Introduction
For thousands of years, Indigenous knowledge has contributed to sustainable water, land, and natural resource conservation on reservation lands (Hoagland, 2016). Modern environmental and ecological issues have become increasingly ubiquitous owing to global warming, hydroclimatological extremes, and population growth. As these issues become increasingly widespread, there is a growing need for data-driven research, in tandem with water rights and policy studies, to highlight the specific risks Native communities face. Gamble et al. (2016) and Fillmore and Singletary (2019) explained that underserved groups, including Indigenous communities, are most vulnerable to climatic changes. Consequently, there is an urgent need for tribes to prepare for and respond to climatic change impacts and risk exposure in their water systems, natural resources, and ecosystems (Cozzetto et al., 2013).
Water resource management of reservation lands is a complex issue because of the involvement of different regulatory systems at the federal, state, and tribal levels (Royster, 2011). A simple review of the existing Indigenous scientific literature on water resources reveals a strong focus on topics related to Indigenous nations’ sovereignty rights and self-determination in water resource protection (Bark et al., 2012; Bulltail, 2024; Wilson et al., 2021), land and natural resource policies (Lertzman and Vredenburg, 2005; Bulltail and Walter, 2020), environmental and ecological data control and sovereignty (Jennings et al., 2023; Reid et al., 2024), and tribal water rights and governance (Cohn et al., 2019; Craig, 2022). Although numerous studies have focused on the policy, legal, and governance aspects of water resources, comparatively few have conducted data-driven and quantitative case studies of reservation lands.
Previous studies have underscored prominent water issues in tribal lands, including climate change and hydroclimatic extremes (Cozzetto et al., 2013; McNeeley, 2017), water contamination (Chief, Meadow and Whyte, 2016), ecosystem and natural resource degradation (Doyle, Redsteer and Eggers, 2013; Gautam, Chief and Smith, 2013), and lack of access to safe and clean piped water (Murphy et al., 2009; VanDerslice, 2011). These issues underscore the need for reliable databases that can support data-driven studies for developing constructive resolutions. However, numerous Indigenous and non-Indigenous researchers interested in addressing these issues have been constrained by the lack of accessible, high-quality, and consistent long- and short-term Water–Climate–Environment (WCE) databases (Chief et al., 2016; Sarzaeim and Bulltail, 2023). To address this knowledge gap, we aimed to evaluate the availability, discoverability, and (re)usability of WCE data provided by Indigenous data repositories, with the goal of empowering scientists to address sustainability barriers in water resource management on US reservations.
Williamson et al. (2023) clarified that unavailable and inaccessible data coupled with poor data management are notable barriers to advancing environmental research on reservation lands. Data scarcity related to reservation lands may also represent one of the main barriers to the active engagement and involvement of Indigenous and local communities in climate research (David-Chavez and Gavin, 2018). To address these challenges, establishing efficient data system management that fosters the democratization of spatiotemporal WCE data collection, validation, documentation, acquisition, verification, and authentication is crucial (Sarzaeim et al., 2023a), as this approach can enhance data accessibility and support data-driven research in Indigenous contexts. Although the rapid growth of digital libraries and web-based data repositories has led to the generation of numerous data packages (Yue et al., 2012), accessible, well-documented, and well-structured datasets are needed to support efficient natural and water resource management on reservation lands. These data resources will equip researchers, scientists, stakeholders, and decision-makers with tools required to advance sustainable development goals.
Scientists have developed several data principles to enhance the discoverability and accessibility of data, thereby improving the infrastructure of scientific data management systems. Wilkinson et al. (2016) stated that while efficient data management is not a standalone goal, it is a leading requirement in digital democracy and knowledge equity. They introduced guiding principles that promote data Findability, Accessibility, Interoperability, and Reusability (FAIR) for both humans and machines to benefit science and researchers. The FAIR principles serve as a guide for data publishers to support data retrieval and reuse by other researchers. Additionally, in the Indigenous context, Carroll et al. (2021) developed the CARE data principles for Indigenous data governance and data protection as a complement to the FAIR principles. The CARE principles support Collective benefits, Authority to control, Responsibility, and Ethics to foster the benefits of data use and protect data, with a focus on the well-being of Indigenous communities at every step of the scientific data lifecycle. The FAIR and CARE principles offer complementary frameworks that collectively maximize the value of data by ensuring the appropriate and responsible (re)use, particularly in the context of Indigenous data (Carroll et al., 2021). In this approach, development of FAIR Indigenous data is the primary step.
