanographic research, climate studies, polar region monitoring [2]
Satellite Relay Data Loggers (SRDL)	Remote data collection and transmission from marine animals [2]	Data compression, satellite transmission, year+ operation [2]	Marine animal tracking, polar region oceanography, inaccessible area monitoring [2]

Biologging research, which uses animal-borne electronic tags to collect data on movement, behavior, physiology, and environment, generates complex datasets that intersect with evolving data protection regulations. For researchers, scientists, and drug development professionals, navigating the compliance landscape is essential for both scientific integrity and legal adherence. The International Bio-Logging Society emphasizes that proper data management maximizes scientific insight and conservation benefits while minimizing impacts on study subjects [7]. Furthermore, funding agencies like the National Institutes of Health (NIH) now require researchers to plan for data management and sharing, submitting formal Data Management and Sharing Plans (DMSPs) with grant applications [8]. This document provides application notes and protocols to align biologging data archiving and sharing with current regulatory requirements, including FAIR principles (Findable, Accessible, Interoperable, Reusable) and specific data retention mandates emerging across U.S. states and federal agencies [9].

Current Regulatory Framework and Quantitative Requirements

Key Data Protection Regulations

The regulatory environment for data protection is rapidly evolving, with significant implications for biologging data that may involve human researchers, location data, or sensitive species information. Several key laws and principles govern this space:

• FAIR Principles: Scientific communities are increasingly adopting FAIR principles to make data easy to "discover, access, interoperate, and sensibly re-use, with proper citation" [9]. For biologging data, this means implementing standard vocabularies, transfer protocols, and aggregation protocols to enable data integration [10].
• NIH Data Sharing Policy: The NIH requires researchers to plan and budget for Data Management and Sharing (DMS), submit a DMSP with grant applications, and comply with the approved plan [8].
• Children's Online Privacy: While less directly applicable to animal research, the trend toward protecting minors' data (e.g., laws in California, Connecticut, Maryland, Nebraska, and Vermont) highlights the regulatory emphasis on privacy that could extend to human-related biologging data [11].
• Biometric Data Regulations: Colorado's updated Privacy Act now covers "Biometric Identifiers," defined as "data generated by the technological processing, measurement, or analysis of an individual's biological, physical, or behavioral characteristics" [12]. This could potentially intersect with certain biologging data types.

Data Retention and Protection Assessment Requirements

Recent state laws have introduced specific requirements for data protection assessments and documentation:

Table 1: Selected 2025 U.S. State Data Protection Assessment Requirements

State/Jurisdiction	Requirement	Effective Date	Key Provisions
Delaware	Data Protection Assessment	July 1, 2025	Required for processing activities presenting heightened risk of harm, including sensitive data processing, targeted advertising, and profiling [12].
Colorado	Biometric Identifiers Policy	July 1, 2025	Written policy required including retention schedules, data incident response, and data deletion procedures [12].
Department of Justice	Bulk Sensitive Data Rule	July 8, 2025	Restrictions on transfer of sensitive data to certain countries; requires demonstration of "good faith efforts" to comply [12].

Experimental Protocols for Compliant Biologging Data Management

Data Standardization and Metadata Protocol

Objective: To create standardized biologging datasets that comply with interoperability requirements and facilitate regulatory compliance.

Materials:

• Biologging data (sensor and positional data)
• Metadata on animal subjects, instruments, and deployments
• Biologging intelligent Platform (BiP) or Movebank database access
• Computational resources for data transformation

Procedure:

• Extract raw data from biologging devices and associated metadata, including:
  • Animal traits (species, sex, body size, health status)
  • Instrument details (device type, manufacturer, firmware version)
  • Deployment information (location, date, attachment method) [2]

• Apply standardization formats:

  • Convert data to standardized column names and formats following international conventions such as Integrated Taxonomic Information System (ITIS), Climate and Forecast Metadata Conventions (CF), and Attribute Conventions for Data Discovery (ACDD) [2].
  • Ensure date-time formats follow ISO8601 standard (YYYY-MM-DD HH:MM:SS).
  • Use consistent taxonomic nomenclature throughout datasets.

• Implement quality control checks:

  • Verify data ranges for all sensor readings (e.g., realistic depth, temperature values).
  • Confirm spatial coordinates use consistent datum references.
  • Validate metadata completeness against required fields.

• Upload to standardized platform:

  • Use BiP's interactive tools to upload sensor data and associated metadata [2].
  • Or prepare data for Movebank using tools like movepub software [13].
  • Select appropriate sharing settings (open or private) based on regulatory and ethical considerations.

• Document the standardization process for compliance reporting, noting any data transformations, quality issues, or exceptions to standard protocols.

Data Protection Assessment Protocol for Sensitive Biologging Data

Objective: To conduct and document formal Data Protection Assessments (DPA) as required by state privacy laws for processing activities that present heightened risks.

Materials:

• Inventory of biologging data types and processing activities
• Data classification guidelines
• Risk assessment framework
• Documentation templates

Procedure:

• Inventory processing activities:
  • Catalog all biologging data collection, storage, and sharing practices.
  • Identify data flows to external collaborators, repositories, or public platforms.
  • Document the purposes for each processing activity.

• Determine assessment requirements:

  • Flag processing activities involving sensitive data (e.g., endangered species locations, proprietary methodologies).
  • Identify activities that could present heightened risks (e.g., profiling animal behavior for commercial applications, sharing data with international collaborators) [12].

• Conduct risk assessment:

  • Evaluate potential risks to individuals, organizations, or conservation efforts.
  • Assess likelihood and severity of potential data breaches or misuse.
  • Consider risks of re-identification even from anonymized datasets.

• Identify mitigation measures:

  • Implement technical safeguards (encryption, access controls).
  • Establish data retention schedules with specific destruction timelines.
  • Develop procedures for secure data sharing, including data use agreements.

• Document the assessment:

  • Record the assessment process, findings, and mitigation measures.
  • Maintain documentation for regulatory review, updating when processing activities change significantly.
  • For Delaware controllers, ensure DPA covers processing of ≥100,000 consumer records [12].

Visualization of Compliant Biologging Data Workflow

The following diagram illustrates the integrated workflow for managing biologging data in compliance with regulatory requirements:

Compliant Biologging Data Workflow: This diagram outlines the pathway from raw data collection to compliant archiving and sharing, showing key stages where regulatory requirements influence data management decisions.

Research Reagent Solutions for Compliant Biologging Research

Table 2: Essential Tools and Platforms for Biologging Data Compliance

Tool/Platform	Function	Compliance Application
Biologging intelligent Platform (BiP)	Integrated platform for sharing, visualizing, and analyzing biologging data [2].	Standardizes data formats and metadata following international conventions; enables controlled data sharing with permission workflows.
Movebank	Global repository for animal tracking data with over 7.5 billion location points [10].	Supports data preservation through public archiving; facilitates FAIR compliance through standardized data protocols.
movepub (R package)	Software tool to prepare Movebank data for publication [13].	Automates data standardization and documentation processes to meet reproducibility requirements.
ETN R package	Access data from the European Tracking Network [13].	Enables interoperable data reuse across research networks while maintaining data integrity.
ESS-DIVE Repository	Department of Energy repository for environmental research data [9].	Provides long-term stewardship for modeling data; implements FAIR principles for terrestrial and biologging data.
Data Protection Assessment Templates	Structured frameworks for evaluating data processing risks [12].	Documents compliance with state privacy laws (DE, CO); demonstrates due diligence for sensitive data processing.

Navigating data retention and regulatory requirements in biologging research requires both technical solutions and institutional commitment. By implementing the standardized protocols, visualization workflows, and toolkits outlined in these application notes, researchers can build compliance into their data management lifecycle rather than treating it as an afterthought. The International Bio-Logging Society's Data Standardisation Working Group continues to develop community standards that both advance scientific goals and address regulatory requirements [13]. As data protection laws continue to evolve, establishing these foundational practices will position biologging researchers to adapt efficiently to new requirements while maximizing the scientific and conservation value of their data.

Bio-logging involves the use of animal-borne electronic tags, or "bio-loggers," to record data about an animal's movements, behavior, physiology, and the environment it experiences [14]. The rapid growth of this field is generating unprecedented volumes of data, offering profound opportunities for ecological research, biodiversity conservation, and environmental monitoring [14] [2] [15]. This application note establishes a standardized protocol for managing the complete lifecycle of bio-logging data, treating these data streams as dynamic digital archives of animal life on Earth [14] [16]. Adhering to this lifecycle is critical for ensuring the long-term scientific value, accessibility, and ethical reuse of these complex datasets.

The Bio-Logging Data Lifecycle

The bio-logging data lifecycle encompasses a series of interconnected stages, from strategic planning and data collection to final archiving and reuse. The following workflow diagram visualizes this comprehensive process and the logical relationships between its key stages.

Stage 1: Project Planning & Data Collection

Effective data management begins before any device is deployed. This stage focuses on defining research objectives and selecting appropriate methodologies to ensure the collection of high-quality, ethically-sound data.

Experimental Protocol: Pre-Deployment Planning

Objective: To define research goals, obtain necessary permits, and select appropriate bio-logging devices and attachment methods that minimize impact on animal welfare [14].

Materials:

• Research proposal and defined hypotheses.
• Ethical review and animal care permits.
• Field permits and land/water access permissions.
• Selection of bio-logging devices (see Section 3.2).

Methodology:

• Define Research Questions: Clearly articulate the primary biological, ecological, or environmental questions. This determines the required sensor types (e.g., GPS, accelerometer, temperature, heart rate), sampling frequency, and deployment duration [15].
• Conduct Ethical Review: Submit a detailed proposal to the relevant Institutional Animal Care and Use Committee (IACUC) or equivalent ethics board. The principles of Replacement, Reduction, and Refinement (the 3Rs) should guide the study design [14].
• Select Devices: Choose devices that are typically <5% of the animal's body mass to minimize burden [14]. Consider the trade-offs between data storage capacity, sampling rate, battery life, and the need for data recovery (retrieval vs. satellite transmission).
• Plan Data Collection: Develop a standardized data dictionary and metadata template (see Section 4.1) before deployment to ensure consistent data recording across multiple researchers or field sites [2].

Research Reagent Solutions: Essential Materials

Table 1: Key tools and platforms for bio-logging data collection and management.

Item Name	Type/Model Examples	Function & Application
Satellite Relay Data Loggers (SRDL)	SMRU Instrumentation SRDLs	Transmit compressed data (e.g., dive profiles, temperature) via satellite; ideal for marine mammals in remote regions [2].
GPS & Accelerometer Tags	Various manufacturers (e.g., TechnoSmart, Ornitela)	Record high-resolution location and tri-axial acceleration data for reconstructing movement paths and classifying behavior [15].
Archival Tags (Data Loggers)	Lotek LAT, Star-Oddi DST	Store data internally for later retrieval; used when high-resolution data is needed and animal recapture is feasible [2].
Bio-Logging Data Platforms	Movebank, Biologging intelligent Platform (BiP), WRAM	Web-based infrastructures for managing, storing, standardizing, and sharing bio-logging data during and after collection [14] [2].
Time-Series Analysis Software	R Statistical Environment (packages: nlme, lme4, glmmTMB)	Statistical toolkits for modeling autocorrelated physiological and movement data from bio-loggers [15].

Stage 2: Data Processing & Standardization

Raw data from bio-logging devices are often not analysis-ready. This stage involves transforming raw data into a structured, standardized, and annotated format.

Experimental Protocol: Data Curation and Standardization

Objective: To clean, calibrate, and annotate raw sensor data with comprehensive metadata, creating a standardized dataset ready for ecological analysis or integration with larger data collections [2].

Materials:

• Raw data files from bio-logging devices.
• Metadata on animal traits, deployment details, and sensor calibration.
• Data standardization platform (e.g., BiP, Movebank).

Methodology:

• Data Ingestion and Decryption: Download data from devices or satellite networks. Decrypt or decompress data using manufacturer-specific software.
• Data Cleaning and Calibration:
  • Filtering: Remove clear outliers (e.g., implausible GPS fixes based on speed).
  • Sensor Calibration: Apply calibration coefficients to convert raw sensor readings (e.g., voltage) into physical units (e.g., °C, m/s²). For accelerometers, calibrate to animal body frame [15].
• Metadata Compilation: Assemble metadata using a standardized framework [2] [13]. Critical components include:
  • Animal Metadata: Species (using ITIS taxonomy), sex, age, mass, breeding status.
  • Deployment Metadata: Deployment date/time, location, attachment method, retrieval date.
  • Instrument Metadata: Device type, serial number, firmware version, sampling schedules.
• Data Standardization: Format the data and metadata according to community standards, such as those proposed by the International Bio-logging Society [13]. This includes using standard column names (e.g., "location-lat"), date-time formats (ISO 8601), and vocabularies.

Stage 3: Data Analysis & Modeling

This stage involves applying statistical and computational models to extract biological and environmental insights from the standardized data.

Experimental Protocol: Analyzing Time-Series Data

Objective: To analyze bio-logging time-series data (e.g., heart rate, depth, acceleration) using appropriate statistical models that account for temporal autocorrelation and complex data structures [15].

Materials:

• Standardized and curated dataset.
• Statistical software (e.g., R, Python).
• Defined research hypotheses.

Methodology:

• Exploratory Data Analysis (EDA): Visualize the data to identify patterns, trends, and potential outliers. Use periodograms to identify cyclical patterns (e.g., diel or tidal cycles).
• Model Selection: Choose a model that accounts for the inherent temporal autocorrelation in the data. Never use simple t-tests or ordinary linear models on raw time-series data, as this greatly inflates Type I error rates [15].
• Model Fitting: Implement a model with a correlation structure. For example:
  • Generalized Least Squares (GLS): Can incorporate an autoregressive (AR) correlation structure to model dependence between consecutive observations [15].
  • Mixed-Effects Models: Allow the inclusion of random effects (e.g., individual animal ID) to account for multiple observations from the same individual.
• Model Validation: Check model assumptions by analyzing the residuals. Ensure residuals are homoscedastic and show no remaining patterns. Use Akaike Information Criterion (AIC) for model comparison.
• Biological Inference: Interpret model coefficients and predictions in the context of the original biological question.

Data Visualization Guidelines

Effective visualization is key to interpreting and communicating results. The following table summarizes best practices for colorizing biological data visualizations.

Table 2: Color application rules for biological data visualization based on data type [17] [18].

Data Type (Measurement Scale)	Description & Examples	Recommended Color Scheme	Rationale & Tips
Nominal/Categorical	Categories without order (e.g., species, behavior type).	Qualitative (distinct hues).	Use distinct colors with easily names hues (e.g., Red, Blue, Green). Avoid fine hue differentiations.
Ordinal	Categories with order but unknown intervals (e.g., low, medium, high).	Sequential (light to dark).	Lightness (Luminance) should increase or decrease with the order. Use a single hue or small set of adjacent hues.
Interval/Ratio (Quantitative)	Numerical values with meaningful intervals (e.g., temperature, depth, speed).	Sequential or Diverging.	Sequential: For data from low to high. Diverging: To highlight deviation from a neutral mid-point (e.g., Purple-Green scheme) [18].
Guidelines for All Types			Check for Deficiencies: Use tools (e.g., Coblis) to simulate Protanopia, Deuteranopia, Tritanopia. Ensure Contrast: Text and symbols must have high contrast against background colors. Use Perceptually Uniform Spaces: Prefer HCL or CIE L*a*b* over RGB [17].

Stage 4: Data Archiving & Sharing

Preserving data in a public, trusted archive ensures its long-term value and supports scientific reproducibility and collaboration.

Experimental Protocol: Preparing Data for Public Archiving

Objective: To deposit curated data and metadata into a suitable long-term repository, making it Findable, Accessible, Interoperable, and Reusable (FAIR) while respecting the CARE principles for Indigenous data governance [14] [19].

Materials:

• Final, cleaned, and standardized dataset.
• Complete metadata record.
• Documentation file (README) describing the dataset.
• Chosen data repository.

Methodology:

• Repository Selection: Choose a recognized repository suitable for bio-logging data, such as Movebank, the Biologging intelligent Platform (BiP), or a government data centre (e.g., Australian Antarctic Data Centre) [14] [2] [19]. The repository should provide a persistent identifier (e.g., DOI).
• FAIR Compliance Check: Verify that the data package meets FAIR principles:
  • Findable: Rich metadata with keywords and a unique persistent identifier.
  • Accessible: Data retrievable by their identifier using a standardized protocol.
  • Interoperable: Data and metadata use formal, accessible, and shared languages and vocabularies.
  • Reusable: Data have multiple, accurate and relevant attributes, a clear usage license, and provenance information [14] [19].
• Apply Usage License: Attach a standard, permissive license to the data, such as CC-BY 4.0, which requires attribution but allows reuse and modification [2].
• Upload and Publish: Upload the data and metadata package to the repository. Once published, the persistent identifier (DOI) should be cited in any related research publications [20].

Integrated Data Platforms and Future Directions

The future of bio-logging relies on integrated data collections that function as dynamic archives. Platforms like Movebank and BiP are leading this effort by providing standardized tools for the entire data lifecycle [14] [2]. BiP's unique integration of Online Analytical Processing (OLAP) tools allows users to calculate environmental parameters, such as surface currents and ocean winds, directly from animal movement data, showcasing the potential for cross-disciplinary secondary use [2].

A critical next step is the establishment of global governance, such as the community-led coordinating body proposed by the International Bio-logging Society, to oversee data standards and ensure the sustainable preservation of these invaluable digital archives of animal life [14] [13]. Widespread adoption of these protocols by researchers, coupled with support from funders, publishers, and data repositories, is essential to fully realize the potential of bio-logging data for addressing fundamental ecological questions and mitigating biodiversity threats.

Biologging, the practice of equipping animals with electronic data loggers, has transcended its origins in behavioral ecology to become a critical source of high-resolution environmental data. This article outlines standardized protocols for repurposing these animal-borne sensor data for applications in oceanography, meteorology, and biomedicine. Framed within the urgent need for robust archiving and sharing approaches, we detail methodologies for data collection, processing, and integration, providing a framework for maximizing the scientific return from biologging investments.

Initially developed to monitor the behavior and physiology of wild animals in their natural habitats, biologging is a Lagrangian observation method that moves with the animal, providing a unique mobile platform for data collection [3]. The field has evolved from basic tracking to the deployment of sophisticated suites of sensors that measure a host of environmental parameters. The resulting data, when properly archived and shared, present a transformative opportunity for secondary use in disparate scientific disciplines. These secondary uses are contingent upon the establishment of FAIR (Findable, Accessible, Interoperable, and Reusable) data principles and community-wide adoption of standardized protocols [10]. This document provides application notes and experimental protocols for leveraging biologging data, emphasizing its role in a broader data-sharing ecosystem essential for addressing global scientific challenges.

Application Notes & Protocols

The secondary use of biologging data follows a structured pipeline, from sensor deployment on animals to the final integration of validated data into cross-disciplinary models.

Oceanographic Data Acquisition Protocol

Objective: To collect in-situ oceanographic data (e.g., temperature, salinity) from marine animals to complement traditional ocean observation systems like Argo floats.

• Step 1: Sensor Selection and Calibration

  • Pressure Sensor (Depth): Must be calibrated for the expected depth range of the target species.
  • Temperature & Salinity Sensors: Should be cross-calibrated against Conductivity-Temperature-Depth (CTD) sensors from research vessels or Argo floats prior to deployment to ensure data fidelity [2].

• Step 2: Animal Deployment

  • Target Species: Select species based on habitat. Phocid seals are ideal for polar regions; sea turtles, sharks, and large fish provide coverage in temperate and tropical waters, and in shallow areas inaccessible to Argo floats [3] [2].
  • Attachment: Deploy Satellite Relay Data Loggers (SRDLs) or similar telemetry-enabled devices using species-appropriate, minimally invasive methods (e.g., attachment to fur, feathers, or via harmless harnesses).

• Step 3: Data Transmission and Processing

  • Data Compression and Transmission: SRDLs store essential data and transmit compressed packets via satellite. The duty-cycling of transmissions is optimized to extend battery life over periods of a year or more [2].
  • Quality Control: Implement automated and manual checks to flag and remove physiologically impossible values (e.g., body temperature readings that match ambient water temperature).
  • Data Validation: Correlate a subset of animal-borne temperature and salinity profiles with concurrent measurements from Argo floats or ship-based CTDs to quantify accuracy and bias [2].

• Step 4: Data Integration

  • Format the quality-controlled data according to international standards (e.g., Climate and Forecast Metadata Conventions) and submit to specialized platforms like the Animal Borne Ocean Sensors (AniBOS) network, which feeds data into the Global Ocean Observing System [2].

Meteorological Parameter Estimation Protocol

Objective: To derive estimates of ocean surface winds, currents, and waves by analyzing the movement patterns of soaring seabirds.

• Step 1: High-Frequency Movement Data Collection

  • Deploy loggers on seabirds (e.g., albatrosses, frigatebirds) that utilize dynamic soaring and slope soaring. The loggers must include:
    • GPS: For high-resolution horizontal positioning.
    • Barometric Pressure Sensor: For accurate altitude measurement.
    • Tri-axial Accelerometer and Magnetometer: To document body posture, wingbeats, and heading [2] [21].

• Step 2: Movement Trajectory Analysis

  • Dead Reckoning: Use the bird's heading (from magnetometer data), airspeed (derived from dynamic body acceleration or pitot tubes), and change in altitude (from pressure data) to reconstruct fine-scale 3D movements between GPS fixes [21].
  • Drift Calculation: The vector difference between the bird's heading-oriented movement (its intended path) and the actual ground-tracked movement (from GPS) is calculated. This drift is attributed to the influence of wind and surface currents [2].

• Step 3: Environmental Parameter Estimation

  • Ocean Wind Estimation: Apply algorithms that model the physics of dynamic soaring. The bird's flight energy gain is balanced against the energy loss from drag, allowing for the estimation of wind speed and direction. These algorithms are integrated into Online Analytical Processing (OLAP) tools on platforms like the Biologging intelligent Platform (BiP) [2].
  • Surface Current Estimation: In calm wind conditions, the drift of swimming or floating animals (e.g., turtles, seals) can be used to estimate surface current velocity.

• Step 4: Data Validation and Sharing

  • Validate derived wind and current estimates against measurements from satellite scatterometers and oceanographic drifters.
  • Share the processed data via meteorological databases and the BiP platform, ensuring it is accessible for operational weather forecasting and climate models [2].

Biomedical Data Integration and Analysis Protocol

Objective: To integrate animal tracking and trait data for ecological forecasting and to develop models with potential translational applications in biomedicine.

• Step 1: Multi-Source Data Aggregation

  • Data Sources:
    • Biologging Repositories: Access tracking and sensor data from platforms like Movebank and BiP [2] [10].
    • Trait Databases: Aggregate species- and individual-level trait data, such as body size, limb length, metabolic rate, and lifespan, from ecological databases [22].

• Step 2: Data Curation and Standardization

  • Standardization: Harmonize data using controlled vocabularies and formats (e.g., ITIS for taxonomy, ISO for metadata) to ensure interoperability [2] [22].
  • Curation: Address data quality issues, including missing values, and ensure trait data are linked to tracking data at the individual level where possible.

• Step 3: Integrated Analysis for Hypothesis Testing

  • Research Questions: Formulate cross-disciplinary questions. For example: "How do physiological traits (e.g., metabolic rate) influence migration distance and energy expenditure?" or "How do biomechanical traits (e.g., limb length) constrain movement speed and habitat use?" [22]
  • Analytical Approach: Use multivariate statistics and machine learning to analyze the combined dataset. For instance, path analysis or structural equation modeling can test causal relationships between animal traits, movement syndromes, and environmental drivers.

• Step 4: Application in Human Health

  • Biomechanical Modeling: Insights from animal locomotion and energetics can inform the design of prosthetics and rehabilitative therapies.
  • Disease Ecology: Models of animal movement and population connectivity can predict the spread of zoonotic diseases, aiding in public health planning and intervention strategies [22].

The Data Management Pipeline: Archiving and Sharing

The value of biologging data for secondary use is fully realized only through a robust and standardized data management pipeline.

The Biologging Data Workflow

The following diagram illustrates the integrated pipeline from data collection to secondary use, highlighting the critical role of archiving and standardization.

Essential Data Types and Platforms for Secondary Use

Table 1: Key data types collected via biologging and their relevance to secondary research disciplines.

Data Type	Sensor(s)	Primary Use	Secondary Use & Discipline
Depth / Pressure	Pressure Sensor	Diving behavior	Oceanographic profile data (Oceanography)
Water Temperature	Thermistor	Thermal niche use	Sea surface & subsurface temperature maps (Oceanography)
Salinity	Conductivity Cell	Habitat preference	Ocean salinity models (Oceanography)
High-Res Location	GPS	Home range, migration	Derivation of surface currents & winds (Meteorology)
Acceleration	Accelerometer	Behavior identification, energetics	Model validation; Biomechanical studies (Biomedicine)
Body Temperature	Thermistor	Physiology, health	Fever response models (Biomedicine)

Table 2: Platforms for archiving and sharing biologging data to enable secondary use.

Platform Name	Primary Focus	Key Feature for Secondary Use	Data License
Biologging intelligent Platform (BiP)	Integrated, multi-sensor data	Online Analytical Processing (OLAP) for environmental parameter estimation [2]	CC BY 4.0
Movebank	Animal tracking	Large-scale data aggregation; Linkage to environmental layers [10]	Varies by owner
AniBOS (Animal Borne Ocean Sensors)	Oceanographic data	Standardizes & delivers biologging data into the Global Ocean Observing System [2]	Follows GOOS policy

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential tools, platforms, and reagents for biologging data research and secondary use.

Item / Solution	Type	Function in Research
Satellite Relay Data Logger (SRDL)	Hardware	Transmits compressed sensor data (depth, temperature, etc.) via satellite, enabling long-term, remote data collection [2].
Biologging intelligent Platform (BiP)	Data Platform	Stores, standardizes, and analyzes biologging data with integrated OLAP tools to estimate environmental parameters [2].
Movebank	Data Platform	A global repository for animal tracking data that facilitates data management, sharing, and analysis [10] [21].
vol2bird Algorithm	Software Algorithm	Processes weather radar data to extract biological signals, generating vertical profiles of bird density, speed, and direction [23].
FAIR Guiding Principles	Data Framework	Ensures data are Findable, Accessible, Interoperable, and Reusable, which is critical for data preservation and secondary use [10].
Integrated Bio-logging Framework (IBF)	Methodological Framework	Aids researchers in designing biologging studies by optimizing the links between biological questions, sensors, data, and analysis [21].

Biologging data are a powerful and growing resource for secondary research, turning animals into intelligent, mobile sensors of our changing planet. The full potential of this data to revolutionize oceanography, meteorology, and biomedicine will only be unlocked through a concerted, global effort to prioritize standardized archiving and open, ethical data sharing. The protocols and platforms outlined here provide a roadmap for researchers to contribute to and draw from this invaluable, expanding horizon of scientific data.

Implementing Modern Archiving Platforms and Data Standards

The expansion of bio-logging technologies has generated unprecedented volumes of data on animal movement, behavior, and physiology. Effective management and sharing of these complex datasets present significant challenges for researchers, requiring robust archival solutions that ensure data persistence, accessibility, and interoperability. This application note provides a comparative analysis of three major platforms—GBIF (Global Biodiversity Information Facility), Movebank, and the Movebank Data Repository (often referenced as BiP in bio-logging contexts)—to guide researchers in selecting appropriate archives for their specific data types and research objectives. Framed within a broader thesis on archiving approaches for bio-logging research, we detail specific workflows for mobilizing data between these platforms to maximize scientific value and compliance with evolving data policies.

GBIF serves as a global data infrastructure for biodiversity occurrence data, integrating species observations from specimens, human observations, and increasingly, machine-generated sources [24]. Its core mission centers on providing free and open access to biodiversity data to support research, policy, and conservation [25]. In contrast, Movebank is a specialized platform for managing, visualizing, and sharing animal tracking and other bio-logging sensor data, built on a flexible data model designed to accommodate diverse tracking technologies and taxa [26]. The Movebank Data Repository (MDR) is a public archive integrated with Movebank that provides long-term preservation and formal publication of curated tracking datasets with persistent identifiers [27].

Table 1: Core Characteristics of Major Bio-Logging Data Archives

Feature	GBIF	Movebank	Movebank Data Repository
Primary Scope	Global biodiversity species occurrence data [28]	Animal tracking & bio-logging sensor data [26]	Published, curated animal movement & bio-logging data [27]
Core Data Model	Darwin Core (DwC) [28]	Movebank Data Model [26]	Movebank Data Model + DwC for integration [29]
Data Publication	Dataset publication via IPT, DOI assignment [25]	User-managed studies, controlled sharing options	Formal deposition, curation, and DOI assignment [27]
Licensing	Creative Commons (CC0, BY, BY-NC)	Various, user-controlled	Creative Commons (CC0, BY, BY-NC) [27]
Key Strength	Unified search for biodiversity data, extensive use in policy & research [25]	Specialist tools for complex movement data, active research collaboration	Data preservation, formal citation, fulfilling journal/funder mandates [27]

Table 2: Quantitative Platform Metrics (Based on Available Data)

Metric	GBIF	Movebank	Movebank Data Repository
Data Volume	~3 billion occurrence records [25]	>4 billion location records, >7,500 studies [29]	Subset of Movebank studies submitted for publication
Taxonomic Coverage	All taxa (specimens, observations) [25]	>1,252 taxa [29]	Varies, primarily animal tracking data
Publisher Workflow	Integrated Publishing Toolkit (IPT) [25]	Movebank website & API	Submission via Movebank, followed by curation [27]

Interoperability and Data Integration Workflows

A critical development in bio-logging data management is the creation of workflows to mobilize data from research platforms like Movebank to global infrastructures like GBIF. This enables movement data to function as species occurrence records, significantly broadening their discoverability and utility for biodiversity modeling, conservation assessments, and policy [30]. The MOVE2GBIF project established a foundational, open-source workflow for this purpose, using an R package (movepub) to transform data formatted in the Movebank model into the Darwin Core standard required by GBIF [29] [31].

Key technical considerations for this data transformation include:

• Data reduction: High-resolution GPS data is often subsampled (e.g., to hourly positions) to align with the typical use cases and data volumes expected in biodiversity portals [30].
• Vocabulary mapping: Attributes specific to the Movebank model (e.g., animal_id, tag_id) must be mapped to appropriate Darwin Core terms (e.g., organismID, occurrenceID) [29].
• Metadata conversion: Comprehensive metadata describing the study must be converted from Movebank's system to the Ecological Metadata Language (EML) for GBIF [29].

The diagram below illustrates the primary data flow models for publishing bio-logging data to GBIF, as demonstrated by the MOVE2GBIF project and analogous efforts for camera trap data.

Data Publishing Workflow Models. This diagram outlines multiple pathways for publishing bio-logging data from primary research platforms to GBIF, including direct archival and transformed publication.

Experimental Protocols for Data Mobilization

Protocol: Publishing Movebank Animal Tracking Data to GBIF

This protocol is adapted from the established MOVE2GBIF workflow [29] [30]. It allows researchers to make their tracking data discoverable alongside other biodiversity records while maintaining a primary, rich dataset in a specialist repository.

I. Pre-publication Data Preparation on Movebank

• Data Cleaning: Ensure all data in your Movebank study is curated. This includes verifying species identifications (linked to a taxonomic authority), validating location coordinates, and defining all custom measurements in the study attributes.
• Metadata Completion: Provide comprehensive study information within Movebank, including detailed descriptions of the project objectives, animal handling and tagging protocols, and data processing steps. This information is crucial for future reuse.
• License Selection: Choose a Creative Commons license (e.g., CC0, CC BY) for your data to explicitly define the terms of reuse [27].

II. Data Transformation to Darwin Core

• Software Installation: Install the movepub R package from its GitHub repository (github.com/inbo/movepub) [29].
• Data Retrieval: Use the package to securely retrieve data from your Movebank study, authenticating with your Movebank credentials.
• Automated Transformation: Execute the package's functions to convert the Movebank data into a Darwin Core Occurrence core file. This process typically involves:
  • Mapping the animal_id to organismID.
  • Mapping the timestamp and location_lat/location_long to eventDate and decimalLatitude/decimalLongitude.
  • Subsampling high-frequency data to a standardized rate (e.g., 1 location per hour) for use as occurrence records [30].
• Metadata Generation: Use the package to generate an EML metadata file based on the information in your Movebank study.

III. Publication and Registration

• Publish Darwin Core Archive: Upload the generated Darwin Core Occurrence file and EML file to an Integrated Publishing Toolkit (IPT) instance [25].
• Register with GBIF: Use the IPT to register the dataset with GBIF, which will assign a unique DOI and begin the harvesting process.
• Archive Full Dataset: For preservation, submit the complete, high-resolution dataset from Movebank to the Movebank Data Repository or a general-purpose repository like Zenodo to receive a separate DOI [27] [30].

The Scientist's Toolkit: Essential Research Reagents and Materials

The following table details key technologies, standards, and software solutions essential for conducting bio-logging research and effectively archiving the resulting data.

Table 3: Key Research Reagents and Solutions for Bio-Logging Data Management

Item Name	Function/Application	Specific Examples / Properties
GPS Loggers	Records fine-scale location data over time.	University of Amsterdam Bird Tracking System (UvA-BITS) used in oystercatcher studies [30].
Movebank Data Model	Standardized vocabulary to describe and structure animal tracking data from diverse sources.	Defines core concepts: Animal, Tag, Deployment, Study, Event, Location [26].
Darwin Core (DwC)	Standardized vocabulary for sharing biodiversity data, essential for GBIF integration.	Includes terms: organismID, eventDate, decimalLatitude, decimalLongitude [28].
movepub R Package	Open-source software to automate transformation of Movebank data to Darwin Core.	Developed in MOVE2GBIF project; prepares data and EML metadata for GBIF publication [29].
Integrated Publishing Toolkit (IPT)	Software application for publishing and registering biodiversity datasets to GBIF.	Hosted by institutions; used to upload Darwin Core Archives and metadata for GBIF harvesting [25].
Camtrap DP	Data standard for camera trap data, enabling sharing through platforms like GBIF.	Used to transform data from platforms like Agouti into Darwin Core [25].
Creative Commons Licenses	Legal tools to grant public permission to share and use data and creative work.	CC0 (public domain), CC BY (attribution), CC BY-NC (attribution, non-commercial) [27].

GBIF, Movebank, and the Movebank Data Repository are complementary pillars in the bio-logging data ecosystem. Movebank excels as an active research platform for complex data management, while its Data Repository provides critical preservation and formal citation services. GBIF serves as a powerful engine for data discovery and reuse across the broader biodiversity community. The emerging workflows that connect these platforms, such as the MOVE2GBIF model, represent a significant advancement in open science. They empower researchers to leverage the strengths of each platform, ensuring that valuable bio-logging data contributes fully to scientific discovery, conservation policy, and the global effort to understand and protect biodiversity.

The burgeoning field of biologging, which uses animal-borne sensors to collect data on movement, behavior, physiology, and the environment, faces a critical juncture. The volume and complexity of data collected have grown exponentially, creating both unprecedented research opportunities and significant data management challenges. Without systematic approaches to data organization and sharing, much of this valuable data becomes rapidly lost to science, undermining research reproducibility, hindering collaborative synthesis, and limiting the potential for secondary applications in fields such as oceanography, meteorology, and conservation science [14]. The power of standardization lies in its ability to transform disparate, idiosyncratic datasets into interoperable, reusable resources that can fuel discovery across disciplinary boundaries.

Adopting established formats like Darwin Core and community-defined protocols addresses these challenges by providing a common framework for data description, sharing, and archiving. Darwin Core, a widely adopted standard for biodiversity data, offers a stable, straightforward, and flexible framework for compiling data from varied sources [32] [33]. Concurrently, the biologging community is developing and refining its own specialized standards to address the unique complexities of sensor data, animal metadata, and deployment information [13] [2]. This article provides detailed application notes and protocols for implementing these standards, framed within a broader thesis on effective archiving and sharing approaches for biologging data research, to empower researchers, scientists, and data managers to harness the full potential of their data.

Core Standards and Their Applicability to Biologging

Darwin Core: A Foundational Framework

Darwin Core (DwC) is a standard maintained by Biodiversity Information Standards (TDWG). It consists primarily of a glossary of terms intended to facilitate the sharing of information about biological diversity. These terms provide identifiers, labels, and definitions and are primarily based on taxa, their occurrence in nature as documented by observations, specimens, and related information [32]. Its primary strength is in enabling interoperability across the broader biodiversity informatics domain.

For biologging data, Darwin Core provides the essential framework for describing the "who" and "where" components—the taxonomic identity of the tracked animal and the spatiotemporal context of its occurrences. Key term classes include:

• Taxon: Terms for describing the species (e.g., dwc:taxonID, dwc:scientificName).
• Occurrence: Terms for documenting an event during which an organism was recorded (e.g., dwc:occurrenceID, dwc:recordedBy).
• Event: Terms for describing the context of an observation or measurement (e.g., dwc:eventID, dwc:eventDate, dwc:decimalLatitude, dwc:decimalLongitude) [32] [34].

The majority of datasets shared through global infrastructures like the Global Biodiversity Information Facility (GBIF) are published using the Darwin Core Archive (DwC-A) format, which packages data and metadata into a standardized, easily shareable folder structure [33].

Emerging Biologging-Specific Standards

While Darwin Core covers basic biodiversity concepts, the specific nature of biologging data requires more specialized descriptive protocols. A community-driven standardization framework has been proposed to advance ecological research and conservation by making bio-logging data Findable, Accessible, Interoperable, and Reusable (FAIR) [13] [14]. This framework encompasses:

• Standardized vocabularies: Controlled terms for sensor types, measured parameters, and animal life-history stages.
• Standardized transfer protocols: Methods for moving data from collection to storage platforms.
• Standardized aggregation protocols: Rules for combining datasets from different sources into unified collections [14].

Platforms like the Biologging intelligent Platform (BiP) implement this vision by storing sensor data alongside detailed, standardized metadata. BiP's metadata schema conforms to international standards like the Integrated Taxonomic Information System (ITIS), Climate and Forecast Metadata Conventions (CF), and ISO standards, ensuring broad interoperability beyond biology [2].

Table 1: Core Standards for Biologging Data Archiving

Standard Name	Primary Scope	Key Applicability to Biologging	Governing Body/Platform
Darwin Core (DwC) [32]	Biodiversity data exchange	Describing taxonomic identity (dwc:taxonID), occurrence events (dwc:occurrenceID), and basic spatiotemporal parameters (dwc:decimalLatitude, dwc:eventDate).	Biodiversity Information Standards (TDWG)
Darwin Core Archive (DwC-A) [33]	Data packaging and publication	A ready-to-publish package containing core data, extension data, and metadata (EML), ideal for sharing occurrence data derived from tracking.	GBIF / TDWG
Biologging Standardization Framework [14]	Sensor data and metadata	A comprehensive framework for standardizing sensor data formats, animal metadata, and deployment information to enable data integration.	International Bio-logging Society
BiP Metadata Schema [2]	Multi-disciplinary data interoperability	Standardizing metadata for animal traits, instrument details, and deployment context using ITIS, CF, and ISO conventions.	Biologging intelligent Platform (BiP)

Application Note: A Practical Protocol for Standardizing and Archiving Biologging Data

This protocol outlines a step-by-step process for preparing a biologging dataset for public archiving, integrating both Darwin Core and community-specific standards. The workflow ensures data is structured, documented, and packaged for maximum usability and long-term preservation.

Step-by-Step Experimental Protocol

Objective: To transform raw, researcher-collected biologging data into a standardized, archived, and FAIR-compliant dataset.

Step 1: Data Audit and Reorganization

• Action: Survey the raw data directory. Adhere to organizational best practices by eliminating redundant folders and ensuring file and folder names are clear, concise, and use only alphanumeric characters and underscores [35].
• Rationale: A logical, flat structure simplifies navigation and automated processing. Avoid deep nesting of folders and very long file names, which hinder usability [35].
• Example: Instead of Project/Data/2024/Site_A/Final/Processed/seal453_tracks_final_v2.csv, use data/raw_tracks/seal_453.csv.