It appears that FAIRification practices have been examined in several scientific domains, including biodiversity science (Lannom, Koureas and Hardisty, 2020), geoscience (Kinkade and Shepherd, 2021), agricultural engineering (Ali and Dahlhus, 2022), and aquatic systems (Bayer et al., 2023). However, a FAIRification framework for Indigenous WCE data that can build a strong foundation for data-driven water research has not been evaluated. Thus, this study aims to assess the level of FAIR principles implemented in Indigenous-led WCE data repositories and propose a specific scientific and community-driven standard FAIRification framework. In this study, Indigenous-led WCE data repositories refer to data repositories that are generated, collected, resourced, or supported by Indigenous communities. To accomplish this goal, we defined three objectives as follows: (1) to explore and review the Indigenous-led WCE data repositories that are publicly available on the Web, (2) to evaluate the data repositories concerning each principle and sub-principle of FAIR, and (3) to recommend a list of FAIRification actions to support the reproducibility of Indigenous WCE data repositories identified in objective 1, for use in water resource research. Figure 1 illustrates the workflow of this study.
Figure 1
Overview of the key stages of this study.
FAIR Data Principles
Advancements in cyberinfrastructure technologies for measuring, collecting, recording, monitoring, storing, archiving, reporting, sharing, and visualizing have generated increasing data across disciplines that require an efficient, rapid-responsive, and user-friendly data management system (Jeong et al., 2019; Roche et al., 2015; Sun et al., 2019; Yang et al., 2010). The inclination toward interdisciplinary studies is another critical stimulus for increasing the effectiveness of multidimensional database management for a broader range of researchers across domains.
In response to large and complex volumes of data, scientific data in any discipline is recommended to follow FAIR principles. The FAIR principles support a basic standard of acceptable data management by data producers/publishers and benefit the entire academic/scholar community regarding knowledge, discovery, and science advancement by ensuring data can be effectively found, accessed, easily integrated, and reused by other researchers (Go-fair.org). Each FAIR primary principle consists of sub-principles developed by Wilkinson et al. (2016) (Figure 2):
Figure 2
Overview of the FAIR principles and their corresponding sub-principles.
Findability requires that (meta)data (i.e., both data and metadata) be easily found by both humans and machines and consists of four sub-principles, including F1: (Meta)data are assigned a globally unique and persistent identifier, F2: Data are described with rich metadata, F3: (Meta)data clearly and explicitly include the identifier of the data they describe, and F4: (Meta)data are registered or indexed in searchable resources.
Accessibility refers to the requirement that once users find the desired data, they should be able to understand how to retrieve the data; this term consists of two sub-principles, including A1: (Meta)data are retrievable by their identifier using a standardized communication protocol that should be open, free, and universally implementable (A1.1) and should allow for authentication/authorization procedures if necessary (A1.2), and A2: Metadata should be accessible, even when the data are no longer available.
Interoperability requires that the (meta)data should be capable of interoperating with other applications/workflows for analysis, storage, and processing and consists of three sub-principles, including I1: (Meta)data should use a formal, accessible, shared, and broadly applicable language for knowledge representation, I2: (Meta)data should use vocabularies that follow FAIR principles, and I3: (Meta)data should include qualified references to other (meta)data.
Reusability refers to (meta)data that must be comprehensively described and documented to ensure that they can be replicated and/or combined in different settings; this function consists of one sub-principle, R1: (Meta)data should be richly described by a plurality of accurate and relevant attributes; in particular, (meta)data should be released with a clear and accessible data usage license (R1.1), (meta)data should be associated with detailed provenance (R1.2), and (meta)data should meet domain-relevant community standards (R1.3).
Methods and Data
Search design
The evaluation of the FAIR implementation level in Indigenous WCE data repositories began by identifying them on the Web (Figure 3). We systematically searched the data using the Google Search Engine in the “Incognito” mode, as recommended by Sarzaeim et al. (2023b), on Google Chrome to prevent interference from cookies and browsing history when finding relevant entries. We used the following strings on Google Search: (‘water’ OR ‘climate’ OR ‘environment’ OR ‘land’) AND (‘Indigenous’ OR ‘Tribal’ OR ‘Native’) AND (‘data’ OR ‘database’ OR ‘information’).