Step 2: Sensor Data Standardization

• Action: Convert sensor data files (e.g., CSV, TXT) into a consistent format.
  • Standardize column headers using community vocabularies (e.g., use latitude and longitude consistently).
  • Ensure all timestamps follow the ISO 8601 standard (e.g., 2024-11-21T14:30:00Z).
  • Use SI units for all measurements.
  • Save the master file in a non-proprietary format like CSV or TXT [2].
• Rationale: Inconsistencies in column names, date formats, and units are a major barrier to data integration and reuse [2].

Step 3: Mapping to Darwin Core Terms

• Action: Create a new table, the "Occurrence Core," that maps your standardized sensor data to Darwin Core terms. This table will form the basis of your Darwin Core Archive.
• Protocol:
  • Each row should represent a unique occurrence (e.g., a location fix).
  • Extract and map the following key information:
    • occurrenceID: A unique identifier for each location fix.
    • scientificName: The full scientific name of the tracked animal.
    • eventID: An identifier linking to the tracking event.
    • eventDate: The timestamp of the location fix.
    • decimalLatitude & decimalLongitude: The geographic coordinates.
    • basisOfRecord: MachineObservation.

Table 2: Example Mapping of Sensor Data to Darwin Core

Original Sensor Data Column	Standardized Name	Darwin Core Term	Example Value
animal_id	tagLocalIdentifier	(Extension)	Seal_453
lat	decimalLatitude	dwc:decimalLatitude	-67.12345
lon	decimalLongitude	dwc:decimalLongitude	142.67890
timestamp_utc	eventDate	dwc:eventDate	2024-02-01T03:45:12Z
LCC4.2	scientificName	dwc:scientificName	Leptonychotes weddellii
-	occurrenceID	dwc:occurrenceID	https://ipt.biodiversity.org/occurrence/12345
-	basisOfRecord	dwc:basisOfRecord	MachineObservation

Step 4: Compile Extended Metadata

• Action: Create three additional tables describing the animal, the device, and the deployment, following the model used by platforms like BiP and Movebank [2] [14].
• Protocol:
  • Animal Metadata: Include individualID, taxonID, sex, lifeStage, and bodyLength.
  • Device Metadata: Include deviceID, deviceType, sensorType, and accuracy.
  • Deployment Metadata: Include deploymentID, deploymentDateTime, retrievalDateTime, attachmentMethod, and locationOfDeployment.

Step 5: Create Human-Readable Documentation

• Action: Write a comprehensive README.txt file and a data dictionary.
• Protocol:
  • The README should describe the project, the experimental design, file structure, and any quality flags or caveats.
  • The data dictionary should define every column in all tables, including units and definitions.

Step 6: Package as a Darwin Core Archive

• Action: Bundle the components into a DwC-A, which is a ZIP file containing:
  • The Occurrence Core table (from Step 3).
  • The Extension tables (from Step 4).
  • A meta.xml file describing the relationships between the core and extensions.
  • An EML (Ecological Metadata Language) file containing full project-level metadata [33].

Step 7: Deposit in a Public Repository

• Action: Choose a suitable repository and upload the DwC-A.
• Repository Selection: Suitable repositories include:
  • Movebank: A free, specialized repository for animal tracking data [14].
  • Biologging intelligent Platform (BiP): A platform designed for standardized biologging data that also offers Online Analytical Processing (OLAP) tools [2].
  • GBIF: An option if the data is primarily to be shared as species occurrence records.
  • Zenodo or Dryad: General-purpose repositories suitable for data linked to a specific publication.

Successful standardization and archiving rely on a suite of digital "reagents" and platforms. The following table details key resources that constitute the modern biologging data scientist's toolkit.

Table 3: Essential Toolkit for Biologging Data Standardization and Archiving

Tool/Resource Name	Type	Function & Application
Darwin Core Terms Guide [34]	Reference Guide	Provides the definitive list and definitions of all Darwin Core terms, essential for correct metadata mapping.
Biologging intelligent Platform (BiP) [2]	Data Platform	An integrated platform for uploading, standardizing, visualizing, and sharing biologging data with integrated metadata.
Movebank [14]	Data Repository & Management Tool	A free global repository for animal tracking data that supports a robust data model and facilitates data management pre- and post-publication.
EML (Ecological Metadata Language) [33]	Metadata Standard	An XML-based standard for describing ecological datasets in a modular fashion; the required metadata component for a Darwin Core Archive.
bandbox [35]	Software Tool	A Python package that assesses data directory organization by flagging issues like redundant directories or invalid characters in filenames before archival.
ETN R Package [13]	Software Tool	An R package designed to access and process data from the European Tracking Network, exemplifying how standardized data enables powerful analytical tools.
AniBOS (Animal Borne Ocean Sensors) [14]	Community Initiative	A global project integrating animal-borne sensor data into the Global Ocean Observing System, demonstrating cross-disciplinary data reuse.

Adopting the power of standardization is not merely a technical exercise but a fundamental shift towards a more collaborative, open, and cumulative science. By implementing formats like Darwin Core and community-defined protocols, biologging data transitions from being a static, private output of a single study to becoming a dynamic, living component of a global digital natural history archive [14]. This transformation enables researchers to address questions at previously impossible scales, from global patterns of animal movement in response to climate change to the development of more robust computational models across ecology, oceanography, and conservation.

The pathway forward requires continued community engagement. Researchers are encouraged to participate in standards bodies like the International Bio-logging Society's Data Standardisation Working Group, demand and use standardized data from platforms like Movebank and BiP, and advocate for resources and policies that support long-term data stewardship [13] [14]. Through these concerted efforts, the biologging community can ensure that the immense value locked in its data is fully realized, both for today's science and for the legacy of future generations.

Step-by-Step Guide to Preparing and Uploading Data to the Biologging intelligent Platform (BiP)

Biologging involves attaching data recorders to animals to monitor their behavior, physiology, and surrounding environment in the wild [36]. The Biologging intelligent Platform (BiP) is an integrated platform designed to address the critical social and academic mission of preserving these valuable datasets for future generations [36]. As a standardized repository, BiP facilitates collaborative research and biological conservation by enabling researchers to share, visualize, and analyze biologging data according to internationally recognized standards for sensor data and metadata storage [36]. This guide provides comprehensive protocols for preparing and uploading data to BiP, framed within the broader context of enhancing data sustainability and interoperability in biologging research.

Prerequisites and Account Setup

Required Information Before Starting

Prior to uploading data to BiP, researchers should gather the following essential information:

• Animal metadata: Species information, individual animal traits, and biological measurements
• Deployment details: Location, date, time, and methodological information
• Instrument specifications: Device type, manufacturer, and sensor configurations
• Data files: Raw or processed sensor data in original format

Registration Process

• Access the BiP Portal: Navigate to the official BiP website at https://www.bip-earth.com [36]
• Complete User Registration: Click on the registration link and provide required information to create a researcher account
• Account Verification: Verify your email address to activate the account
• Login Credentials: Use your newly created credentials to access the BiP platform

Data Preparation Protocols

Metadata Standardization

BiP requires metadata to be structured according to international standards including the Integrated Taxonomic Information System (ITIS), Climate and Forecast Metadata Conventions (CF), Attribute Conventions for Data Discovery (ACDD), and International Organization for Standardization (ISO) [36]. The platform supports three primary metadata categories:

Table 1: Required Animal Metadata

Metadata Category	Specific Elements	Format Standards
Individual Traits	Sex, body size, life history stage, breeding status	Controlled vocabularies
Taxonomic Information	Species name, taxonomic classification	Integrated Taxonomic Information System (ITIS)
Biological Measurements	Weight, length, health indicators	Numeric values with standardized units

Table 2: Required Instrument Metadata

Metadata Category	Specific Elements	Format Standards
Device Specifications	Manufacturer, model, sensor types, firmware version	Structured text fields
Calibration Data	Calibration dates, methods, reference values	ISO standard formats
Technical Parameters	Sampling rates, resolution, accuracy	Numeric values with standardized units

Table 3: Required Deployment Metadata

Metadata Category	Specific Elements	Format Standards
Deployment Context	Researcher, institution, project name	ACDD conventions
Temporal Information	Deployment date/time, retrieval date/time	ISO 8601 format
Geographical Context	Deployment location, habitat type	Decimal degrees, CF conventions

Sensor Data Formatting

BiP handles diverse sensor data types while promoting standardization for improved interoperability:

• Data Column Standardization:

  • Use consistent column names across datasets (e.g., "latitude" not "lat")
  • Apply standardized date-time formats (ISO 8601: YYYY-MM-DD HH:MM:SS)
  • Maintain consistent file structures and header lines

• Supported Sensor Parameters:

  • Positional data: Latitude, longitude
  • Environmental data: Depth, water temperature, salinity, atmospheric pressure
  • Movement data: Speed, acceleration, angular velocity
  • Physiological data: Body temperature, heart rate
  • Other parameters: Geomagnetism, light intensity

• File Format Considerations:

  • Original device files can be uploaded
  • BiP standardizes formats during processing
  • CSV or TXT formats are preferred for direct upload

Figure 1: Data preparation workflow for Biologging intelligent Platform (BiP) showing sequential steps from metadata collection to upload readiness.

Data Upload Procedure

Interactive Upload Process

• Initiate Upload Session:

  • Login to BiP and navigate to the "Upload Data" section
  • Create a new dataset record and provide basic description

• Metadata Entry:

  • Complete interactive forms for animal metadata (Table 1)
  • Input device specifications and calibration data (Table 2)
  • Provide deployment context information (Table 3)
  • Use standardized dropdown menus and controlled vocabularies where provided

• File Upload:

  • Select sensor data files from local storage
  • Upload supporting documentation if available
  • Monitor upload progress through completion indicators

• Data Standardization:

  • BiP automatically processes uploaded files to standardized formats
  • Review automated format conversions for accuracy
  • Confirm data integrity after standardization

Quality Control Checks

After upload, implement these verification protocols:

• Visualization Review:

  • Examine automatically generated route maps for positional accuracy
  • Review sensor data plots for anomalies or gaps
  • Verify temporal sequences for consistency

• Metadata Validation:

  • Confirm all required metadata fields are complete
  • Verify taxonomic information against ITIS database
  • Check unit consistency across measurements

• Data-Metadata Alignment:

  • Ensure sensor data timestamps align with deployment periods
  • Verify geographical coordinates match deployment locations
  • Confirm individual animal identifiers match metadata records

Data Sharing and Access Configuration

Access Level Settings

BiP provides flexible data sharing options to accommodate different research needs:

Table 4: Data Access Levels in BiP

Access Level	Visibility	Download Permissions	Use Cases
Open Data	Public	Free download under CC BY 4.0	Published research, collaborative projects
Private Data	Restricted	Owner permission required	Ongoing studies, sensitive locations
Embargoed Data	Metadata only	Available after embargo period	Planned publications, thesis research

License Selection

• Default License: BiP uses Creative Commons Attribution 4.0 (CC BY 4.0) for open data

• License Conditions:

  • Permits copying, redistribution, and modification
  • Requires attribution to original data creators
  • Mandates adherence to metadata credit requirements

• Access Requests:

  • Private datasets display metadata and visualizations
  • Interested users can contact data owners directly
  • Data owners manage permission requests through BiP interface

Advanced Analytical Features

Online Analytical Processing (OLAP) Tools

BiP's unique OLAP capabilities enable derived data products:

• Environmental Parameter Calculation:

  • Surface currents derived from animal movement data
  • Ocean wind estimates from bird flight patterns
  • Wave characteristic calculations from marine animal behavior

• Behavioral Parameter Estimation:

  • Dive classification and analysis for marine animals
  • Flight energy expenditure calculations for birds
  • Movement ecology metrics across taxa

• Algorithm Integration:

  • Published algorithms from previous studies are embedded in OLAP
  • Standardized processing ensures reproducible results
  • Custom parameter calculations available for advanced users

• DOI Integration:

  • Link datasets to publications using Digital Object Identifiers
  • Search for datasets using DOI of associated papers
  • Track data citations and reuse through DOI resolution

• Multi-Repository Storage:

  • BiP supports data exchange with complementary databases
  • Enhanced sustainability through distributed preservation
  • Interoperability with systems like Movebank for expanded reach

Essential Research Reagent Solutions

Table 5: Key Research Reagents and Materials for Biologging Studies

Reagent/Material	Function	Application Context
Satellite Relay Data Loggers (SRDL)	Transmit compressed data via satellite	Long-term marine mammal studies in remote regions
Animal-Borne Cameras	Visual documentation of behavior and environment	Fine-scale foraging ecology and species interactions
Acceleration Data Loggers	Monitor fine-scale movements and behaviors	Classification of foraging attempts, grooming, resting
Depth-Temperature Recorders	Profile dive behavior and thermal environment	Oceanographic data collection in ice-covered regions
Heart Rate Monitors	Estimate energy expenditure and physiological stress	Flight energy calculations in seabirds, swimming costs
Geolocation Sensors	Track position using light-based algorithms	Migration mapping for small-bodied species

The Biologging intelligent Platform represents a significant advancement in ecological data archiving by providing standardized protocols for data preservation, sharing, and reuse. By adhering to the preparation and upload procedures outlined in this guide, researchers contribute to a sustainable future for biologging data that transcends individual research projects and disciplinary boundaries. The platform's commitment to international standards, coupled with its advanced analytical capabilities and flexible sharing models, addresses the critical need for interoperable data frameworks in movement ecology and environmental monitoring. As biologging continues to expand across taxonomic groups and research questions, BiP offers a robust infrastructure for preserving these valuable datasets for future scientific discovery and conservation applications.

In the expanding field of biologging, where animal-borne electronic tags generate vast datasets on wildlife movements, behavior, and physiology, comprehensive metadata provides the essential context that transforms raw sensor readings into meaningful, reusable scientific data. The rapid growth of this discipline offers unprecedented opportunities for collaborative research and biological conservation but also presents significant challenges in data integration, sharing, and preservation [10]. Establishing robust archiving and sharing approaches is no longer a secondary concern but a foundational requirement for advancing biologging science. This application note delineates the critical metadata standards and protocols necessary for creating dynamic, interoperable archives of biologging data, ensuring their utility for future scientific discovery across multiple disciplines, from ecology to oceanography [2].

The Core Metadata Framework

A structured metadata framework is indispensable for ensuring data interoperability, facilitating discovery, and enabling accurate interpretation. The following tables summarize the essential metadata elements, categorized by domain, required for effective biologging data archiving.

Table 1: Essential Animal-Based Metadata This table details the fundamental metadata related to the study subject, which is crucial for interpreting behavioral and physiological data in a biological context.

Metadata Field	Description	Standard/Format Recommendation
Species	Scientific name of the animal.	Integrated Taxonomic Information System (ITIS) [2]
Common Name	Common name of the animal.	Automatically populated from ITIS selection [2]
Sex	Biological sex of the individual.	Controlled vocabulary (e.g., M, F, Unknown)
Life Stage	Age class or life stage.	Controlled vocabulary (e.g., Adult, Subadult, Juvenile)
Body Size	Morphometric measurements (e.g., weight, length).	Numerical value with standardized unit (e.g., kg, cm)
Breeding Status	Reproductive condition at time of deployment.	Controlled vocabulary (e.g., Breeding, Non-breeding)

Table 2: Essential Instrument-Based Metadata This table outlines the technical metadata describing the data-logging device, which is vital for understanding sensor capabilities, accuracy, and limitations.

Metadata Field	Description	Standard/Format Recommendation
Device Type	Category of the instrument (e.g., GPS logger, SRDL).	Controlled vocabulary [2]
Device ID	Unique manufacturer serial number.	Alphanumeric string
Manufacturer	Name of the device manufacturer.	Free text
Sensors	List of sensors integrated (e.g., depth, temperature, accelerometer).	Controlled vocabulary [2]
Calibration Dates	Dates when sensors were calibrated.	ISO 8601 (YYYY-MM-DD)
Sampling Rate	Frequency at which data is recorded.	Numerical value with unit (e.g., Hz)

Table 3: Essential Deployment-Based Metadata This table describes the context of how the instrument was attached to the animal, providing necessary information for analyzing data onset and assessing potential impacts on the animal.

Metadata Field	Description	Standard/Format Recommendation
Deployment ID	A unique identifier for the deployment event.	Alphanumeric string
Attachment Method	Technique used to affix the device (e.g., harness, glue).	Controlled vocabulary
Deployment DateTime	The precise date and time the device was attached.	ISO 8601 (YYYY-MM-DDThh:mm:ssZ)
Deployment Location	Geographic coordinates of the deployment site.	Decimal degrees (Lat, Lon)
Recapture DateTime	The date and time the device was recovered (if applicable).	ISO 8601 (YYYY-MM-DDThh:mm:ssZ)
Investigator	Name of the researcher who performed the deployment.	Free text

Experimental Protocols for Data Standardization and Archiving

Protocol: Data and Metadata Submission to a Standardized Platform

This protocol provides a step-by-step methodology for researchers to prepare and upload biologging data to a centralized, standards-compliant platform like the Biologging intelligent Platform (BiP) [2].

1. Pre-Deployment Registration: - Action: Prior to field deployment, register a new deployment event within the platform (e.g., BiP). - Methodology: Populate the online form with the planned deployment metadata, drawing from the fields defined in Tables 1-3. Utilize pull-down menus for controlled vocabularies to ensure consistency and minimize entry errors [2].

2. Sensor Data Collection and Export: - Action: Retrieve data from the biologging instrument after recovery or via remote transmission. - Methodology: Use the manufacturer's software to export the raw sensor data. Preserve the original data format. Common parameters include timestamp, latitude, longitude, depth, temperature, and acceleration [2].

3. Data Format Standardization: - Action: Convert the raw sensor data into a standardized format. - Methodology: Leverage the platform's (e.g., BiP) integrated tools to map raw data columns (e.g., "Lat," "Latitude") to standardized column names. Standardize date-time formats to ISO 8601 and ensure numerical values have consistent units [2].

4. Metadata Association and Validation: - Action: Link the standardized sensor data file with the complete deployment, animal, and instrument metadata. - Methodology: The platform will present the pre-registered metadata for confirmation and completion. Finalize all entries. The system should run an automated validation check to ensure all required fields are populated and conform to the agreed standards [2].

5. Licensing and Sharing Policy Selection: - Action: Define the terms of use for the dataset. - Methodology: Select a license, such as CC BY 4.0, which permits sharing and adaptation with appropriate attribution. Choose the data visibility (e.g., fully open, available on request) as required by the research funders and ethical permits [2].

Workflow Visualization: Biologging Data Archiving Pipeline

The following diagram illustrates the logical flow and decision points within the data standardization and archiving protocol.

The Scientist's Toolkit: Essential Research Reagent Solutions

Table 4: Key Platforms and Databases for Biologging Data Management This table lists essential resources for storing, visualizing, and analyzing biologging data, forming the core infrastructure for the field.

Resource Name	Type	Primary Function	Key Feature
Biologging intelligent Platform (BiP)	Integrated Data Platform	Standardized storage, visualization, and analysis of sensor data and metadata [2].	Online Analytical Processing (OLAP) tools for estimating environmental parameters [2].
Movebank	Web-Based Database	Management, sharing, and analysis of animal tracking data [2].	Largest database, containing billions of location points across numerous taxa [2].
AniBOS (Animal Borne Ocean Sensors)	Observation Network	Establishing a global ocean observation system using animal-borne sensors [2].	Focuses on gathering physical environmental data (e.g., temperature, salinity) for oceanography [2].
Seabird Tracking Database	Taxonomic-Specific Database	Hosting and managing tracking data for seabirds [10].	Enables meta-analyses of seabird behavior and distribution for conservation [2].
OTN (Ocean Tracking Network)	Acoustic Tracking Network	Global data repository for aquatic animal acoustic telemetry data [10].	Coordinates a network of acoustic receivers to track animal movements over long distances.
International Bio-Logging Society	Coordinating Body	Community-led organization promoting best practices and standardization [10].	Launches working groups (e.g., Data Standardisation WG) to develop community standards [10].