Figure 3
Systematic search flowchart for the active Indigenous Water-Climate-Environment (WCE) data repositories using the Google Search Engine on Google Chrome.
In our last search in February 2024, Google returned over two billion results. Because the search engines ranked the results based on relevance, we screened the first 20 entries for further evaluation. We sought possible databases consisting of digital, numerical, and geospatial data on water quality/quantity variables, climate/weather time series, and ecological and land-use datasets located within the reservation lands. We limited our screening strategy to the first 20 entries after confirming that no further active WCE data beyond this point were found for inclusion as case studies in our research.
For the selection of the relevant web entries, we reviewed the “Home Page” or “About Us” pages of each result and filtered out the data not focused on the US. Next, we categorized the results into two groups: (1) Indigenous and (2) non-Indigenous data repositories. Indigenous data repositories specifically focus on data from tribal/reservation lands, whereas non-Indigenous data repositories are broader and may include Indigenous data as a part of a more general dataset. Given the study’s scope, we excluded the “non-Indigenous” data repositories. After applying the filters, two primary case studies remained for detailed examination: (a) Native Climate (NC) and (b) Native Land Information System (NLIS). Upon the initial review of the NC and NLIS data repositories, we identified overall information related to the scope of the data, including the data types, variables, and developers/sponsors, as outlined below:
Native Climate (NC)
The Native Climate (NC) project was established to support climate change adaptation within tribal communities by rebuilding connections between Indigenous knowledge and Western scientific data (https://native-climate.com). Led by the Desert Research Institute, University of Nevada-Reno Tribal Extension, the Montana Climate Office at the University of Montana, the University of Arizona Tribal Extension, and the USDA Climate Hub, the project is funded by the United States Department of Agriculture-National Institute for Food and Agriculture (USDA-NIFA) and Native Waters on Arid Lands. The NC project provides access to multiple historical and projected climate and weather variables, primarily focusing on temperature and precipitation time series for US reservation lands.
In the NC data repository, under the ‘Climate Data/Projections’ tab, the data folders directory can be accessed via the ‘DOWNLOAD AND ACCESS THE DATA’ button. The data were systematically structured and distributed across 633 zip folders, each corresponding to a distinct tribally controlled region developed by one of 35 Tribal Colleges/Universities (TCUs) responsible for dataset development, based on the last modified date in July 2024. Each .zip folder was then downloaded and extracted. Within each folder, the dataset was stored in an Excel file with the format of ‘.xlsx’ and accompanied by a metadata file in “.pdf” format. The metadata provides details on the data attributes and includes various time series plots for climate projections under different representative concentration pathways for greenhouse gas emissions, covering the projection period from 2015 to 2100, as well as historical simulations from 1950 to 2014. Table 1 presents the key features of the NC data repository.
Table 1
Summary of the NC and NLIS data repositories’ features.
| DATA REPOSITORY | DATA TYPES | VARIABLES | DEVELOPERS | SPONSORS | WEBSITE |
|---|---|---|---|---|---|
| Native Climate(NC) | Agricultural climate variables | Annual precipitation, spring precipitation, summer precipitation, fall precipitation, winter precipitation, average precipitation on wet days, average precipitation on wet days (trace), maximum 3-day precipitation, number of wet days, number of wet days (trace), number of dry days, number of dry days (trace), day of first snow, average temperature, number of days ≥ 100 F, number of days with heat index hazard, growing degree days, frost free days, first day of growing season, last day of growing season, length of growing season, average surface wind speed, number of days with average wind speed ≥ 20 mph | Desert Research Institute, University of Nevada-Reno Tribal Extension, Montana Climate Office, University of Arizona Tribal Extension, and USDA Climate Hub | US Department of Agriculture-National Institute of Food and Agriculture, Native Waters on Arid Lands | native-climate.com |
| Native Land Information System(NLIS) | Surface water, groundwater, hydroclimatic variables, environmental risks and land cover variables | Mean monthly and annual surface flow, lake average storage and capacity, aquifer characteristics, wells characteristics and water table, Palmer Drought Severity Index (PDSI), Extreme Heat Days (EHD), precipitation, Wildfire Hazard Potential (WHP), land cover types and acres, projected mean annual precipitation, change of snow-to-precipitation ratio | Native Land Advocacy Project | Indian Land Tenure Foundation, Native American Agriculture Fund, Village Earth | nativeland.info |
Native Land Information System (NLIS)
The Native Land Information System (NLIS) is facilitated by the Native Land Advocacy Project based in Fort Collins, Colorado, to support Indigenous communities’ sovereign rights to protect their own data, land, and resources (https://nativeland.info). In the NLIS data repository, under the ‘Data Tools/Dashboards’ tab, a diverse collection of data is embedded in separate dashboards for Native demographics, agricultural systems, land use and cover, soil properties and capabilities, water resources, environmental risks, and energy and mineral data. Each dashboard presents the data in various formats, including maps, plots, tables, and images. Additionally, each dashboard is accompanied by descriptive metadata that provide contextual information, statistical summaries, and notes on the limitations or potential shortcomings of the presented data.