Within modern biologging research, a critical challenge lies in transforming vast archives of raw animal-borne sensor data into accessible, quantitative environmental knowledge. The Biologging intelligent Platform (BiP), a platform designed for sharing and analyzing biologging data, addresses this through its integrated Online Analytical Processing (OLAP) tools [36]. These tools are instrumental for a research thesis focused on novel archiving and sharing approaches, as they enable the secondary use of animal behavior data to estimate physical environmental parameters [36]. This application note details the protocols for using OLAP to calculate key environmental variables such as surface currents, ocean winds, and waves from biologging datasets, thereby illustrating how shared data can create value across disciplines like oceanography and meteorology [36].

Biologging data, when processed through OLAP, provides critical environmental measurements that complement traditional observation systems like Argo floats and meteorological satellites [36]. The following parameters can be derived, offering high temporal resolution and coverage in regions difficult for conventional methods to access.

Table 1: Core Environmental Parameters Derivable via OLAP Analysis

Environmental Parameter	Data Source (Animal Subjects)	Typical Sensor Data Required	Complement to Traditional Observation
Surface Currents	Seabirds, Marine Reptiles [36]	Horizontal position (GPS), timestamps [36]	Provides data in shallow waters and ice-covered regions unsuitable for Argo floats [36].
Ocean Winds & Waves	Seabirds [36]	Flight dynamics, acceleration [36]	Offers higher temporal resolution data at the ocean-atmosphere boundary [36].
Water Temperature Profiles	Seals, Sea Turtles, Sharks [36]	Depth, temperature [36]	Collects data in polar and eastern Pacific regions with high density, filling spatial gaps [36].
Salinity Profiles	Phocid Seals (e.g., Elephant Seals) [36]	Conductivity, temperature, depth (CTD) [36]	Data volume in the Antarctic is comparable to that from Argo floats [36].

Table 2: Comparison of Observation Systems

Observation System	Spatial Coverage	Temporal Resolution	Key Limitations
Meteorological Satellites	Large areas [36]	Limited frequency [36]	Cannot penetrate saltwater; only monitors surface conditions [36].
Argo Floats	Global oceans (deep waters) [36]	~10 days per profile [36]	Unsuitable for shallow waters; limited sub-surface temporal resolution [36].
Animal-Borne Sensors (via OLAP)	Polar, Temperate, Tropical regions [36]	High (e.g., continuous profiles during dives) [36]	Coverage depends on animal movement and distribution [36].

Experimental Protocols for OLAP Workflow

This section provides a detailed, step-by-step methodology for researchers to estimate environmental parameters from biologging data using the OLAP tools within the BiP platform. The workflow encompasses data preparation, upload, processing, and analysis.

Protocol 1: Data Preparation and Standardization

Objective: To format raw biologging data and its associated metadata according to international standards to ensure compatibility with the BiP platform and OLAP tools.

Materials:

• Raw sensor data files (e.g., CSV, TXT) from animal-borne devices.
• Metadata on animal subjects, instruments, and deployments.
• Computer with internet access and a modern web browser.

Procedure:

• Collate Sensor Data: Gather all raw data files from the biologging devices. Ensure files include essential parameters such as:
  • Time stamp: Use a standardized datetime format (e.g., ISO 8601: YYYY-MM-DD HH:MM:SS).
  • Location data: Latitude and longitude.
  • Environmental readings: Depth, temperature, acceleration, etc. [36].
• Compile Metadata: Prepare three categories of metadata in separate tables, adhering to international standards like the Integrated Taxonomic Information System (ITIS) and Climate and Forecast (CF) Metadata Conventions [36]:
  • Animal Metadata: Includes individual traits such as species (with taxonomy), sex, body size, and breeding status [36].
  • Instrument Metadata: Details about the deployed device, including manufacturer, model, sensor types, accuracy, and sampling frequency [36].
  • Deployment Metadata: Information on the deployment event, including attachment method, date and time of deployment, retrieval status, and location of deployment [36].
• Standardize Column Headers: Rename data columns to consistent, platform-friendly headers (e.g., use "lat" and "lon" consistently instead of mixing "Latitude" and "longitude") to prevent processing errors [36].
• Data Validation: Check for and document any gaps or anomalies in the sensor data. Verify that all metadata fields are complete.

Protocol 2: Data Upload and OLAP Processing on BiP

Objective: To upload the standardized dataset to the BiP platform and execute OLAP tools to estimate environmental parameters.

Materials:

• The standardized dataset and metadata from Protocol 1.
• User account on the BiP platform (https://www.bip-earth.com).

Procedure:

• Platform Access: Navigate to the BiP website and log in with user credentials. New users must complete a registration process [36].
• Create a New Dataset: Initiate a new dataset upload within the user dashboard.
• Upload Metadata: Use interactive forms to input the prepared animal, instrument, and deployment metadata. The platform will guide the entry of required fields [36].
• Upload Sensor Data: Attach the standardized sensor data file(s). The platform will parse the data based on the column headers and format.
• Data Standardization: Within BiP, use the built-in functions to finalize the data standardization, ensuring the entire dataset conforms to the platform's internal format for analysis [36].
• Execute OLAP Analysis:
  • Navigate to the Online Analytical Processing (OLAP) tools section.
  • Select the target dataset and choose the specific environmental algorithm to run (e.g., "Surface Currents from Bird Movement" or "Water Temperature Profile Extraction") [36].
  • Initiate the processing job. The system will integrate published algorithms to calculate the selected environmental parameters from the animal movement and sensor data [36].
• Output Retrieval: Once processing is complete, download the results. The output will typically be a new data file containing the estimated environmental variables (e.g., current velocity, wind speed, wave height) alongside the original sensor data.

Workflow Visualization

The following diagram illustrates the complete experimental protocol from data collection to analysis, providing a logical overview of the process.

The Scientist's Toolkit: Essential Research Reagents & Materials

Successful execution of these protocols requires a suite of specialized materials and digital resources. The following table details the key "research reagent solutions" essential for biologging research and environmental data analysis.

Table 3: Essential Research Reagents and Materials for Biologging Analysis

Item Name	Function/Application	Specifications & Examples
Animal-Borne Data Loggers	Primary data collection units attached to animals to record movement, behavior, and environmental data.	Includes Satellite Relay Data Loggers (SRDLs) for marine mammals [36], and devices for birds, reptiles, and fish. Measures depth, temperature, acceleration, etc.
Biologging Intelligent Platform (BiP)	The core platform for data archiving, standardization, and analysis. Hosts the OLAP tools for environmental parameter estimation.	Web-based platform (https://www.bip-earth.com). Supports FAIR data principles, CC BY 4.0 licensing for open data, and stores sensor data with standardized metadata [36].
OLAP (Online Analytical Processing) Tools	Integrated algorithms within BiP that calculate environmental parameters from animal movement and sensor data.	Algorithms published in peer-reviewed studies are integrated to estimate surface currents, ocean winds, and waves [36].
Metadata Standards	Defined vocabularies and formats to ensure data interoperability and reuse across disciplines.	Conforms to international standards: Integrated Taxonomic Information System (ITIS), Climate and Forecast (CF), and ISO conventions [36].
Data Visualization & Statistical Tools	Software for exploring, analyzing, and communicating biological and environmental data findings.	R (with ggplot2 package) and Python (with Matplotlib, Seaborn) for creating publication-quality plots [37] [38]. Tools like GraphPad Prism for biostatistics [38].

Overcoming Common Pitfalls in Data Management and Machine Learning

Addressing Global Biases and Gaps in Biologging Data Collection

Biologging, the practice of attaching data recorders to animals to monitor their behavior, physiology, and surrounding environment, has transformed ecological research and conservation [2]. This methodology provides unprecedented, real-time insights into animal lives in wild settings, offering data that is critical for understanding ecological dynamics and informing conservation strategies [39]. However, the deployment of these technologies has not been uniform across global ecosystems. Recent research led by the University of California, Berkeley, reveals substantial global biases and gaps in biologging data collection, with the majority of data originating from remote or suburban regions in Europe and the United States [39]. This disparity leaves critical knowledge gaps, particularly for highly urbanized areas and regions in the Global South that are experiencing rapid environmental change. These biases hinder the development of effective, global biodiversity conservation strategies. This document outlines standardized protocols and archival approaches to mitigate these disparities and promote a more equitable, comprehensive global biologging data infrastructure.

Quantifying Global Biases in Biologging Data

The following table summarizes the primary quantitative findings on global biologging data collection biases, as identified in a 2025 analysis of existing tracking data [39].

Table 1: Documented Biases in Global Biologging Data Collection

Bias Dimension	Overrepresented Regions/Contexts	Underrepresented Regions/Contexts
Geographical	Remote and suburban regions in Europe and the United States [39]	Highly urbanized areas globally; vulnerable regions across the Global South [39]
Environmental Context	Pristine or protected national park areas [39]	Human-dominated landscapes, areas undergoing rapid environmental change [39]
Oceanographic Data	Antarctic, Arctic, and the eastern Pacific Ocean (data primarily from pinnipeds) [2]	Temperate and tropical regions (despite use of turtles, sharks, and fish) [2]

These biases mean that our understanding of animal behavior and ecology is spatially and contextually limited. As noted by researchers, scientists often prioritize tracking animals in remote national parks, leaving a critical gap in knowledge about how animals live and cope in cities and other human-modified landscapes [39]. This information is vital for designing effective conservation strategies for a planet increasingly shaped by human activity.

Standardized Protocol for Equitable Biologging Data Collection

To address the documented gaps and biases, a standardized, proactive methodology for biologging studies is essential. The protocol below provides a framework for planning and executing a biologging deployment with an emphasis on equitable data collection and sharing.

Diagram 1: Standardized workflow for equitable biologging studies.

Protocol Steps in Detail

• Study Design & Ethical Review: Formulate a clear biological question. Crucially, apply the 5R principle (Replace, Reduce, Refine, Responsibility, and Reuse) to enhance animal welfare and data quality from the outset [40]. Submit the study plan for approval by an institutional animal ethics committee.

• Technology Selection & Calibration: Choose biologging devices (e.g., GPS loggers, satellite relay data loggers - SRDLs, accelerometers, proximity loggers like the Encounternet system) based on the scientific question, target species, and habitat [2] [41]. Calibrate sensors, such as those measuring Received Signal Strength Indication (RSSI) in proximity loggers, to ensure accurate distance estimation during encounters [41].

• Field Deployment: Deploy tags on animals using species-appropriate attachment methods (e.g., weak-link harnesses for birds) [41]. Actively prioritize deployments in underrepresented regions, including the Global South and highly urbanized environments, to directly counter existing geographical biases [39].

• Data Acquisition & Processing: Collect raw data from the devices or via basestations. For proximity loggers, this results in "encounter logs" containing IDs, timestamps, and RSSI values [41]. Process the data to filter records by signal strength, amalgamate temporally clustered logs, and examine tag reciprocity to ensure data quality.

• Data Standardization & Archiving: Standardize the processed sensor data and associated metadata according to international conventions (e.g., Integrated Taxonomic Information System (ITIS), Climate and Forecast (CF) Metadata Conventions) [2] [13]. This step is critical for making data interoperable and reusable across different disciplines.

• Data Analysis & Sharing: Conduct biological and ecological analyses. To maximize impact and collaboration, share the standardized dataset via public platforms like the Biologging intelligent Platform (BiP) or Movebank, ideally under a permissive license such as CC BY 4.0 [2].

The Scientist's Toolkit: Essential Materials and Reagents

Table 2: Key Research Reagent Solutions for Biologging Studies

Item Name	Function / Application	Key Features / Standards
Satellite Relay Data Logger (SRDL)	Transmits compressed data (e.g., dive profiles, temperature) via satellite; used for long-term tracking of marine mammals in remote areas [2].	Enables oceanographic data collection in ice-covered regions inaccessible to ships and Argo floats [2].
Encounternet Proximity Logger	A digital proximity-logging system for direct mapping of animal social encounters by recording reciprocated contacts between tagged individuals [41].	Logs raw signal-strength (RSSI) data for distance estimation; supports tag-to-tag communication over distances >10m [41].
Biologging intelligent Platform (BiP)	An integrated online platform for sharing, visualizing, and analyzing standardized biologging data [2].	Adheres to international metadata standards; includes Online Analytical Processing (OLAP) tools to calculate environmental parameters [2].
Movebank Database	A web-based platform for managing, sharing, and analyzing animal tracking and other biologging data [13].	One of the largest biologging databases; supports collaboration and data reuse across the research community [2].
Animal Borne Ocean Sensors (AniBOS)	A global observation network that uses animal-borne sensors to gather physical oceanographic data [2].	Complements traditional ocean observation systems like Argo floats and satellites, especially in shallow waters [2].

Data Archiving and Sharing Workflow

A robust and standardized archival process is the cornerstone of overcoming data gaps and fostering collaborative, interdisciplinary science. The workflow for this process is outlined below.

Diagram 2: Data archiving and sharing workflow.

• Data Standardization: The raw sensor data must be converted into standardized formats. This involves using consistent column names, date-time formats (e.g., ISO8601), and file structures to overcome inconsistencies that vary by sensor type, manufacturer, or software version [2].
• Metadata Annotation: Comprehensive metadata is what makes sensor data meaningful and reusable. This includes detailed information about the individual animal (e.g., sex, body size), the attached instrument, and the deployment circumstances (who, when, where, how) [2]. Platforms like BiP use pull-down menus aligned with international standards to minimize errors and inconsistencies during this annotation process [2].
• Platform Upload & Access Setting: The standardized data and metadata are uploaded to a public repository. Data owners can then set the access level, choosing to make data fully open under licenses like CC BY 4.0, which permits copying and modification, or keeping it private, accessible only upon request [2]. This flexibility helps balance open science with data ownership concerns.

Addressing the profound global biases in biologging data is not merely a technical challenge but an ethical and strategic imperative for conservation. By adopting the standardized protocols, equitable deployment strategies, and robust data archiving frameworks outlined in these application notes, the research community can transform biologging from a patchwork of disparate studies into a truly global, collaborative, and actionable system for understanding and protecting biodiversity.

The exponential growth of bio-logging research—the use of animal-borne electronic tags—has generated unprecedented volumes of data on animal movement, behavior, physiology, and environments [10]. These datasets constitute invaluable dynamic archives of animal life on Earth, with tremendous potential for addressing biodiversity threats and expanding digital natural history collections [10] [42]. However, this opportunity is tempered by a significant challenge: extreme heterogeneity in data formats, column headers, and structural schemas across independent research initiatives. This heterogeneity creates substantial bottlenecks in data integration, analysis, and reuse, hindering the transformative potential of bio-logging science. Data harmonization—the practice of reconciling disparate data types, levels, and sources into compatible and comparable formats—emerges as a critical methodology for unlocking the full value of these ecological archives [43]. This article establishes comprehensive protocols for standardizing diverse data formats and column headers, framed within the imperative to create accessible, preserved bio-logging data collections for the global research community.

Understanding the Dimensions of Data Heterogeneity

Data harmonization resolves heterogeneity across three primary dimensions, each presenting distinct challenges for bio-logging data integration.

Syntax Heterogeneity (Data Format)

Syntax refers to the technical format of data files (e.g., .csv, JSON, HDF5, SQL databases) [43]. Bio-logging data arrives in myriad formats from different sensor systems and proprietary software, requiring initial conversion to workable formats before harmonization can proceed.

Structural Heterogeneity (Conceptual Schema)

Structure concerns how variables relate within datasets, ranging from highly organized tables to unstructured data streams [43]. Bio-logging data manifests structural variance in:

• Temporal representation: Event-based formats (single row per animal movement episode) versus panel data (multiple rows per animal across time points) [43].
• Header organization: Multi-index column headers from aggregation operations versus flattened structures [44].
• Data granularity: Varying temporal or spatial resolution across tracking datasets.

Semantic Heterogeneity (Intended Meaning)

Semantics involves the intended meaning of terms and variables [43]. This presents particularly subtle challenges in bio-logging, where identical terminology may measure different concepts (e.g., "migration duration" defined differently across studies) or different terms may describe identical concepts (e.g., "location quality" versus "positional precision").

Table 1: Dimensions of Data Heterogeneity in Bio-Logging Research

Dimension	Definition	Bio-Logging Examples	Impact on Analysis
Syntax	Technical file format and encoding	CSV, JSON, HDF5, proprietary binary formats	Prevents immediate data loading and combination
Structure	Conceptual schema and variable relationships	Event data vs. panel data formats; multi-index headers	Requires structural transformation before analysis
Semantics	Intended meaning of terms and variables	Differing operational definitions of "foraging behavior"	Leads to erroneous comparisons and conclusions

Experimental Protocols for Data Harmonization

Protocol 1: Standardizing Column Headers and Structure

The following protocol provides a systematic methodology for transforming heterogeneous column headers into a standardized schema suitable for bio-logging data integration.

Research Reagent Solutions

Table 2: Essential Computational Tools for Header Standardization

Tool/Resource	Function	Application Context
Pandas Library	Python data manipulation toolkit	Primary engine for data transformation and header operations
String Methods	(.str.upper(), .str.replace())	Case normalization and character replacement in headers
Regular Expressions	Pattern matching for complex string operations	Identifying and transforming patterned header elements
Custom Mapping Functions	(lambda functions)	Applying complex transformation logic to header sets
Semantic Type Libraries	Pre-defined value dictionaries (e.g., species taxonomies)	Standardizing content within columns after header stabilization

Step-by-Step Workflow

• Convert Headers to String Type

  • Rationale: Ensures consistent string operations regardless of original header format (e.g., numeric indices) [44].
  • Implementation:

• Case Normalization

  • Rationale: Resolves case sensitivity issues in column referencing [44].
  • Implementation (multiple approaches):

• Remove Extraneous Whitespace

  • Rationale: Eliminates hidden formatting differences and reference errors [44].
  • Implementation:

• Character Standardization

  • Rationale: Replaces problematic characters (spaces, hyphens) with consistent separators for programmatic access [44].
  • Implementation:

• Structural Flattening

  • Rationale: Converts multi-index headers from aggregated data into single-level headers [44].
  • Implementation:

• Semantic Mapping

  • Rationale: Aligns heterogeneous terminology to common ontology using a controlled vocabulary [10] [45].
  • Implementation:

The following workflow diagram illustrates the complete header standardization process:

Protocol 2: Value Standardization and Semantic Harmonization

Beyond header formatting, standardizing values within columns is essential for meaningful data integration in bio-logging collections.

Fuzzy Matching for Value Standardization

• Algorithm Selection

  • Technique: Apply the Levenshtein distance algorithm to find closest valid values for invalid entries [46].
  • Process: The algorithm computes character differences between strings, identifying nearest matches through tokenized comparison.

• Threshold Configuration

  • Implementation: Set matching thresholds to balance precision and recall [46]:
    • High (90%+): For critical classifications (e.g., species identification)
    • Default (80%+): For standard descriptive fields (e.g., sensor types)
    • None: Automatic replacement with closest match for non-critical metadata

• Dictionary-Based Validation

  • Implementation: Validate against controlled vocabularies and semantic types [46]:

Semantic Harmonization for Conceptual Alignment

The ECHO-wide Cohort approach demonstrates successful large-scale semantic harmonization through Common Data Models (CDMs) and harmonization protocols [47]. This involves:

• Construct Identification: Define core scientific constructs (e.g., "foraging behavior," "migration event") [47].
• Measure Mapping: Document how different measures across datasets operationalize these constructs [45].
• Transformation Rules: Develop algorithms to map study-specific data to standardized variable formats [45] [47].

Table 3: Semantic Harmonization Framework for Bio-Logging Data

Harmonization Phase	Process	Output
Concept Definition	Define core constructs with domain experts	Standardized ontology for bio-logging phenomena
Measure Inventory	Catalog existing measures for each construct across datasets	Crosswalk between measure-specific and common variables
Transformation Specification	Develop algorithms to map specific measures to common format	Processing scripts and validation checks
Validation	Assess harmonized data for conceptual equivalence	Quality metrics and harmonization assessment report

Implementation Framework for Bio-Logging Data Collections

Organizational Infrastructure

Successful large-scale harmonization requires coordinated community effort and infrastructure [10]:

• Common Data Models (CDMs): Standardized schemas defining essential data elements, measures, and formats for bio-logging data [47].
• Coordinating Body: Community-led organizations (e.g., International Bio-Logging Society) to govern standards and drive adoption [10].
• Centralized Platforms: Data repositories supporting preservation and access (e.g., Movebank, Seabird Tracking Database) [10].

Workflow Integration

The complete data harmonization pipeline integrates both technical and conceptual processes:

Quantitative Assessment of Harmonization Outcomes

Table 4: Metrics for Evaluating Harmonization Success in Bio-Logging Data

Evaluation Dimension	Pre-Harmonization State	Post-Harmonization Target	Measurement Approach
Syntax Compatibility	Multiple proprietary formats	≤ 3 standardized, open formats	Percentage of data in standard formats
Header Consistency	Heterogeneous case, spacing, separators	100% consistent formatting	Automated header validation checks
Semantic Interoperability	Variable operational definitions	Common ontology with mapping	Cross-dataset comparability index
Processing Efficiency	Manual, one-off transformations	Automated, reproducible workflows	Reduction in data preparation time
Repository Compliance	Dataset-specific schemas	Common Data Model adherence	Validation against CDM specification

Harmonizing diverse data formats and column headers transcends technical exercise to become foundational for realizing the potential of bio-logging data as dynamic archives of animal life [10]. The protocols outlined here provide a rigorous methodology for standardizing syntax, structure, and semantics—enabling integration across disparate datasets. When embedded within broader community initiatives for data standardization, preservation, and access [10] [42], these approaches support the transformation of fragmented individual datasets into unified resources. Through committed adoption of harmonization practices, the bio-logging research community can ensure these vital digital archives continue to illuminate wildlife biology and inform conservation strategies for years to come.

Overfitting represents a fundamental challenge in developing robust behavioral classification models, particularly within the domain of biologging data research. It occurs when a machine learning model fits its training data too closely, learning both the underlying patterns and the irrelevant noise or random fluctuations [48] [49]. In the context of behavioral classification, an overfitted model essentially "memorizes" the specific examples in its training dataset rather than learning the generalizable features that distinguish behavioral states [49] [50]. This results in a model that performs exceptionally well on its training data but fails to generalize to new, unseen data—a critical flaw for models intended for real-world scientific application.

The epidemic nature of overfitting in behavioral classification is particularly concerning. A systematic review of animal accelerometer-based behavior classification literature revealed that 79% of examined studies (94 papers) did not adequately validate their models to robustly identify potential overfitting [51]. This prevalence underscores the need for heightened awareness and improved methodological rigor throughout the research community. When overfitted models are deployed or shared without detection, they produce misleading results that can misdirect scientific conclusions and waste valuable research resources.

The consequences of overfitting are especially profound within biologging research, where models are increasingly used to draw ecological inferences and inform conservation strategies. An overfitted behavioral classifier may appear highly accurate during development but will perform poorly when applied to data from new individuals, different populations, or varying environmental conditions [51] [52]. This limitation fundamentally undermines the scientific value of shared biologging datasets and hampers reproducibility across studies.

Detecting Overfitting: Key Indicators and Diagnostic Protocols

Performance Discrepancy Analysis

The most straightforward method for detecting overfitting involves comparing model performance between training and validation datasets:

Table 1: Performance Indicators of Model Fit Status

Model Status	Training Performance	Validation Performance	Performance Gap
Overfitting	High accuracy (e.g., >95%) [53]	Significantly lower accuracy (e.g., <70%) [53]	Large gap (>20 percentage points)
Underfitting	Low accuracy [49]	Similarly low accuracy [49]	Minimal gap
Well-fit	High accuracy	Similarly high accuracy	Small gap (<10 percentage points)

To implement this diagnostic approach:

• Data Partitioning: Reserve a portion of your labeled dataset (typically 20-30%) as a validation set before any model training begins [48] [54]. This data must not be used in training or feature selection.
• Performance Monitoring: Track appropriate evaluation metrics (e.g., accuracy, F1-score, AUC-ROC) on both training and validation sets throughout the training process [50] [55].
• Gap Analysis: Calculate the performance difference between training and validation sets. A significant and growing gap indicates overfitting [53].

Cross-Validation Protocols

K-fold cross-validation provides a more robust approach for detecting overfitting than a single train-validation split [48] [55]:

Table 2: K-fold Cross-Validation Implementation

Step	Procedure	Purpose
1	Randomly shuffle dataset and split into k equally sized folds (typically k=5 or k=10)	Ensure random representation across folds
2	Iteratively use k-1 folds for training and the remaining fold for validation	Maximize data usage for both training and validation
3	Repeat process until each fold has served as validation once	Eliminate bias from single data split
4	Calculate mean performance across all folds	Obtain stable estimate of generalization performance
5	Compare training and validation performance for each fold	Identify consistency of performance gaps

The following workflow represents the complete model development and validation process for behavioral classification:

For behavioral classification tasks with temporal dependencies (e.g., accelerometry time series), special consideration is needed during data splitting. Standard random splitting may create data leakage through temporal autocorrelation [51]. Instead, implement:

• Grouped splits by individual animal to ensure the same individual doesn't appear in both training and validation sets
• Temporal splits where training and validation data come from different time periods
• Stratified splits to maintain similar distribution of behavioral classes across splits

Experimental Protocols for Overfitting Detection

Learning Curve Analysis Protocol

Learning curves provide powerful visual diagnostic tools for detecting overfitting:

Materials Required:

• Labeled behavioral dataset with sufficient samples for incremental training
• Computational environment for model training and evaluation
• Plotting library (e.g., matplotlib, ggplot) for visualization

Procedure:

• Start with a small subset (e.g., 10%) of your training data
• Train the model on this subset and evaluate performance on both the training subset and a validation set
• Gradually increase the training data size in steps (e.g., 20%, 30%, ..., 100%)
• At each step, record training and validation performance
• Plot learning curves with dataset size on the x-axis and performance metric on the y-axis

Interpretation:

• Well-fit model: Both curves converge to similar performance values
• Overfitting: Significant gap between curves, with training performance remaining notably higher
• Underfitting: Both curves converge at low performance values

Model Complexity Analysis Protocol

This protocol evaluates how model complexity affects generalization:

Materials Required:

• Framework for controlling model complexity (e.g., tree depth, network layers, regularization parameters)
• Validation dataset as described in Section 2.1

Procedure:

• Select a complexity parameter to vary (e.g., decision tree depth, number of neural network layers)
• Train multiple models with increasing complexity levels
• Evaluate each model on training and validation datasets
• Plot complexity versus performance for both datasets

Interpretation:

• Initially, both training and validation performance should improve with increasing complexity
• At the overfitting point, validation performance plateaus or deteriorates while training performance continues to improve
• The optimal complexity lies just before this inflection point

Prevention Strategies: The Researcher's Toolkit

Technical Approaches to Mitigate Overfitting

Table 3: Overfitting Prevention Techniques and Applications

Technique	Mechanism	Implementation Examples
Regularization	Adds penalty for complexity to loss function [56]	L1 (Lasso), L2 (Ridge), ElasticNet [56] [53]
Early Stopping	Halts training when validation performance stops improving [48] [50]	Monitor validation loss; stop when no improvement for N epochs
Data Augmentation	Artificially increases dataset size and diversity [48] [54]	For accelerometry: add noise, time-warping, rotation [55]
Ensemble Methods	Combines multiple models to reduce variance [48]	Random Forests, Gradient Boosting Machines (XGBoost) [55]
Dimensionality Reduction	Reduces feature space to most informative variables [56]	PCA, feature selection based on importance scores
Dropout	Randomly disables neurons during training (for neural networks) [50] [55]	Typically disable 20-50% of neurons per layer

Data-Centric Strategies

The quality and quantity of training data fundamentally influence overfitting risk:

Increasing Dataset Size and Diversity:

• Collect data from multiple individuals across different conditions [54] [53]
• Ensure balanced representation of all behavioral classes
• For biologging studies, include individuals from different populations, seasons, and environmental contexts

Data Augmentation for Behavioral Data:

• For accelerometry data: apply random rotations, additive noise, time warping, or scale adjustments [55]
• Ensure augmented data maintain behavioral semantics (e.g., don't augment resting behavior to appear as movement)
• Validate that augmentation improves generalization to truly unseen data, not just validation performance

The following diagram illustrates the relationship between model complexity, dataset size, and overfitting risk:

Implications for Biologging Data Archiving and Sharing

The overfitting epidemic directly impacts practices for archiving and sharing biologging data. When models overfit to specific datasets, they fail to generalize across studies, diminishing the value of shared data resources.

Archiving Best Practices to Support Robust Modeling

Comprehensive Metadata Documentation:

• Document all preprocessing steps, sensor specifications, and labeling protocols
• Record demographic information about subjects (species, sex, age, physiological state)
• Include environmental context (habitat type, season, weather conditions)
• Detail annotation procedures, including annotator expertise and validation methods

Stratified Data Archives:

• Structure datasets to facilitate proper train-validation splits by individual, location, and time
• Flag data characteristics that might create spurious correlations (e.g., all resting behavior at night)
• Provide multiple sampling scenarios to support cross-population validation

Model Sharing Protocols

When sharing pre-trained behavioral classifiers:

Required Documentation:

• Complete performance metrics on both training and validation datasets
• Details of the validation protocol, including splitting methodology
• Description of data characteristics the model was trained on
• Clear statements about limitations and appropriate use cases

Standardized Reporting: Adopt reporting checklists that explicitly address overfitting risks:

• Data splitting methodology
• Cross-validation procedures
• Hyperparameter tuning protocols
• Final evaluation on completely held-out test data

Essential Research Reagents and Computational Tools

Table 4: Key Research Reagents and Solutions for Behavioral Classification

Resource Category	Specific Tools/Approaches	Function in Overfitting Prevention
Regularization Implementations	L1/L2 in scikit-learn, Dropout in TensorFlow/PyTorch [50] [55]	Explicitly penalize model complexity to improve generalization
Cross-Validation Frameworks	Scikit-learn KFold, StratifiedKFold, GroupKFold [51]	Robust performance estimation and hyperparameter tuning
Automated ML Platforms	Amazon SageMaker, Azure Automated ML [54] [53]	Built-in overfitting detection and regularization
Model Interpretation Tools	SHAP, LIME [55]	Identify feature reliance patterns suggestive of overfitting
Data Augmentation Libraries	Albumentations, SciPy signal processing	Increase effective dataset size and diversity
Ensemble Methods	XGBoost, Random Forests, Stacking ensembles [55]	Reduce variance through model averaging

The overfitting epidemic in behavioral classification represents a critical challenge that demands systematic attention throughout the research pipeline. From experimental design through data sharing, each stage offers opportunities to detect and prevent overfitting. The protocols and frameworks presented here provide concrete strategies for developing more robust, generalizable behavioral classifiers.

For the biologging research community, addressing overfitting is particularly essential for realizing the full potential of data archiving and sharing initiatives. Only when models can generalize across datasets can we build a cumulative science of animal behavior. By adopting these practices, researchers can contribute to more reproducible, reliable behavioral classification that advances our understanding of animal movement and ecology.

Within biologging research, which involves collecting detailed data from animal-borne sensors on movement, physiology, and environmental parameters, the challenge of ensuring data longevity is paramount [2]. The massive volumes of complex sensor data generated are not only critical for behavioral ecology but are also increasingly valuable for secondary applications in oceanography and meteorology [2] [3]. Effective archiving and sharing of these datasets are essential. This document outlines application notes and protocols for migrating and refreshing outdated storage media, providing a structured approach to safeguard biologging data against technological obsolescence and physical degradation.

Quantitative Comparison of Archival Storage Media

A strategic approach to data longevity begins with selecting appropriate storage media. The table below compares key characteristics of traditional and emerging options relevant to biologging data archiving.

Table 1: Comparative Analysis of Long-Term Archival Storage Media

Storage Medium	Estimated Lifespan	Capacity/Density	Key Advantages	Key Challenges / Refresh Triggers
Magnetic Tape (LTO)	15 - 30 years [57]	High (Terabytes per cartridge)	Proven, cost-effective for large volumes [57].	Technology obsolescence (drive compatibility); requires controlled environment.
Hard Disk Drives (HDD)	3 - 5 years (active use)	High	Fast read/write speeds; suitable for active archives.	Mechanical failure; high power consumption for always-on systems.
Optical Discs (Archival Grade)	50 - 100+ years (theoretical)	Low to Moderate	Passive, durable media; immune to electromagnetic effects.	Susceptible to physical scratches, UV light; low capacity.
Synthetic DNA Data Storage	Thousands of years [58] [59]	Extremely High (theoretical)	Unparalleled density and longevity; ultrastable passive medium [58] [59].	Very high write cost and latency; early-stage technology; specialized retrieval [58].

Experimental Protocols for Data Preservation

Protocol: Integrity Check and Migration for Traditional Digital Media

This protocol details the process for verifying data on existing storage and migrating it to new media, a cornerstone of a robust data refresh cycle.

3.1.1. Materials and Reagents

• Source Media: The legacy storage media (e.g., HDD, tape cartridge).
• Target Media: New storage media with sufficient capacity.
• Checksum Verification Tool: Software (e.g., md5deep, sha256sum).
• Data Transfer Station: A secure computer with interfaces for both source and target media.
• Data Management Plan (DMP): A formal document outlining data types, formats, and governance [57].

3.1.2. Procedure

• Pre-Migration Integrity Check:
  • Generate checksum hashes (e.g., SHA-256) for all files on the source media using the verification tool.
  • Record these hashes in a separate, secure log file that will not be migrated during the process.
• Data Transfer:
  • Copy all data and associated metadata from the source media to the target media.
  • Ensure the file structure and permissions are preserved.
• Post-Migration Verification:
  • Generate new checksum hashes for all files on the target media.
  • Systematically compare the new hashes against the log of original hashes. Any mismatch indicates a corruption during transfer, and the affected files must be re-transferred.
• Validation and Decommissioning:
  • Perform spot checks by opening a random sample of files from the target media to confirm usability.
  • Update the Data Management Plan (DMP) and any catalogs with the new media location and migration date.
  • Only after successful verification should the source media be securely erased or decommissioned.

Protocol: End-to-End Workflow for DNA-Based Data Archiving

DNA storage is an emerging platform for "cold" archives, offering exceptional longevity. This protocol describes the workflow from digital file to synthetic DNA and back [58] [59].

3.2.1. Research Reagent Solutions

Table 2: Essential Reagents for DNA Data Storage Protocols

Item	Function / Description
Oligonucleotide Pool	Short, synthetic DNA strands encoding the digital data, typically purchased from commercial synthesis providers.
DNA Stabilization Matrix (e.g., Silica)	Protects DNA molecules from hydrolysis and other environmental damage, enabling room-temperature storage for centuries [58].
Polymerase Chain Reaction (PCR) Reagents	Enzymes and nucleotides for targeted amplification of specific DNA sequences, enabling large-scale random access and retrieval [58].
Next-Generation Sequencing (NGS) Kit	Reagents for reading the nucleotide sequence of the retrieved DNA strands to reconstruct the original digital data.

3.2.2. Procedure

Part A: Encoding and Writing (Digital-to-Biological)

• Encoding and Error Correction: Convert the digital file (e.g., a biologging dataset) from binary code (0s and 1s) into a nucleotide sequence (A, C, G, T) using a robust encoding algorithm (e.g., DNA Fountain [58]). Integrate error-correcting codes to correct for synthesis and sequencing errors [58] [59].
• Oligo Design and Synthesis: Fragment the encoded sequence into short, overlapping oligonucleotides. Include addressing information in each oligo to facilitate random access and reassembly. Send the designed sequences to a commercial vendor for chemical synthesis.
• Storage Preparation: Upon receipt, the synthesized DNA pool is encapsulated in a protective matrix like silica beads to ensure chemical stability for long-term archival [58].

Part B: Retrieval and Decoding (Biological-to-Digital)

• Random Access and Amplification: To retrieve a specific file, use Polymerase Chain Reaction (PCR) with primers matching the file's unique address to selectively amplify only the relevant DNA strands from the entire pool [58].
• Sequencing: Use a Next-Generation Sequencing (NGS) platform, such as a nanopore sequencer, to read the nucleotide sequences of the amplified DNA strands [58].
• Decoding and Reconstruction: Translate the sequenced nucleotides back into binary data using the same encoding scheme. Apply error-correcting codes to fix any errors, finally reconstructing the original digital file.

Diagram 1: DNA data storage and retrieval workflow.

The Scientist's Toolkit: Biologging Data Archiving Essentials

Successful long-term preservation of biologging data relies on a combination of technical infrastructure, standardized practices, and strategic planning.

Table 3: Essential Tools and Practices for Biologging Data Archiving

Tool / Practice	Function in Ensuring Data Longevity
Data Management Plan (DMP)	A formal document outlining the lifecycle management of data, including storage formats, refresh schedules, metadata standards, and sharing policies [57].
Common Data Elements (CDEs) & Ontologies	Standardized terms and definitions (e.g., from ITIS, CF/ACDD conventions) that ensure data consistency and interpretability over time and across research groups [2] [57].
Standardized Metadata	Detailed information about animal traits, instrument details, and deployment context, stored in internationally recognized formats to facilitate future understanding and reuse [2].
Trusted Data Repositories	Platforms like Movebank or the Biologging intelligent Platform (BiP) that provide structured environments for storing, sharing, and preserving biologging data with standardized formats [2].
Checksum Verification Tools	Software utilities that generate and verify digital fingerprints of files, critical for ensuring data integrity during migration and refresh cycles.

Strategic Roadmap and Implementation

Integrating these protocols requires a forward-looking strategy. For traditional media, establish a regular migration schedule based on media lifespan (e.g., every 3-5 years for HDDs) [57]. For emerging technologies, DNA storage is projected to evolve from prototypes to a practical supplementary archive for "cold" biologging data, such as raw sequencing archives and de-identified records, between 2025 and 2030 [58]. A hybrid approach is recommended: use disk or tape for active projects and frequent access, while planning for DNA or other advanced media for final, irreplaceable dataset archiving.

Diagram 2: A tiered data archiving strategy.

The expansion of biologging research generates vast datasets detailing animal movement, behavior, and physiology. Effectively archiving and sharing this data is crucial for ecological discovery and conservation, yet it presents a fundamental challenge: balancing the imperative for open data sharing with the ethical and practical need to protect sensitive information. This Application Note provides structured protocols for implementing managed access frameworks that support collaborative biologging science while safeguarding data integrity, individual animal welfare, and stakeholder interests. Adopting these approaches ensures data flow from acquisition to repository is efficient, standardized, and secure, enabling data to function as a living archive of animal life on Earth [13].

Data Classification and Sensitivity Tiers

Establishing a clear data classification system is the foundational step in managing access. This involves categorizing datasets based on their sensitivity and potential risks, which in turn dictates the appropriate level of access control.

Table 1: Data Sensitivity Tiers and Corresponding Access Protocols

Sensitivity Tier	Data Description	Primary Risks	Recommended Access Protocol
Public	Processed, non-sensitive movement paths; summary metrics; species occurrence data.	Misinterpretation; lack of attribution.	Open Access (e.g., CC BY 4.0 license); immediate download upon registration [2].
Protected	High-resolution movement paths; behavioral data; locations of sensitive species (e.g., endangered).	Ecological disturbance; poaching; harassment.	Embargoed Access; requires formal data request and justification; possible time-bound embargo [13].
Restricted	Precise locations of nesting sites, dens, or breeding colonies; data on species of high conservation concern.	Population-level harm; habitat disruption.	Managed Access; requires specific data use agreement; project-specific restrictions; may involve data owner coordination [2].
Private	Raw, unprocessed sensor data; data subject to ongoing analysis for thesis or publication.	Pre-publication scooping; invalid conclusions from unvetted data.	Private/Closed; metadata may be public but data is accessible only to the owner and designated collaborators [2].

Technological Implementation and Platform Solutions

Modern data platforms provide the technical infrastructure to enforce the data sensitivity tiers described above. These platforms facilitate standardization, storage, and granular access control.

Standardised Platforms for Biologging Data

The Biologging intelligent Platform (BiP) is an integrated system designed to store standardized sensor data alongside rich metadata. A key feature of BiP is its flexible access model:

• Data owners can interactively upload data and choose between open and private settings.
• Metadata and visualized route maps are accessible to all users, regardless of the data's open or private status, enhancing discoverability.
• For private datasets, interested users can contact the data owner directly to request permission, creating a managed access channel [2].