In this study, we selected five dashboards relevant to water resource research, including (a) the ‘Surface Waters on Native Lands’ dashboard, featuring four tabs of ‘Map’, ‘Derived Mean Monthly Flow Graph’, ‘Measured Mean Annual Flow Table’, and ‘Lake Storage’, (b) the ‘Ground Water on US Native Lands’ dashboard, featuring three tabs of ‘Map’, ‘Aquifer Data Table’, and ‘Well Data Table’, (c) the ‘Environmental Risks for US Native Lands’ dashboard, featuring four tabs of ‘Environmental Maps’, ‘Timeline’, ‘Metadata’, and ‘Raw Data’, (d) the ‘Land Covers on Native Lands in the Coterminous US’ dashboard, featuring three tabs of ‘Map + Pie’, ‘Change Over Time’, and ‘Table Explorer’, and (e) ‘Precipitation Projections and Winter Trends on US Native Lands’, featuring three tabs of ‘Precipitation Projection Map’, ‘Snow-to-Precipitation Map’, and ‘Data Table’. This selection strategy was pivotal, as it covered a broad spectrum of climatological, hydrological, and environmental parameters and variables and was essential for comprehensively documenting water resource dynamics and interactions on Native lands. Table 1 outlines the features of the NLIS dashboards used in this study.
FAIR assessment
In the following steps, we evaluated the (meta)data of NC and NLIS for compliance with all of the FAIR principles and sub-principles following the FAIR metrics framework formulated by Wilkinson et al. (Go-fair.org; 2018). This qualitative framework is a questionnaire-based approach originally designed to evaluate the utility of the FAIR metrics themselves, rather than to evaluate different data resources. However, for evaluating the NC and NLIS (meta)data, we designed a questionnaire focused on answering this question, ‘What must be provided’? (see Supplementary File 1: Appendix. FAIR Assessment Questionnaire). We employed this questionnaire to evaluate NC and NLIS by comparing the elements that were provided or missing in each data repository with the requirements for FAIR compliance. We formulated our responses, represented in the Results section, using the explanations, meanings, and examples for each FAIR principle and sub-principle provided on Go-fair.org. This approach allowed the authors of this work to perform a detailed assessment of each individual digital resource’s compliance with FAIR principles/sub-principles to identify the strengths and limitations and suggest a specific set of FAIRifications to improve data discoverability, dataset formatting, metadata structure, access protocols, and (meta)data interpretability and comprehensibility.
Results
Findability
F1: The NC and NLIS do not explicitly assign unique and persistent identifiers to each dataset they support.
An identifier is defined as ‘a globally unique name for a resource that can be used to retrieve a description of the resource or to identify an entity like a file or dataset without dereferencing the identifier’ (Ananthakrishnan et al., 2020), such as the Digital Object Identifier (DOI).
Given the rapid global increase in dataset searches, assigning a unique, persistent, resolvable, and thoroughly described identifier that supports both human and machine readability facilitates the discovery process, enables citations in scientific publications, and forms a fundamental part of an efficient data management system (Ananthakrishnan et al., 2020; Berber and Yahyapour, 2017).
F2: The NC and NLIS contain descriptive metadata but not in a consistent, machine-readable format.
In the NC, the metadata file was released in both a ‘.pdf’ format and within the first sheet of the ‘.xlsx’ data file, labeled ‘Metadata’, if the user downloaded the data zip folder for a desired reservation. The NLIS data repository lacks a section explicitly labeled as ‘metadata’ in all the selected dashboards, except for the ‘Environmental Risks for US Native Lands’ dashboard, where a brief plain text description is provided.
Zhu and Cole (2022) argued that formats such as PDF are not designed for machine reading or software processing. Similarly, Awad et al. (2020) criticized Excel spreadsheets as a low-quality format for data, as they are proprietary, the sheets may be interconnected, and they may contain hidden contents and visual effects. Instead, non-proprietary machine-readable formats, such as JSON-DL, XML, and RDF, should be used, as they can be indexed by search engines more effectively and consequently improve data discoverability (Nair and Jeevan, 2004; Wilkinson et al., 2018).
Regarding metadata richness, our observations showed that in NC, both the PDF and ‘.xlsx’ metadata provided information on data provenance, data generation methods, temporal coverage, and contact email addresses. The PDF metadata were more explanatory and included more details on methodologies, data provenance, variable definitions and units, and data sponsors. Additionally, the PDF metadata included a wide range of variable time series plots.
In NLIS, the ‘Metadata’ tab of the ‘Environmental Risks for US Native Lands’ dashboard provides an overview of the methods, data provenance, spatiotemporal coverage, and the main references for data development. In the other dashboards, text-based metadata are provided on the same data pages explaining the methods, importance of the datasets, and data limitations.
Metadata plays a crucial role in fostering user understanding of the data, including relevant processes, techniques, tools, and instruments used for data generation, projection, collection, validation, and curation, along with other data features like purpose and applications. Details about space and time coverage and resolution are particularly essential for spatiotemporal datasets that are employed in water resource research (Ge et al., 2019). Rich, highly explanatory, complete, and thoroughly documented metadata embedded in a machine-readable format substantially simplify data interpretations for users and help save time and data analysis resources.
F3: In both NC and NLIS, metadata do not include identifiers.
As described in the evaluation of the F1 sub-principle, the NC and NLIS data repositories do not assign unique and persistent identifiers to each dataset; consequently, the metadata do not include such identifiers. The metadata should explicitly contain a data identifier to improve the searchability, indexing, and discoverability of the data through search engines and to enable the data discovery from the metadata directly.
F4: The NC and NLIS data repositories are indexed in a searchable resource.
As confirmed by our discovery of these data repositories during the web-based search process outlined in the methodology, both NC and NLIS are indexed in the Google Search engine’s database (see Table 2).
Table 2
Summary of FAIR-F evaluation; italic text displays FAIR non-compliance.
| DATA REPOSITORY | FINDABILITY |
|---|---|
| NC | F1. No unique and persistent identifiers have been explicitly assigned for each dataset. F2. Rich metadata, no ideal machine-readable metadata format has been provided. F3. Metadata does not include the identifier. F4. Indexed in the Google search engine’s database. |
| NLIS | F1. No unique and persistent identifiers have been explicitly assigned. F2. Rich metadata, no consistent machine-readable metadata format has been provided. F3. Metadata does not include the identifier. F4. Indexed in the Google search engine’s database. |
Accessibility
A1: The datasets in both the NC and NLIS data repositories are accessible through Hypertext Transfer Protocols (HTTPs).
HTTP(s) is a standard, open, free, and universally implementable protocol (A1.1) that also supports authentication/authorization when necessary (A1.2), depending on the access policies established by the data developers.
In our most recent access to the NC and NLIS websites, neither site required authentication/authorization, ensuring that the data are openly accessible to all users. However, according to the FAIR-A1.2 sub-principle, FAIR data do not necessarily refer to open-access data. Instead, data publishers can specify the conditions under which data are accessible based on user rights; in other words, highly protected data can nonetheless meet FAIR standards (Go-fair.org).
A2: In both NC and NLIS, there is no public evidence indicating the metadata preservation.