Similar capabilities are embedded within platforms like Movebank, which manages billions of animal location records. The technical protocol for setting access rights on such platforms typically involves a dropdown menu or checkbox interface during the upload or dataset management phase, allowing the owner to select from visibility states such as "Public," "Embargoed," or "Private."

The Critical Role of Metadata Standardization

Interoperability and effective data sharing rely on community-approved standards. Adhering to frameworks like those proposed by the International Bio-logging Society's Data Standardisation Working Group is essential [13]. Standardized metadata ensures that data, whether public or private, can be discovered, understood, and reused.

Table 2: Essential Metadata Classes for Managed Access Biologging Datasets

Metadata Class	Key Elements	Standards & Formats	Function in Access Management
Animal Traits	Species, sex, body mass, life history stage.	ITIS (Integrated Taxonomic Information System)	Enables filtering and responsible use; e.g., hiding data for sensitive demographic groups.
Instrument & Deployment	Device type, manufacturer, attachment method, deployment location/date.	Custom, but aligned with Climate and Forecast (CF) conventions.	Critical for data quality assessment and interpreting sensor limitations.
Data Collection	Sensor parameters, sampling frequency, calibration information.	Attribute Convention for Data Discovery (ACDD)	Allows for correct data fusion and analysis across different studies and devices.
Project & Access	Principal investigator, funding source, data license, embargo period.	ISO (International Organization for Standardization)	Directly encodes the terms of use and access restrictions for the dataset.

The movepub software package, for instance, provides a practical tool for preparing Movebank data for publication, ensuring these metadata standards are met before a dataset is shared publicly or with restricted groups [13].

Experimental Protocol: Implementing a Tiered Data Access Workflow

The following protocol outlines the steps for a research group to process a raw biologging dataset and publish it under a managed access model on a platform like BiP or Movebank.

Objective: To transform a raw device output into a standardized, archived dataset with appropriate access controls. Primary Output: A discoverable dataset with rich metadata, where data access is tiered according to its sensitivity.

Protocol Steps

• Data Download and Initialization

  • Retrieve the raw data from the biologging device using the manufacturer's software.
  • Create a new project directory on a secure server with the structure: /Raw_Data, /Processing_Scripts, /Metadata, /Outputs.
  • Safety Consideration: Store raw data in the /Raw_Data folder, which should have restricted access (e.g., only accessible to the core research team).

• Data Cleaning and Standardization

  • Using a script (e.g., in R or Python), read the raw data file.
  • Perform quality checks: filter implausible coordinates (e.g., speed filters), identify and handle gaps in data, and correct time zones to a universal standard (UTC).
  • Standardize column names and data formats to align with community standards (e.g., "location-lat" and "location-long" as per the framework in Sequeira et al. [60]).
  • Export the cleaned, standardized data to a new file in the /Outputs directory.

• Metadata Compilation

  • Complete three standardized metadata tables (as in BiP's model [2]):
    • Animal Metadata: Species (using ITIS taxonomy), sex, body mass, age class.
    • Device Metadata: Device type, manufacturer, serial number, calibration dates.
    • Deployment Metadata: Deployment timestamp, retrieval timestamp, attachment method, geographical location of deployment.
  • Save these tables as CSV files in the /Metadata folder.

• Sensitivity Assessment and Access Tier Assignment

  • Convene a meeting with the research team and relevant stakeholders (e.g., conservation managers).
  • Review the cleaned data and its context against the criteria in Table 1.
  • Decision Point:
    • If the data involves a critically endangered species or a precise breeding site, assign to Restricted.
    • If the data is from a sensitive species but from a non-breeding season, assign to Protected.
    • If the data is from a common species with no sensitive context, assign to Public.
  • Document the justification for the chosen tier in the project's data management plan.

• Platform Upload and Access Configuration

  • Log in to the chosen data repository (e.g., BiP, Movebank).
  • Create a new study and upload the cleaned data file from the /Outputs directory.
  • Input the compiled metadata using the platform's forms, utilizing pull-down menus where available to ensure consistency [2].
  • In the platform's "Sharing" or "Access" settings, select the appropriate tier:
    • For Public: Apply an open license (e.g., CC BY 4.0).
    • For Protected/Embargoed: Set an embargo period (e.g., 2 years from collection) and require user registration.
    • For Restricted: Set the dataset to "Private" and configure the system to allow users to contact you for access requests.

The Scientist's Toolkit: Essential Research Reagent Solutions

Successful implementation of managed access frameworks relies on a combination of software tools, data standards, and policy documents.

Table 3: Key Research Reagent Solutions for Managed Data Access

Tool/Reagent	Type	Primary Function	Access Management Relevance
Biologging intelligent Platform (BiP)	Software Platform	Integrated platform for storing, sharing, visualizing, and analyzing biologging data.	Provides built-in functionality for setting datasets as open or private, and manages user requests for access [2].
Movebank	Data Repository	A global repository for animal tracking and other biologging data.	Offers granular, study-level permissions for defining user roles (viewer, downloader, editor) and setting data visibility [13].
ETN R Package	Software Tool	An R package to access data from the European Tracking Network.	Demonstrates how standardized API access can be programmed for authorized users, a model for managed data retrieval [13].
movepub	Software Tool	An R package to prepare Movebank data for publication.	Ensures data and metadata meet quality and standardization requirements before being shared under any access level [13].
CC BY 4.0 License	Legal Document	A creative commons license requiring attribution.	The standard license for "Public" tier data, permitting reuse while ensuring attribution [2].
Data Use Agreement (DUA)	Legal Document	A formal contract outlining the terms and conditions for using a dataset.	The primary mechanism for governing access to "Restricted" and some "Protected" tier data, enforcing responsible use [2].

Balancing data openness and protection is not a one-size-fits-all endeavor but a dynamic process requiring careful judgment. By implementing the tiered sensitivity framework, utilizing modern data platforms, and adhering to community standards outlined in this protocol, researchers can navigate this complex landscape. This approach maximizes the collaborative and conservation potential of biologging data while rigorously upholding responsibilities to protect sensitive biological information and respect the contributions of data creators.

Ensuring Data Quality, FAIRness, and Cross-Platform Interoperability

Rigorous Validation Techniques for Supervised Machine Learning Models

The application of supervised machine learning (ML) to biologging data, particularly animal-borne accelerometry, has revolutionized our ability to decipher fine-scale animal behaviors at unprecedented scales [61] [51]. However, this powerful approach brings forth significant challenges in model reliability and generalizability. Within the broader context of archiving and sharing biologging data, rigorous validation is not merely a technical step but a fundamental requirement for ensuring that shared data and models are robust, reproducible, and truly useful for the scientific community.

A core challenge is overfitting, where a model over-adapts to the training data, memorizing specific instances rather than learning the underlying generalizable patterns [51]. Such models may appear highly accurate during training but perform poorly on new, unseen data, directly undermining the value of shared models and datasets. A systematic review of 119 studies using accelerometer-based supervised ML revealed that 79% did not employ adequate validation techniques to robustly identify potential overfitting [61] [51]. This highlights a critical gap in the current research practices and underscores the urgent need for standardized protocols to ensure the quality and reliability of biologging research.

Core Principles: Overfitting and Data Leakage

Understanding Overfitting

In machine learning, overfitting occurs when a model's complexity approaches or surpasses the complexity of the data itself [51]. This causes the model to capture not only the underlying signal but also the noise and specific nuances of the training dataset. The tell-tale sign of an overfit model is a significant drop in performance between the training set and an independent test set, indicating low generalizability to new data [51].

The Peril of Data Leakage

Data leakage arises when the evaluation set is not kept fully independent from the training process, allowing information from the test set to inadvertently influence the model during training [51]. This compromises the validity of the performance evaluation because the model is assessed on data that is more similar to the training data than truly unseen data would be. Consequently, data leakage masks the effects of overfitting and leads to a significant overestimation of a model's real-world performance [51].

Table 1: Common Pitfalls in Machine Learning Validation for Biologging

Pitfall	Description	Consequence
Non-independent Test Set	Test data is not properly isolated during training and/or feature selection.	Masks overfitting, inflates performance estimates.
Non-representative Data Splitting	Training and test sets do not represent the overall data distribution (e.g., splitting data from a single individual).	Poor generalizability to new individuals or conditions.
Incorrect Hyperparameter Tuning	Hyperparameters are tuned directly on the test set rather than a dedicated validation set.	Optimistic bias, as the test set is no longer independent.
Inappropriate Performance Metrics	Reliance on metrics that do not reflect the class imbalance or the biological question.	Misleading interpretation of model utility.

Standardized Validation Protocols

To address these challenges, we propose the following standardized workflow for the rigorous validation of supervised ML models in biologging. This protocol is designed to be applicable to a wide range of biologging data, including accelerometry, and aligns with efforts to promote data standardization and sharing [13].

Workflow for Rigorous Model Validation

The following diagram outlines the core workflow for partitioning data to ensure a robust validation process.

Detailed Experimental Protocol

This section provides a step-by-step methodology for implementing the validation workflow described above.

Protocol 1: Nested Cross-Validation with a Holdout Set

1. Objective: To train a supervised ML model for behavior classification from biologging data (e.g., accelerometry) and obtain a robust, unbiased estimate of its performance on unseen data.

2. Experimental Principles: The protocol is designed to prevent overfitting and data leakage by strictly separating data used for training, model selection, and final evaluation. It combines the strengths of cross-validation for reliable model development and a holdout set for final performance assessment.

3. Reagents and Materials:

• Labeled Biologging Dataset: A dataset containing sensor data (e.g., tri-axial acceleration) and corresponding ground-truthed behavior labels.
• Computing Environment: Python with scikit-learn, R, or equivalent ML framework.
• Hardware: Standard desktop or high-performance computing cluster, depending on data size and model complexity.

4. Procedure:

5. Data Analysis:

• Report performance metrics (e.g., Accuracy, Precision, Recall, F1-Score) for both the cross-validation and the final holdout test.
• A significant performance drop from cross-validation to the holdout test is a strong indicator of overfitting or data leakage during the earlier stages.
• Generate a confusion matrix based on the holdout test set predictions to identify specific behaviors the model struggles to distinguish.

The Scientist's Toolkit: Essential Research Reagents and Materials

The following table details key solutions and materials required for implementing rigorous ML validation in biologging research.

Table 2: Essential Research Reagents and Computational Tools

Item Name	Function/Description	Example/Note
Labeled Accelerometer Data	The fundamental reagent for supervised learning. Consists of raw acceleration signals paired with ground-truthed behavior labels.	Labels obtained via direct observation, video synchronisation, or captive surrogate training [51].
Data Standardization Tools	Software and protocols to format, annotate, and share data consistently, enabling reuse and collaboration.	Tools from the Data Standardisation Working Group [13]; ETN package for tracking data [13].
Computational Framework	The programming environment for implementing ML algorithms and validation routines.	Python's scikit-learn, R's caret or tidymodels.
Hyperparameter Tuning Libraries	Tools that automate the search for optimal model parameters.	Scikit-learn's GridSearchCV or RandomizedSearchCV.
Performance Metric Suite	A set of functions to quantitatively evaluate model performance from different angles.	Includes accuracy, precision, recall, F1-score, and Cohen's kappa.
Data Reuse Information (DRI) Tag	A machine-readable metadata tag for public data, indicating the creator's preference for contact before reuse [62].	Associated with an ORCID, the DRI tag facilitates equitable data reuse and collaboration [62].

Integration with Bio-logging Data Archiving and Sharing

The push for rigorous ML validation must be integrated into the broader movement for standardized data archiving and sharing in biologging. Inadequate validation not only produces unreliable models but also degrades the value of shared datasets, as subsequent users cannot trust the models built upon them.

The proposed Data Reuse Information (DRI) tag is a key innovation in this space [62]. By associating a machine-readable tag with public sequence data (and potentially other data types), it clarifies the conditions for reuse and provides a mechanism for data consumers to contact creators. This fosters a collaborative environment where rigorous validation is the norm, as it builds trust between data creators and consumers. Adopting such standards, alongside the FAIR data principles, ensures that biologging data collections function as living archives of animal life on Earth [13] that support reproducible and impactful science.

The guidelines and protocols outlined here provide a concrete path toward this goal, ensuring that the machine learning models powering the next generation of biological discovery are as robust and reliable as the data they analyze.

The FAIR Guiding Principles—Findable, Accessible, Interoperable, and Reusable—were established in 2016 to provide a framework for enhancing the reusability of scientific data holdings and improving the capacity of computational systems to automatically find and use data [63]. In the specific context of biologging research, which generates complex multi-modal data through animal-attached tags and sensors, implementing FAIR principles addresses the critical challenge of ensuring that valuable tracking, physiological, and environmental datasets remain discoverable, interpretable, and useful beyond their initial collection purpose [64]. The FAIR principles emphasize machine-actionability—the capacity of computational systems to find, access, interoperate, and reuse data with minimal human intervention—which is particularly important given the increasing volume, complexity, and creation speed of biologging data [65].

Effective data management is crucial for scientific integrity and reproducibility, especially in biologging research where studies often involve intricate protocols, extensive metadata, and large datasets [64]. Biologging data reuse is essential for various purposes including repurposing, meta-analyses, longitudinal studies, predictive modeling, and training machine learning algorithms [64]. Proper FAIR data management supports these goals by embedding metadata, provenance, and context to help research teams track how data was collected, processed, and interpreted [63].

Core FAIR Principles and Their Implementation

The Four Components of FAIR

The FAIR principles are defined by four interdependent components, each with specific requirements for implementation:

• Findable: Data should be easy for both humans and computers to discover. This requires assigning globally unique and persistent identifiers (such as DOIs) to all datasets and ensuring they are indexed with rich, machine-actionable metadata [65] [63]. The first step in (re)using data is finding them, making findability an essential foundation for all other FAIR principles.

• Accessible: Data must be retrievable by users through standardized communication protocols, even when behind authentication and authorization layers [65] [63]. The accessible principle emphasizes that once a user finds the required data, they should know how to access them, potentially including authentication and authorization procedures.

• Interoperable: Data needs to be machine-readable and compatible with other systems and formats beyond the initial experimental environment [65] [63]. This requires describing data using standardized vocabularies and ontologies, and storing them in formats that can be seamlessly combined with other datasets, which is particularly important for integrating diverse biologging data types.

• Reusable: Data must be easily replicated and studied in new contexts [65] [63]. This necessitates clarity on licensing and usage rights, robust documentation of data provenance and quality, and annotation with rich, well-described metadata. Reusability represents the ultimate goal of FAIR, optimizing the potential for data to be repurposed in different settings.

FAIR vs. Open Data

A critical distinction exists between FAIR data and open data. FAIR data is focused on making data findable, accessible, interoperable, and reusable, not necessarily publicly available [63]. FAIR principles aim to ensure data is well-structured, richly described, and machine-actionable to maximize utility in complex research environments.

In contrast, open data is made freely available for anyone to access, use, modify, and share without restrictions [63]. While open data serves the public good through free accessibility, it may not always be compatible with privacy rules, intellectual property protection, and other governance restrictions—particularly relevant for biologging data involving endangered species or sensitive locations.

Table 1: Comparison of FAIR Data and Open Data

Aspect	FAIR Data	Open Data
Primary Focus	Data structure, metadata, and machine-actionability	Unrestricted public access
Access Restrictions	Can include authentication and authorization	Generally none
Main User	Computational systems and researchers	Anyone
Metadata Requirements	Rich, structured, and standardized	Variable
Compatibility with Privacy Rules	Yes, through access controls	Limited

FAIR Assessment Framework and Metrics

Implementing a FAIR Assessment Tool

Systematic assessment of FAIR implementation requires structured evaluation tools. Research has demonstrated the development of an 11-item questionnaire with strong internal consistency (Cronbach's α = 0.85-0.87) to evaluate the FAIRness of research data [66]. This tool groups questions according to the four FAIR attributes, providing a quantitative means to assess compliance levels.

The assessment framework can be implemented through a structured workflow that guides researchers through the evaluation process for each FAIR component, identifies gaps, implements improvements, and verifies compliance through iterative refinement.

Quantitative FAIR Metrics

Implementation of FAIR principles can be measured using standardized metrics that assess compliance across the four FAIR components. The following table summarizes key metrics and assessment methods for evaluating FAIR implementation in biologging data:

Table 2: FAIR Implementation Metrics and Assessment Methods

FAIR Principle	Assessment Metric	Measurement Method	Target Compliance
Findable	Persistent identifiers	Check for DOI or other persistent ID	100% of datasets
	Rich metadata	Metadata completeness score	>80% required fields
	Searchable indexing	Repository indexing verification	Fully indexed
Accessible	Standardized retrieval	Protocol compliance check	HTTPS/API available
	Authentication clarity	Access procedure documentation	Clear access pathway
	Metadata persistence	Metadata accessibility after data	Always available
Interoperable	Vocabulary standards	Ontology usage audit	Standard terms >90%
	Format compatibility	Machine-readability test	Fully machine-readable
	Qualified references	Related resource links	All references valid
Reusable	Usage licenses	License presence and clarity	Clear license present
	Provenance documentation	Provenance completeness	Full workflow documented
	Community standards	Domain standard adherence	Full compliance

Research indicates that implementing such assessment frameworks reveals significant variability in FAIR compliance across different data types, with metadata richness and vocabulary standardization often presenting the greatest challenges [66]. The same study found that structured assessment tools demonstrated strong internal consistency across all FAIR domains (Cronbach's α: Findable=0.85, Accessible=0.87, Interoperable=0.86, Reusable=0.85), supporting their reliability for evaluating FAIR implementation [66].

Experimental Protocol: Implementing FAIR for Biologging Data

Data Management Planning for Biologging Studies

Purpose: To establish a comprehensive Data Management Plan (DMP) that ensures biologging data compliance with FAIR principles throughout the research lifecycle.

Materials and Reagents:

• Computational Infrastructure: Secure data storage with backup capabilities, preferably with 3-2-1 backup strategy (3 copies, 2 media types, 1 offsite)
• Metadata Standards: Domain-specific metadata standards (REMBI for imaging, DwC for biodiversity)
• Persistent Identifier Service: Data repository providing DOIs or other persistent identifiers
• Data Documentation Tools: Electronic lab notebook, readme file templates, ontology browsers

Procedure:

• Pre-Data Collection Planning

  • Identify appropriate metadata standards specific to biologging data types (e.g., REMBI for bioimaging data) [67]
  • Define project-specific vocabulary and ontologies to ensure semantic consistency
  • Establish data organization structure with consistent naming conventions
  • Document anticipated data types, formats, and volumes

• Metadata Specification

  • Create structured metadata templates capturing essential elements:
    • Study metadata: Research questions, hypotheses, experimental design
    • Biosample metadata: Subject information, handling procedures
    • Specimen metadata: Detailed characteristics of tagged subjects
    • Data acquisition metadata: Instrument specifications, calibration data
    • Processing metadata: Algorithms, parameters, software versions [67]

• Data Collection and Documentation

  • Implement automated metadata capture where possible
  • Document all data transformations and processing steps
  • Record instrument calibration and maintenance logs
  • Maintain version control for scripts and analysis code

• Quality Control and Validation

  • Perform regular data quality assessments
  • Verify metadata completeness against templates
  • Conduct syntax validation for structured metadata
  • Implement anomaly detection for data outliers

• Data Publication and Sharing

  • Select appropriate FAIR-compliant repositories
  • Apply persistent identifiers to datasets
  • Define clear usage licenses and access restrictions
  • Publish comprehensive data papers where appropriate

Protocol for FAIRification of Legacy Biologging Data

Purpose: To transform existing biologging data collections into FAIR-compliant resources, maximizing their potential for reuse and integration.