According to the FAIR-A2 sub-principle, metadata should be permanently accessible even when the data are no longer available. This requirement implies that the metadata should be persistent, accessible, and stored separately from the data (Go-fair.org). If the primary data are missing, lost, or inactive, a separate and persistent digital metadata resource allows users to understand that a specific dataset with a set of features was previously available for research purposes. Providing metadata in appropriate formats (F2), with persistent identifiers (F3) registered in searchable resources (F4), is essential for meeting the FAIR requirement of preserved metadata accessibility (see Table 3).
Table 3
Summary of FAIR-A evaluation; italic text displays FAIR non-compliance.
| DATA REPOSITORY | ACCESSIBILITY |
|---|---|
| NC | A1. Retrievable in an open, free access protocol, and no authentication/authorization is currently required. A2. No guarantee for metadata preservation. |
| NLIS | A1. Retrievable in an open, free access protocol, and no authentication/authorization is currently required. A2. No guarantee for metadata preservation. |
Interoperability
I1: Both NC and NLIS partially use formal, accessible, shared, and broadly applicable language for knowledge representation.
Although both NC and NLIS use standard and widely accepted scientific languages in water science to describe the (meta)data, the NC metadata are released in ‘.pdf’ and ‘.xlsx’ formats, and the data are published in the ‘.xlsx’ format following a consistent (meta)data structure. In the NLIS, (meta)data are provided in plain text-based descriptors across all selected dashboards, where data are downloadable in the ‘.csv’ format in one single dashboard (i.e., ‘Precipitation Projections and Winter Trends on the US Native Lands’), following an inconsistent (meta)data structure.
To enhance interoperability, data should be released in a machine-friendly format, including but not limited to JSON-DL, XML, or RDF, as previously discussed in F2. For geospatial data, formats such as GeoJSON, Shapefile, and NetCDF are also recommended because of their effectiveness in map generation, geospatial analysis platforms, terrestrial modeling tools, and geospatial data archives (Butler et al., 2016; Hassell et al., 2017; Jhummarwala, Prajapati and Potdar, 2017). Additionally, the metadata can be structured using standards such as the Dublin Core Schema to ensure consistency and interoperability.
I2: Both NC and NLIS partially use FAIR vocabularies.
The absence of adequate documentation, clear definitions, and direct links for some variables, tribes, and references within NC and NLIS metadata and datasets hinders an efficient understanding, interpretations, and use of the data. Using formal FAIR languages in (meta)data facilitates the effective discovery, interpretation, validation, and processing information by both humans and machines. This sub-principle ensures that each concept, term, and vocabulary used in the (meta)data is effectively documented, findable through unique and persistent identifiers (F1), accessible by standard communication protocols (A1), and interoperable by employing a widely applicable language for knowledge communication (I1) (Go-fair.org; Wilkinson et al., 2018).
I3: Both the NC and NLIS provide links to the relevant digital resources.
In both data repositories, resolvable links to the original data products on which they are built, reference articles describing the data development methodology, and relevant digital resources, portals, and data partners are properly cited and acknowledged.
Enriching the (meta)data by referencing the resources and articles through which the data is developed is essential for understanding the data development process, understanding data origins, and ensuring data validity (dos Santos, Peroni and Mucheroni, 2022) (see Table 4).
Table 4
Summary of FAIR-I evaluation; italic text displays FAIR non-compliance.
| DATA REPOSITORY | INTEROPERABILITY |
|---|---|
| NC | I1. (Meta)data are provided in a partially formal, accessible, shared, and broadly applicable language formats. I2. (Meta)data partially use FAIR vocabularies. I3. The metadata provides the key references. |
| NLIS | I1. (Meta)data are provided in a partially formal, accessible, shared, and broadly applicable language formats. I2. (Meta)data partially use FAIR vocabularies. I3. The metadata provides the key references. |
Reusability
R1: In both NC and NLIS, metadata do not contain the data usage license, provide details on data provenance, and partially align with community standards.
NC provides important data descriptions, including data provenance, developers/sponsors, development methods, spatiotemporal coverage, variable definitions and units, and contact information for any inquiries. Similarly, the NLIS offers a necessary data description, including data provenance, developers/sponsors, development methods, spatiotemporal coverage, data importance and limitations, and a contact email address for any inquiries.