Duration: 4-8 weeks depending on data volume and complexity

Procedure:

• Data Inventory and Assessment

  • Catalog all existing datasets with basic characteristics
  • Assess current storage formats and structures
  • Identify gaps in documentation and metadata
  • Prioritize datasets based on reuse potential and risk of degradation

• Metadata Enhancement

  • Map existing metadata to standardized schemas
  • Extract additional metadata from primary files where possible
  • Create comprehensive readme files following domain templates
  • Standardize vocabulary using community ontologies

• Format Standardization

  • Convert proprietary formats to open, community-standard formats
  • Ensure machine-readability of all data files
  • Implement consistent structural formatting
  • Validate format compatibility with target repositories

• Identifier Assignment and Repository Deposition

  • Select appropriate target repositories for each data type
  • Prepare dataset packages according to repository specifications
  • Obtain persistent identifiers (DOIs) for published datasets
  • Establish clear access protocols and licensing terms

Troubleshooting:

• Incomplete Metadata: Use inference from related datasets and literature to fill critical gaps
• Proprietary Formats: Implement open-source conversion tools with validation checks
• Vocabulary Inconsistencies: Develop crosswalk tables to standardize terminology

Essential Research Reagent Solutions for FAIR Implementation

Successful implementation of FAIR principles requires both technical infrastructure and methodological tools. The following table catalogs essential "research reagent solutions" for biologging researchers implementing FAIR data practices:

Table 3: Essential Research Reagent Solutions for FAIR Data Management

Solution Category	Specific Tools/Resources	Function in FAIR Implementation	Application Notes
Metadata Standards	REMBI [67], DwC, Darwin Core	Provide structured frameworks for metadata annotation	REMBI specifically designed for biological images; essential for interoperability
Persistent Identifiers	DOI, UUID, ARK	Create permanent, unique references to datasets	DOI most widely recognized; required by many repositories and publishers
Data Repositories	BioImage Archive [67], EMPIAR, Zenodo [66]	Provide preservation, access control, and indexing	Domain-specific repositories often provide enhanced curation and standardization
Ontology Services	OBO Foundry, EDAM, EBI Ontology Lookup Service	Standardize terminology and enable semantic interoperability	Critical for cross-dataset integration and machine-actionability
Data Management Tools	Electronic Lab Notebooks, REDCap [66], DataSifter [68]	Support documentation, organization, and privacy-preserving sharing	REDCap provides secure data collection; DataSifter enables privacy protection
Synthetic Data Generators	DataSifter, Synthetic Data Vault (SDV) [68]	Create "digital twin" datasets for sharing while minimizing re-identification risk	DataSifter shows particular strength with longitudinal data [68]
Assessment Tools	F-UJI, ARDC FAIR Self-Assessment Tool [66]	Evaluate FAIR compliance and identify improvement areas	Essential for benchmarking and continuous improvement

Data Management Workflow for Biologging Studies

Implementing FAIR principles requires integration throughout the entire data lifecycle. The following workflow diagram illustrates a comprehensive data management process tailored to biologging research, from project initiation through to data sharing and reuse:

Challenges and Solutions in FAIR Implementation

Common Implementation Barriers

Implementing FAIR principles in biologging research presents several significant challenges:

• Fragmented Data Systems and Formats: Biologging research often involves multiple instruments, platforms, and measurement systems generating data in disparate formats [63]. This fragmentation creates substantial interoperability challenges when attempting to integrate datasets for analysis.

• Lack of Standardized Metadata or Ontologies: Different research teams frequently use different vocabularies for the same concepts, creating semantic mismatches and ontology gaps that hinder data integration [63]. Without consistent terminology, automated systems cannot properly interpret or combine datasets.

• High Cost and Time Investment: Transforming legacy data into FAIR-compliant formats requires substantial resources [63]. The curation process involves both technical expertise and domain knowledge, creating resource constraints particularly for smaller research groups.

• Cultural Resistance and Awareness Gaps: Research teams may lack awareness of FAIR principles or perceive them as burdensome administrative requirements rather than scientific enablers [66] [63]. Traditional academic reward systems often prioritize publication over data sharing, reducing motivation for FAIR implementation.

Practical Implementation Strategies

Addressing these challenges requires systematic approaches:

• Adopt Interoperable Standards and Platforms: Implement standardized protocols, metadata frameworks, and data formats to ensure interoperability across systems [69]. Leverage common platforms that facilitate data analysis, visualization, and integration.

• Develop Comprehensive Data Management Plans: Create detailed Data Management Plans (DMPs) that describe systems used, data flow, management roles and responsibilities, plus methods for back-ups, storage and archiving while ensuring anonymization and privacy [66].

• Utilize Privacy-Preserving Techniques: For sensitive biologging data, employ techniques like statistical obfuscation and synthetic data generation. DataSifter has demonstrated strong privacy protection (0.83 privacy score) while preserving key statistical signals (83.1% confidence interval overlap in regression models) [68].

• Implement Incremental FAIRification: Rather than attempting complete FAIR compliance immediately, prioritize high-value datasets and implement improvements progressively. This iterative approach builds experience while delivering tangible benefits.

The implementation of FAIR principles represents a fundamental shift in how biologging research data is managed, shared, and utilized. By making data Findable, Accessible, Interoperable, and Reusable, researchers can accelerate scientific discovery, enhance research reproducibility, and maximize the value of increasingly complex and costly biologging datasets. The protocols and frameworks presented in this article provide practical pathways for researchers to implement FAIR principles in their specific contexts.

As the volume and complexity of biologging data continue to grow, FAIR implementation will become increasingly essential for extracting maximum scientific insight from research investments. By adopting these practices, biologging researchers can contribute to a more open, collaborative, and efficient research ecosystem that benefits the entire scientific community and ultimately enhances our understanding of biological systems.

The expansion of biologging—the use of animal-borne data loggers—has generated immense volumes of data on animal movement, behavior, physiology, and the surrounding environment. Effectively archiving and sharing this data is critical for advancing ecological research, informing conservation policy, and enabling cross-disciplinary science. This application note examines the capabilities of modern biologging data platforms, with a focus on the types of data managed, the analytical functions provided, and the flexibility of data sharing protocols. We present structured comparisons of platform attributes, detailed protocols for data standardization and benchmarking, and visual workflows to guide researchers in selecting and utilizing these platforms. The findings underscore that adherence to standardized data and metadata formats, coupled with integrated analytics and flexible sharing options, is paramount for transforming dispersed biologging data into a cohesive, living archive of life on Earth [10] [13].

Biologging data offers unparalleled insights into animal life, serving fields from ecology to oceanography. However, the heterogeneity of data formats, sensor types, and collection protocols poses a significant challenge to its integration and reuse. The vision for the future is to establish biologging data collections as dynamic, living archives [10]. Realizing this vision depends on robust benchmarking of the platforms that store, process, and disseminate this data. Key capabilities to assess include the diversity of data types a platform can ingest, the analytical tools it provides for data processing and environmental parameter estimation, and the flexibility it offers for data sharing and access control [2] [3]. This document provides a detailed examination of these capabilities, offering application notes and protocols for researchers engaged in biologging data management.

Platform Capabilities and Data Handling

A critical function of any biologging platform is its ability to handle a wide array of data and metadata types in a standardized manner. This ensures interoperability and facilitates secondary use across disciplines.

Data and Metadata Types

Biologging platforms manage two primary classes of information: raw sensor data and contextual metadata. The sensor data constitutes the core measurements, while metadata provides the essential context that makes the sensor data interpretable and reusable.

Table 1: Common Data Types in Biologging Platforms

Data Category	Specific Data Types	Description
Spatial Data	Latitude, Longitude, Altitude/Depth	Horizontal position and vertical dimension data [2].
Movement Data	Speed, Acceleration (3-axis), Angular Velocity	Kinematic measurements of animal movement and behavior [2].
Environmental Data	Water Temperature, Salinity, Atmospheric Pressure, Light Intensity	Parameters characterizing the animal's physical environment [2] [3].
Physiological Data	Body Temperature, Heart Rate	Metrics reflecting the internal state and physiology of the animal [2].

Table 2: Essential Metadata for Biologging Data

Metadata Category	Examples	Purpose
Animal Metadata	Species, Sex, Body Size, Breeding History	Provides biological context for interpreting sensor data [2].
Instrument Metadata	Device Type, Manufacturer, Sensor Specifications	Details the data source and its technical parameters [2].
Deployment Metadata	Deployment Date/Location, Recapture Date, Researcher Contact	Documents the experimental context and provenance [2].

Platforms like the Biologging intelligent Platform (BiP) address the challenge of format inconsistency by enforcing international standards for both data and metadata, such as the Integrated Taxonomic Information System (ITIS) and Climate and Forecast (CF) Metadata Conventions [2]. This standardization is a foundational step for data integration and preservation.

Data Sharing and Access Flexibility

The value of biologging data is maximized when it can be shared and accessed by a broad community, while respecting the needs and rights of data owners.

Platforms implement different sharing modes:

• Open Data: Freely downloadable datasets, often under licenses like CC BY 4.0, which require attribution [2].
• Private Data: Accessible metadata and visualized route maps, but raw data download requires direct permission from the data owner [2].
• Multi-repository Storage: Linking with other databases to enhance data sustainability and accessibility [2] [42].

This flexible framework supports the FAIR (Findable, Accessible, Interoperable, and Reusable) principles, ensuring data can be both protected and widely utilized [70].

Analytical Capabilities

Beyond data storage, modern platforms integrate analytical tools that extract meaningful biological and environmental information from raw sensor data.

Online Analytical Processing (OLAP)

A key feature of advanced platforms like BiP is the inclusion of OLAP tools. These tools calculate higher-order environmental and behavioral parameters by applying published algorithms to the raw sensor data [2]. For instance:

• Oceanographic Parameters: Surface currents, ocean winds, and wave characteristics can be estimated from the movement data of seabirds and marine animals [2] [3].
• Behavioral Classification: Dive profiles from marine mammals can be automatically classified into foraging, traveling, or resting behaviors based on depth and acceleration patterns.

This integrated analysis transforms raw telemetry data into actionable knowledge for fields like conservation biology and physical oceanography.

Benchmarking as an Analytical Practice

Benchmarking is the conceptual framework for evaluating the performance of computational methods against a defined ground truth [70]. In the context of biologging, this can involve comparing different analytical algorithms for classifying animal behavior or estimating environmental variables.

The process requires:

• Gold Standard Data: A reference dataset with known outcomes, which can be derived from trusted technology (e.g., Sanger sequencing), integration/arbitration of multiple methods, or synthetic mock communities [71].
• Performance Metrics: Well-defined statistical measures (e.g., precision, recall, F1 score) to quantify tool performance [72].
• Formal Workflows: The use of standardized workflow systems (e.g., Nextflow, Snakemake) to ensure the benchmarking process is reproducible and extensible [70].

Table 3: Common Performance Metrics for Benchmarking

Metric	Calculation	Interpretation
Precision	True Positives / (True Positives + False Positives)	The proportion of identified items that are correct.
Recall	True Positives / (True Positives + False Negatives)	The proportion of true items that were successfully identified.
F1 Score	2 * (Precision * Recall) / (Precision + Recall)	The harmonic mean of precision and recall.

Experimental Protocols

This section provides detailed methodologies for key procedures in biologging data management, from initial standardization to performance benchmarking.

Protocol 1: Data Standardization and Upload to a Platform (e.g., BiP)

Objective: To prepare raw biologging data and associated metadata for upload to a standardized platform, ensuring interoperability and reuse.

Materials:

• Raw data files from biologging devices (e.g., CSV, TXT).
• Metadata on animal traits, instrument details, and deployment information.
• Access to a standardized platform (e.g., BiP, Movebank).

Procedure:

• Data Format Check: Review raw data files for consistency. Check column headers, date-time formats (prefer ISO 8601: YYYY-MM-DD HH:MM:SS), and file structure.
• Metadata Compilation: a. Collect Animal Metadata (see Table 2) using platform pull-down menus where available to minimize errors. b. Collect Instrument Metadata including device model, firmware version, and sensor calibration data. c. Record Deployment Metadata such as attachment method, release condition, and geospatial coordinates of deployment.
• Platform Upload: a. Log in to the platform and initiate a new dataset creation. b. Upload the raw sensor data file. The platform may automatically parse and standardize the format. c. Input the compiled metadata into the designated web forms.
• Sharing Settings: a. Designate the dataset as Open or Private based on data ownership and ethical considerations. b. For private data, ensure contact information is accurate for permission requests.
• Validation: Use the platform's visualization tools (e.g., interactive route maps) to perform a sanity check on the uploaded data.

Protocol 2: Benchmarking an Analytical Tool

Objective: To objectively evaluate the performance of a computational tool (e.g., a behavior classifier) against a gold standard dataset.

Materials:

• Gold standard dataset (e.g., a manually annotated dataset of animal behaviors).
• The computational tool(s) to be benchmarked.
• A workflow management system (e.g., Nextflow, Snakemake).
• Computing infrastructure (e.g., high-performance computing cluster).

Procedure:

• Benchmark Definition: a. Formally define the task, input data, and ground truth. b. Specify the software environment (e.g., Docker container) and parameters for each tool. c. Define the performance metrics (e.g., those in Table 3) for evaluation.
• Workflow Execution: a. Execute the tools on the gold standard dataset using the defined workflow. b. Record all outputs, including runtime and resource consumption.
• Results Comparison: a. Compare the tool's output against the ground truth. b. Calculate the predefined performance metrics.
• Analysis and Reporting: a. Aggregate results across multiple datasets or tool versions. b. Generate summary visualizations (e.g., FunkyHeatmaps) to communicate performance [70]. c. Document the entire process, ensuring provenance is tracked for full reproducibility.

Visualization of Workflows

The following diagrams, generated using Graphviz, illustrate the logical relationships and data flows in the key processes described.

Biologging Data Management and Use Workflow

Diagram 1: This workflow outlines the journey of biologging data from its raw state to becoming part of a living archive. Raw data and metadata are ingested by a standardized platform, which enables both analytical processing (OLAP) and flexible data sharing, ultimately contributing to a preserved data collection.

Benchmarking Process for Analytical Tools

Diagram 2: This flowchart details the formal process of benchmarking an analytical tool. It begins with a clear definition of the benchmark and uses gold standard data within a controlled workflow execution to generate performance results, which are then compiled into a report.

The Scientist's Toolkit

Successful biologging data management and analysis relies on a suite of key resources, platforms, and reagents.

Table 4: Essential Research Reagents and Solutions for Biologging Data Science

Item Name	Function	Example/Reference
BiP (Biologging intelligent Platform)	A standardized platform for data storage, sharing, visualization, and analysis (OLAP).	[2]
Movebank	A global database for animal tracking data, supporting management and analysis of massive datasets.	[10] [13]
International Bio-logging Society Standards	Community-developed frameworks for data standardization, ensuring interoperability.	[13]
Gold Standard Datasets	Reference data (e.g., from GIB Consortium, mock communities) for validating and benchmarking tools.	[71] [72]
Workflow Management Systems	Software for creating reproducible and scalable data analysis pipelines (e.g., Nextflow, Snakemake).	[70]
Containerization (Docker/Singularity)	Technology for packaging software and dependencies to guarantee reproducible computational environments.	[70]

Establishing Minimum Reporting Standards for Animal Welfare and Data Quality

Biologging, the use of animal-borne electronic tags to document movements, behavior, physiology, and environments, has rapidly expanded over the past six decades [73]. This growth presents a critical opportunity to build digital archives of animal life but also introduces significant challenges concerning animal welfare and data quality. Device impacts can alter animal physiology, behavior, and demography—the very metrics researchers aim to measure [73]. The establishment of minimum reporting standards emerges as a low-cost, high-impact strategy to address these challenges, promoting both ethical research practices and the collection of high-quality, reusable data [73].

Framed within a broader thesis on archiving and sharing approaches, these standards are foundational for creating interoperable data collections. They ensure that biologging data can fulfill its potential in ecological discovery, conservation, and contributing to fields like oceanography and meteorology [2] [10]. This document outlines application notes and protocols to implement these standards effectively.

Quantitative Foundations: The Impact of Device Deployment

The development of reporting standards is informed by empirical studies on how biologging devices affect studied subjects. The following table summarizes key quantitative findings and impact relationships identified from a comprehensive review of the literature.

Table 1: Quantified Impacts of Biologging Devices and Related Findings

Subject of Measurement	Key Quantitative Finding or Relationship	Source / Context
General Device Impacts	Review of 175 biologging impact studies over 25 years reveals broad, multispecies connections between instrument characteristics and animal physiology, behavior, and/or demography.	[73]
Data Archiving Quality (Completeness)	56.4% of 362 assessed open datasets in ecology and evolution were complete (mean score of 3.4/5).	[74]
Data Archiving Quality (Reusability)	45.9% of the same 362 datasets were reusable (mean score of 3.1/5).	[74]
Temporal Trend in Reusability	Datasets associated with more recent studies were slightly more reusable than older studies, though improvements are slow.	[74]
Data Volume in Movebank	As of January 2025, Movebank, a primary biologging database, contained 7.5 billion location points and 7.4 billion other sensor records across 1,478 taxa.	[2]

Minimum Reporting Standards: A Machine-Readable Checklist

Informed by the documented impacts on welfare and data quality, a minimum reporting standard was distilled into eight best practices [73]. The core of this standard can be implemented as a machine-readable checklist for researchers to include with manuscripts and data submissions.

Table 2: Minimum Reporting Standard Checklist for Biologging Studies

Category	Reporting Field	Required?	Description & Examples	Alignment with Broader Standards
Animal Morphology	Species, Sex, Age, Body Mass	Required	Key traits influencing device impact; use controlled vocabularies (e.g., ITIS).	Integrated Taxonomic Information System (ITIS) [2] [75].
Device Properties	Mass, Dimensions, Attachment Method	Required	Device-to-body mass ratio, attachment type (harness, collar, glue), dimensions.	Informs animal welfare assessment [73].
Deployment Details	Deployment Date, Location, Duration	Required	Context for animal disturbance and data interpretation.	Darwin Core for location data [76] [77].
Data Collection Parameters	Sampled Sensors, Resolution	Required	Sensors used (GPS, accelerometer), sampling frequency, duty cycling.	Movebank Vocabulary [75].
Animal Welfare Assessment	Post-Release Behavior, Device Effects	Recommended	Qualitative/quantitative notes on behavior post-handling, any visible effects.	Promotes ethical review and refinement [73].
Metadata for Reuse	Data Dictionary, License	Required	Explanation of column headers, abbreviations, units; license for reuse (e.g., CC BY).	FAIR Principles [77] [74].

Experimental Protocol: Implementing the Minimum Reporting Standard

This protocol provides a detailed methodology for implementing the minimum reporting standard throughout a biologging study's lifecycle, from pre-deployment planning to data archiving.

Pre-Fieldwork Planning and Preparation

• Device Selection: Select the smallest and lightest feasible device. Document the device mass, dimensions (length, width, height), and attachment method. Adhere to established guidelines (e.g., the 3% rule as a starting point, but recognize its limitations and species-specificity) [73].
• Ethical Review: Include in permit applications a detailed plan for device attachment and monitoring, strategies for device recovery (if applicable), and a protocol for post-release welfare assessment.
• Metadata Schema Definition: Pre-define data tables and a data dictionary using a standardized vocabulary (e.g., the Movebank Attribute Dictionary) to ensure consistency during data collection [75].

In-Field Data Collection and Documentation

• Animal Handling: Record the species (using scientific name), sex, age class, and body mass at capture. Photograph the device attached to the animal for documentation.
• Deployment Log: For each deployment, record the deployment timestamp, geographic coordinates (in decimal degrees), and animal identifier. Note any deviations from the planned procedure.
• Welfare Monitoring: Observe and note the post-release behavior of the instrumented animal (e.g., immediate locomotion, reunion with conspecifics) for a defined period. This serves as a critical, though often qualitative, welfare check.

Post-Fieldwork Data Processing and Archiving

• Data Curation: Clean and format sensor data into a "tidy" structure where each row is an observation and each column is a variable [76]. Use the pre-defined data dictionary to annotate all columns, units, and abbreviations.
• Compile Reporting Checklist: Complete the minimum reporting standard checklist (Table 2) with all required and applicable recommended fields.
• Public Archiving: Deposit the sensor data, complete metadata, and reporting checklist in a FAIR-aligned public repository (e.g., Movebank, BiP, Zenodo) under an open license (e.g., CC BY 4.0) [77] [2] [10]. Assign a persistent identifier (DOI) to the dataset.