However, both NC and NLIS have considerable potential to improve (meta)data reusability for research applications. These improvements include specifying a clear data usage license for which the data can be used (R1.1), providing data provenance in both human- and machine-readable formats (R1.2), and ensuring that the (meta)data aligns with the scientific community standards (R1.3). For the latter task, any specific standards established by Indigenous communities also should be considered. A consistent (meta)data structure in format, file naming, variable definitions, and data applications and limitations is also highly useful in enhancing data reusability (see Table 5).
Table 5
Summary of FAIR-R evaluation; italic text displays FAIR non-compliance.
| DATA REPOSITORY | REUSABILITY |
|---|---|
| NC | R1. No clear data usage license is provided, details on data provenance is provided, and partially follows the community standards. |
| NLIS | R1. No clear data usage license is provided, details on data provenance is provided, and partially follows the community standards. |
Although both the NC and NLIS repositories offer valuable Indigenous WCE datasets, a specific set of recommendations has been developed as a part of FAIRification to address the identified gaps and non-compliances with FAIR principles. These recommendations provide a first-level set of actions for data managers/developers to implement:
NC
Assigning unique and persistent identifiers for each dataset and including them in the metadata.
Ensuring that the preserved metadata remain as comprehensive as possible, including but not limited to data accuracy, data limitation, spatiotemporal coverage updates, resolution, and identifiers in both human- and machine-readable formats.
Providing the datasets in appropriate human- and machine-readable formats.
Registering the (meta)data in various searchable resources for broader discoverability.
Specifying a clear data usage license.
Maintaining (meta)data with both WCE data and Indigenous community standards.
NLIS
Assigning unique and persistent identifiers for each dataset and including them in the metadata.
Ensuring that the preserved metadata remain as comprehensive as possible, including but not limited to data accuracy, variable definitions and units, spatiotemporal coverage updates, resolution, and identifiers in a dedicated persistent metadata digital resource in both human- and machine-readable formats in all dashboards.
Providing the datasets in consistent, both human- and machine-readable formats in all dashboards.
Registering the (meta)data in various searchable resources for broader discoverability.
Specifying a clear data usage license.
Maintaining (meta)data with both WCE data and Indigenous community standards, with a consistent (meta)data structure.
Discussion
The FAIRification of Indigenous data repositories is the first step that Carroll et al. (2021) recommended for operationalizing FAIR-CARE principles in the entire data lifecycle for already existing and newly created data in both exclusively Indigenous and a mix of Indigenous and non-Indigenous datasets. Thus, the development of a data management plan should be encouraged through the proposal phase of Indigenous database development projects. In turn, funding agencies should consider and provide the financial support to guarantee the implementation of FAIR principles. Funding sources typically require a data management plan in research proposals but may not weigh evaluation scoring rubrics to recognize the FAIRification of the data products. Additionally, proposal budget justifications can include resources to support professional development opportunities and training for projects and community members to manage data. If the management plans implement these principles in the WCE data, evaluators should recognize this practice in the evaluation scoring rubric. The implementation of sound data management plans will contribute to the overall research project quality. This implementation plan should be considered as an ongoing effort in each stage of the data lifecycle, including design, infrastructure, collection, storage, and reuse for sustained FAIR compliance (Jennings et al., 2025). Both NC and NLIS are acknowledged for providing a wide range of WCE datasets across reservations in the US, and after implementing the FAIRification processes recommended above, subsequent datasets will provide a step forward in enhancing visibility and reusability for data-driven water resource research on reservations.
Notably, FAIR data principles are not designed to examine data quality or reliability; instead, the evaluation of these aspects is the responsibility of the data user and is based on a specific data application (Go-fair.org). In this research, we aimed to highlight the importance of implementing FAIRification for Indigenous WCE data publishers and users through emphasizing the need to enhance data discoverability and usability; such an approach would facilitate the broader use of data across a wide range of data-driven topics in water resource research, including, but not limited to, surface and groundwater modeling and management, water balance analysis, flood control and risk assessment, water quality modeling, supply and demand management and operation, ecological modeling and assessment, and quantifying the effects of climate change on water resources.