Table 3: Key Resources for Biologging Standards and Data Management

Tool / Resource Name	Type	Function & Key Features	Access / Reference
Movebank	Data Repository	Global platform for managing, sharing, and analyzing animal tracking data. Harmonizes data to a shared vocabulary.	[2] [10] [75]
Biologging intelligent Platform (BiP)	Data Repository	Integrated platform for sharing, visualizing, and analyzing standardized biologging data and metadata. Includes Online Analytical Processing (OLAP) tools.	[2]
Minimum Reporting Standard Checklist	Reporting Tool	Machine-readable checklist to standardize reporting of device, deployment, and animal details to promote welfare and data quality.	[73]
Wildlife Disease Data Standard	Data Standard	Example of a parallel minimum data standard with 40 data fields, demonstrating application of the "tidy data" principle for disaggregated data.	[76] [77]
International Bio-Logging Society (IBLS) Data Standardisation WG	Community Body	Community-led group coordinating the development and adoption of data standards and protocols for the biologging community.	[13] [10]
ETN / movepub / etn R package	Software Tools	Example software packages for preparing, accessing, and working with standardized biologging data, particularly from Movebank and the European Tracking Network.	[13]

The field of biologging, which involves attaching data recorders to animals to monitor their movements, behavior, and physiology, generates vast amounts of complex data. The transition from isolated, proprietary data formats to internationally standardized data protocols has been a critical evolution, transforming this data from a specialized biological resource into a powerful, cross-disciplinary asset. This case study examines how standardized biologging data, facilitated by platforms like Movebank and the Biologging intelligent Platform (BiP), now directly fuels advanced ecological research, oceanographic and meteorological science, and evidence-based conservation policy. By establishing common data formats and rich, structured metadata, researchers can now integrate diverse datasets for large-scale meta-analyses, while also providing policymakers with clear, actionable insights derived from robust, reproducible data.

The Evolution and Imperative of Data Standardization in Biologging

The Challenge of Heterogeneous Data

Before widespread standardization, biologging data was characterized by heterogeneity. Inconsistencies in data formats, such as different column names for the same sensor data (e.g., "Latitude" vs. "lat"), variations in date-time formats, differing file types, and disparate numbers of header lines, created significant barriers to collaborative research and secondary data use [2]. These discrepancies often varied by sensor manufacturer, device type, or even software version, making integration and large-scale analysis a manual and error-prone process.

The Drive for Standards

Recognizing these limitations, the international biologging community, spearheaded by the International Bio-logging Society's Data Standardisation Working Group, initiated a concerted effort to develop and promote common protocols [13]. The working group's objective is to "progress standardisation of data protocols used within the bio-logging community, with a view to making databases interoperable" [13]. This has resulted in proposed frameworks and demonstrated standards, such as those outlined by Sequeira et al. (2021), which provide a foundation for storing diverse data types along with associated metadata in a consistent manner [13] [2].

Key Platforms and Protocols Enabling Standardization

The theoretical framework for standardization is implemented through specific platforms and protocols that serve as the infrastructure for cross-disciplinary data sharing.

Comparative Analysis of Major Biologging Data Platforms

Platform Name	Primary Function	Key Standardization Features	Data Access Policy
Biologging intelligent Platform (BiP) [2]	Integrated platform for sharing, visualizing, and analyzing biologging data.	Adheres to international standards for sensor data and metadata (e.g., ITIS, CF, ACDD, ISO); integrated Online Analytical Processing (OLAP) tools.	CC BY 4.0 license for open data; private data requires owner permission; metadata and route maps are publicly viewable.
Movebank [2]	Large-scale database for animal tracking and sensor data.	Manages over 7.5 billion location points with standardized taxon classifications; supports tools like movepub for data publication preparation [13].	Data visibility and access controlled by data owner; used for large-scale collaborative studies and distribution mapping.
European Tracking Network (ETN) [13]	Centralized access to data from the European aquatic animal tracking network.	Provides standardized data access via the etn R package, ensuring consistent data retrieval and formatting for users.	Data access typically requires registration and adherence to network data policies.

Core Metadata Standards Protocol

For data to be truly interoperable, it must be accompanied by rich, structured metadata. The following table details the core metadata classes required for a standardized biologging data submission, as implemented by platforms like BiP [2].

Metadata Class	Description	Example Fields	Relevant Standard
Animal Metadata	Records traits of the individual animal studied.	Species (Scientific Name, Common Name), Sex, Body Mass, Life Stage, Breeding Status.	Integrated Taxonomic Information System (ITIS)
Instrument Metadata	Describes the data-logging device used.	Device Manufacturer, Model, Serial Number, Sensor Types (e.g., GPS, accelerometer, depth).	Climate and Forecast (CF) Metadata Conventions
Deployment Metadata	Documents the context of the device attachment.	Deployment DateTime, Location, Retrieval DateTime, Attachment Method, Data Processing Steps.	Attribute Conventions for Data Discovery (ACDD)

Experimental Protocols for Cross-Disciplinary Data Utilization

Protocol 1: Utilizing Animal-Borne Sensors for Oceanographic Data Collection

Objective: To collect in-situ physical oceanographic data (e.g., temperature, salinity) from marine animals to complement traditional ocean-observation systems [2].

• Device Selection and Configuration:

  • Select a Satellite Relay Data Logger (SRDL) or similar device capable of measuring the desired parameters (e.g., depth-temperature profiles).
  • Configure the device's sampling regime (e.g., record a temperature-depth profile every 4 hours). Devices must be calibrated according to manufacturer and oceanographic community standards.

• Animal Deployment:

  • Deploy the configured device on a target marine species (e.g., seals, sea turtles, sharks) using species-appropriate, ethically-approved attachment methods.
  • Record all relevant Animal, Instrument, and Deployment Metadata as defined in Section 2.2.

• Data Transmission and Archiving:

  • The device stores and compresses essential data, transmitting it via satellite while the animal is at the surface.
  • Upon receipt, raw data is ingested into a standardized platform like BiP or the Animal Borne Ocean Sensors (AniBOS) project database.
  • Data is automatically standardized and linked with its full metadata.

• Data Processing and Quality Control:

  • Use integrated OLAP tools within BiP or other platforms to calculate derived environmental parameters.
  • Apply quality control flags to identify and filter potentially erroneous data points (e.g., data from periods when the animal is on land).
  • The final, quality-controlled dataset is made available for oceanographic model assimilation and analysis.

Protocol 2: Meta-Analysis of Animal Movement for Conservation Policy

Objective: To synthesize biologging data from multiple studies and species to identify critical habitats, migration corridors, and anthropogenic threat areas to inform marine spatial planning and policy [2].

• Research Question and Data Discovery:

  • Define a clear policy-relevant question (e.g., "Identify areas of high overlap between shipping lanes and whale shark aggregations in the Pacific").
  • Use platforms like Movebank and BiP to discover relevant, standardized datasets. Search using filters for species, region, and sensor type.

• Data Access and Integration:

  • For datasets with open access (e.g., CC BY 4.0), download directly.
  • For private datasets, contact the data owner to request access, outlining the project's policy goals.
  • Integrate datasets using their standardized formats. The common structure of fields (e.g., individual_id, timestamp, location_lat, location_long) allows for seamless merging.

• Spatio-Temporal Analysis:

  • Calculate utilization distributions and kernel density estimates to map critical habitats and movement corridors for each species.
  • Perform overlap analysis with human-use layers (e.g., shipping lanes, fishing effort maps) in a Geographic Information System (GIS).

• Policy Reporting and Visualization:

  • Synthesize findings into a policy brief with clear, accessible visualizations, such as heatmaps of animal density and conflict zones.
  • Ensure all data sources are cited appropriately, giving credit to both the original publications and the data packages, as per best practices for data re-use [78].

Essential Research Reagent Solutions

The following table lists key hardware and software "reagents" essential for conducting modern, standardized biologging research.

Item Name / Category	Function / Application	Key Specifications
Satellite Relay Data Logger (SRDL) [2]	Transmits compressed data (dive profiles, temperature) via satellite; enables long-term, remote data collection without device retrieval.	Sensors: Depth, Temperature, Salinity; Communication: Argos/Iridium; Deployment Duration: >1 year.
ORI400-PD3GT (Little Leonardo) [79]	High-resolution data logging for marine species. Records data for later retrieval.	Sensors: Swimming Speed, Depth, Temperature, 3-Axis Acceleration; Memory: Internal Storage.
LoggLaw G2 (Biologging Solutions) [79]	Terrestrial/freshwater animal tracking with direct communication via cellular networks.	Sensors: GPS; Communication: LTE-M; Power: Rechargeable battery.
DEBUT FLEX II (Druid tech) [79]	Flexible tracking device for a variety of mid-sized terrestrial animals.	Sensors: GPS, Temperature, Illuminance, ODBA; Communication: 4G.
movepub R Package [13]	Software tool to prepare and standardize tracking data from Movebank for publication and archiving.	Function: Data cleaning, format standardization, metadata generation.
etn R Package [13]	Provides programmatic, standardized access to data from the European Tracking Network (ETN).	Function: Data extraction, integration, and analysis within the R environment.

Data Flow and Interoperability Architecture

The following diagram illustrates the workflow from raw data collection to cross-disciplinary application, enabled by standardization.

Impact and Future Directions

The implementation of these standardized protocols has tangible impacts. Biologging data from seals and turtles now provides ocean temperature data comparable in volume to Argo floats in specific regions like the Antarctic and Arctic, filling critical observational gaps [2]. Initiatives like the AniBOS project formalize this contribution as part of the Global Ocean Observing System. In conservation, the ability to integrate dozens of datasets has enabled comprehensive mapping of species distributions for identifying Marine Protected Areas [2]. The future of the field depends on sustaining this momentum through continued community coordination, development of sustainable funding models for long-term data curation, and the creation of robust incentive structures that reward researchers for publishing high-quality, standardized data [13].

Conclusion

Effective archiving and sharing are no longer optional but fundamental to unlocking the full potential of biologging data. By adopting standardized formats, utilizing dedicated platforms like BiP and Movebank, and implementing rigorous validation protocols, researchers can transform isolated datasets into a cohesive, global resource. These practices directly address current challenges, from model overfitting to geographic data biases, thereby enhancing the reliability and scope of research findings. The future of biologging lies in its integration with global biodiversity monitoring frameworks and its expanding utility in fields like drug discovery, where animal-borne environmental data can provide unique insights. Embracing these collaborative and robust data management strategies is paramount for accelerating scientific discovery, informing effective conservation policies, and fostering a truly open scientific ecosystem.

References

• 1. Establishing bio-logging data collections as dynamic ... [https://www.vlaanderen.be/inbo/en-GB/publications/establishing-bio-logging-data-collections-as-dynamic-archives-of-animal-life-on-earth]
• 2. Biologging intelligent Platform (BiP): an integrated and ... [https://movementecologyjournal.biomedcentral.com/articles/10.1186/s40462-025-00551-8]
• 3. Biologging as a potential platform for resolving ocean ... [https://www.sciencedirect.com/science/article/pii/S2772735125000265]
• 4. Protecting Earth's oldest data system: the case for biodiversity [https://news.mongabay.com/2025/10/protecting-earths-oldest-data-system-the-case-for-biodiversity/]
• 5. Biomedical Data Repository Concepts and Management ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11176378/]
• 6. Protocol sharing for open research | Nature Cell Biology [https://www.nature.com/articles/s41556-025-01666-1]
• 7. Best practices – International Bio-Logging Society [https://bio-logging.net/wg/best-practices/]
• 8. NIH's Public Access & Data Sharing Policies [https://mblwhoilibrary.org/nihs-public-access-data-sharing-policies-what-the-woods-hole-science-community-needs-to-know/]
• 9. Guidelines for Publicly Archiving Terrestrial Model Data to ... [https://datascience.codata.org/articles/10.5334/dsj-2022-003]
• 10. Establishing bio-logging data collections as dynamic ... [https://pubmed.ncbi.nlm.nih.gov/39753915/]
• 11. 2025 Mid-Year Review: US State Privacy Law Updates ... [https://www.mayerbrown.com/en/insights/publications/2025/10/2025-mid-year-review-us-state-privacy-law-updates-part-2]
• 12. July 2025 Privacy Compliance Countdown: Key Deadlines ... [https://www.hinshawlaw.com/en/insights/privacy-cyber-and-ai-decoded-alert/pcad-july-2025-state-privacy-law-compliance-doj-bulk-sensitive-data-rule]
• 13. Data Standardisation Working Group [https://bio-logging.net/wg/data-standardisation/]
• 14. Establishing bio-logging data collections as dynamic ... [https://www.nature.com/articles/s41559-024-02585-4]
• 15. A brief introduction to the analysis of time-series data from ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8237163/]
• 16. FEM Leads in Bio-Logging Data Collection, Dynamic ... [https://fmach.it/en/Communication/Press-releases/FEM-Leads-in-Bio-Logging-Data-Collection-Dynamic-Archives-of-Animal-Life-on-Earth3]
• 17. Ten simple rules to colorize biological data visualization - PMC [https://pmc.ncbi.nlm.nih.gov/articles/PMC7561171/]
• 18. Five Steps to a Diverging and Contrasting Color Scheme for ... [https://nightingaledvs.com/five-steps-to-a-diverging-and-contrasting-color-scheme-for-data-visualization/]
• 19. Biodiversity data sharing platforms are vital for the ... [https://link.springer.com/article/10.1007/s10531-025-03058-1]
• 20. Data archiving in ecology and evolution: best practices [https://www.sciencedirect.com/science/article/abs/pii/S0169534710002697]
• 21. Optimising the use of bio-loggers for movement ecology ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7041970/]
• 22. Integrating animal tracking and trait data to facilitate global ... [https://pubmed.ncbi.nlm.nih.gov/39973193/]
• 23. Biological data derived from European weather radars [https://www.nature.com/articles/s41597-025-04641-5]
• 24. Data models and standards for improved usability [https://discourse.gbif.org/t/data-models-and-standards-for-improved-usability/3476]
• 25. Global Biodiversity Information Facility (GBIF): Supporting ... [https://speakerdeck.com/peterdesmet/global-biodiversity-information-facility-gbif-supporting-policy-through-data]
• 26. The Movebank data model for animal tracking [https://www.sciencedirect.com/science/article/abs/pii/S1364815210003257]
• 27. MDR :: FAQ - Movebank Data Repository [https://datarepository.movebank.org/help/faq]
• 28. Which data can be shared through GBIF and what cannot [https://data-blog.gbif.org/post/data-shareability/]
• 29. Mobilizing animal GPS tracking data to Movebank and GBIF [https://zenodo.org/records/7408643]
• 30. Archiving animal movements as a record of biodiversity [https://oscibio.inbo.be/blog/move2gbif/]
• 31. Mobilizing animal GPS tracking data to Movebank and GBIF [https://pureportal.inbo.be/en/projects/move2gbif-mobilizing-animal-gps-tracking-data-to-movebank-and-gbi/]
• 32. Darwin Core [https://www.tdwg.org/standards/dwc/]
• 33. Data standards [https://www.gbif.org/standards]
• 34. Darwin Core [https://dwc.tdwg.org/]
• 35. Ten recommendations for organising bioimaging data ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10938051/]
• 36. ( Biologging ): an integrated and standardized... intelligent Platform BiP [https://pmc.ncbi.nlm.nih.gov/articles/PMC11959716/]
• 37. 3 Data visualization – Modern Statistics for Modern Biology [https://www.huber.embl.de/msmb/03-chap.html]
• 38. Effective Data Visualization in Life Sciences: Tools, charts ... [https://lifesciences.enago.com/blogs/effective-data-visualization-in-life-sciences-tools-charts-and-best-practices]
• 39. Transforming conservation with real-time data of individual ... [https://ourenvironment.berkeley.edu/news/transforming-conservation-real-time-data-individual-animals]
• 40. Burden of a failed error culture in biologging [https://www.sciencedirect.com/science/article/pii/S000334722500315X]
• 41. Processing and visualising association data from animal ... [https://animalbiotelemetry.biomedcentral.com/articles/10.1186/s40317-015-0065-4]
• 42. Newly published: Preserving bio-logging data for the future of ... [https://www.sequeiralab.com/news/preserving-bio-logging-data-for-the-future-of-wildlife/]
• 43. A General Primer for Data Harmonization | Scientific Data [https://www.nature.com/articles/s41597-024-02956-3]
• 44. Pandas - Different Ways of Formatting Column Headers [https://medium.com/swlh/pandas-different-ways-of-formatting-column-headers-aec6889d1881]
• 45. Harmonization Protocols [https://www.maelstrom-research.org/harmonized-datasets]
• 46. Automatically standardizing values in a column [https://help.qlik.com/talend/en-US/data-preparation-user-guide/8.0/automatically-standardizing-values-in-column]
• 47. Approaches to protocol standardization and data ... - PMC [https://pmc.ncbi.nlm.nih.gov/articles/PMC11245389/]
• 48. What is Overfitting? | IBM [https://www.ibm.com/think/topics/overfitting]
• 49. Underfitting vs. Overfitting in Machine Learning [https://www.ayadata.ai/a-guide-to-overfitting-and-underfitting-in-machine-learning/]
• 50. How to Fix Overfitting and Underfitting | by Tahir | Oct, 2025 [https://medium.com/@tahirbalarabe2/how-to-fix-overfitting-and-underfitting-4cf8f383ea96]
• 51. Practical guidelines for validation of supervised machine ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12214441/]
• 52. Inferring an animal's environment through biologging [https://movementecologyjournal.biomedcentral.com/articles/10.1186/s40462-020-00228-4]
• 53. Prevent overfitting and imbalanced data with Automated ML [https://learn.microsoft.com/en-us/azure/machine-learning/concept-manage-ml-pitfalls?view=azureml-api-2]
• 54. Overfitting in Machine Learning Explained [https://aws.amazon.com/what-is/overfitting/]
• 55. Overfitting and Underfitting 2025 Avoid the Biggest Mistakes [https://w3schools.cloud/avoid-the-biggest-overfitting-and-underfitting/]
• 56. Predictive overfitting in immunological applications: Pitfalls ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10498807/]
• 57. Long-Term Digital Storage and Usage of Research Data [https://link.springer.com/chapter/10.1007/978-3-031-62332-5_9]
• 58. The Future of DNA Storage in Revolutionizing Biological ... [https://www.icck.org/article/abs/jaib.2025.924847]
• 59. Design considerations for advancing data storage with ... [https://www.sciencedirect.com/science/article/pii/S2590006422001041]
• 60. Standardisation of bio-logging data is key to advance ... [https://megamove.org/standardisation-of-bio-logging-data-is-key-to-advance-global-conservation-of-marine-megafauna/]
• 61. Practical guidelines for validation of supervised machine ... [https://pubmed.ncbi.nlm.nih.gov/40387610/]
• 62. A roadmap for equitable reuse of public microbiome data [https://www.nature.com/articles/s41564-025-02116-2]
• 63. FAIR data principles: What you need to know [https://tiledb.com/blog/fair-data-principles-explained]
• 64. Best practices for data management and sharing in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11380994/]
• 65. FAIR Principles [https://www.go-fair.org/fair-principles/]
• 66. FAIR data management: a framework for fostering data literacy ... [https://bmcmedresmethodol.biomedcentral.com/articles/10.1186/s12874-024-02404-1]
• 67. FAIR Data Principles [https://globalbioimaging.org/international-training-courses/repository/fair-data]
• 68. Medical data sharing and synthetic clinical data generation [https://www.nature.com/articles/s41746-025-01935-1]
• 69. Data Sharing & Collaboration in Biotech to Drive innovation [https://www.sparity.com/blogs/data-sharing-collaboration-in-biotech-driving-innovation-forward/]
• 70. Building a continuous benchmarking ecosystem in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12614554/]
• 71. Systematic benchmarking of omics computational tools [https://www.nature.com/articles/s41467-019-09406-4]
• 72. Comprehensive benchmarking and ensemble approaches ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5609029/]
• 73. Minimum reporting standards can promote animal welfare ... [https://ecoevorxiv.org/repository/view/6784/]
• 74. Slow improvement to the archiving quality of open datasets ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9114975/]
• 75. Integrating animal tracking and trait data to facilitate global ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11883293/]
• 76. A minimum data standard for wildlife disease research and ... [https://www.nature.com/articles/s41597-025-05332-x]
• 77. Standardizing Wildlife Disease Data for Global Health ... [https://globalbiodefense.com/2025/06/24/wildlife-disease-data-standard/]
• 78. Data Archiving in Ecology and Evolution: Best Practices [https://pubmed.ncbi.nlm.nih.gov/21159406/]
• 79. "BiP Up 2025" Application Guidelines [https://help.bip-earth.com/en/bip-up-application/]