Despite there being a remarkable potential for technical studies and data-driven research to address the water-related problems of Native communities in several cases, data scarcity poses barriers for scholars, water managers, policymakers, and stakeholders (Young, Colby and Thompson, 2019). Only a limited number of data-driven case studies have been conducted; those that do exist highlight the critical role of valid data generation and application in research, with a focus on enhancing the well-being of Indigenous peoples. Examples include analyzing the geospatial distribution of groundwater contaminants across the Navajo Nation (Hoover et al., 2017), health risk assessment and water treatment performance evaluation in the Crow Reservation (Eggers et al., 2018), surface water quality sampling across the Yukon River Basin conducted in collaboration with the Alaska Native Tribes and Canadian First Nations (Wilson et al., 2018), groundwater quality assessments in the Omaha and Santee Sioux Reservations (McGinnis and Davis, 2001), and groundwater sampling in Alaska (Rowles III et al., 2020). These data-driven studies can contribute to informed decision-making regarding water and natural resource management in tribal communities.
The ‘availability’ of data falls out of the FAIR data principles’ scope, and in the case of Indigenous data, decisions regarding data publication and accessibility are in the scope of the sovereign right of the tribes, who determine whether and how their data should be shared. The core of CARE data principles is founded on data protection practices to preserve Indigenous data and ensure that data usage aligns with the interests of Native communities. However, publishing, sharing, and preserving non-sensitive Indigenous WCE data that follows FAIR principles in research data repositories, in tandem with tribal agreements, can remarkably advance data-driven water resource research focused on Native communities.
This study focused on Indigenous-led WCE data repositories. Data repositories managed by federal and state organizations, which record, document, and publish Indigenous data as part of larger databases, were not included in this analysis.
Conclusions
The lack of FAIR WCE data is a notable barrier in conducting data-driven water resource research relevant to tribal communities. The FAIRification of Indigenous data is an initial step toward operationalizing the integrated FAIR-CARE principles that support the development, sharing, and protection of these data. Our assessment of publicly available Indigenous WCE databases such as NC and NLIS outlines key areas for improving the data’s findability, accessibility, interoperability, and reusability. The NC and NLIS data repositories provide valuable spatiotemporal hydrological, climatic, and environmental datasets across US reservations in various data structures and formats. However, these data repositories do not fully align with the FAIR guiding principles, as outlined below:
The absence of unique and persistent identifiers hinders data discovery and tracing.
Inconsistent and nonmachine-readable (meta)data formats make working with these datasets challenging and time-consuming.
Metadata could be considerably more explanatory regarding data accuracy and limitations, variable definitions, updated information on spatiotemporal and resolution coverage, and data usage terms and conditions.
Registering datasets published by the NC and NLIS on multiple data-search platforms is strongly recommended to increase their visibility to researchers.
Specifying a clear data usage license and permissions can substantially enhance research efficiency, facilitate data reuse, promote transparency, and protect the rights of both data developers and users.
The FAIR Indigenous WCE data can considerably enhance data discoverability and usability for both Indigenous and non-Indigenous scholars, thereby broadening the scope and impact of data-driven water resource research relevant to tribal communities. Because this research was limited to the web-based, Indigenous-led WCE data repositories, an evaluation of FAIR principles in Indigenous WCE datasets documented and released by federal and state institutions, as well as non-web-based data, is recommended for future work. Additionally, FAIRness requirements may be different in different communities due to the community-specific standards. Therefore, evaluating the FAIR data metrics in the context of the Indigenous community could be a valuable topic for further research.
Additional Files
The additional files for this article can be found as follows:
Acknowledgements
We thank the Native Climate and Native Land Information System for providing the preliminary data used in this study. We also thank Cathy Middlecamp for her insightful comments on improving our manuscript. AI has been used in some cases for grammar checks.
Competing Interests
The authors have no competing interests to declare.
Author Contributions
P.S. conceptualized the study, carried out the implementation, and prepared the manuscript, and G.B. conceptualized the study, supervised the research, and reviewed and prepared the manuscript.
Author Information
P.S. is a Research Associate at the Nelson Institute for Environmental Studies at the University of Wisconsin-Madison. Her main research focuses on water resource management, database management, and data science.
G.B. is a member of the Crow Tribe and an Assistant Professor at the Nelson Institute for Environmental Studies at the University of Wisconsin-Madison. Her main research focuses on developing resource sovereignty frameworks for Indigenous communities.