Challenges of Documenting Historic Water Systems

ABSTRACT Geospatial research in archaeology often relies on datasets previously collected by other archaeologists or third-party groups, such as state or federal government entities. This article discusses our work with geospatial datasets for identifying, documenting, and evaluating prehistoric and historic water features in the western United States. As part of a project on water heritage and long-term views on water management, our research has involved aggregating spatial data from an array of open access and semi-open access sources. Here, we consider the challenges of working with such datasets, including outdated or disorganized information, and fragmentary data. Based on our experiences, we recommend best practices: (1) locating relevant data and creating a data organization method for working with spatial data, (2) addressing data integrity, (3) integrating datasets in systematic ways across research cohorts, and (4) improving data accessibility.

During research for the Water Heritage Anthropological Project (WHAP), we have encountered several obstacles working with historic water infrastructure, and we have developed several solutions to help mitigate obsolete datasets. In this article, we discuss archaeological geospatial datasets (i.e., datasets that reference a specific geographical area or location) in northern Utah, a part of the Great Basin that is rapidly developing and modifying water features such as canals, which are being piped and modified for more efficient water use at an alarming rate (Walton 2021). As part of the multidisciplinary WHAP, we use geospatial data to document historic irrigation systems to provide insight into water heritage (that is, the emotional and physical relationships between water, culture, and community, as described in Hein 2020) and to assess impacts under the National Historic Preservation Act. Geospatial datasets, along with archaeological fieldwork and archival research, are essential to the development of a historical context that presents a detailed, longitudinal view of water use in Utah, as well as to reinforce the importance of water heritage to traditional communities. Historic contexts, as well as multiple property and historic district nominations to the NRHP, are crucial to our understanding of the past and create a framework for interpreting and defining the significance of historic and archaeological sites (Hardesty and Little 2009;Lees and Noble 1990;Morgan 2022). The first step to achieving this goal is the collection of available data and information on water infrastructure from precontact, ethnohistoric, and modern periods. Our experience demonstrates that water feature documentation is often fragmentary, which makes it difficult to evaluate features within a holistic context.
In this article, we first provide an overview of geospatial datasets applicable to historical water management in Utah and highlight some of our project research goals. We then discuss the methods we used for dataset location, collection, storage, and interpretation. We show that many of these datasets are only accessible in gray literature or partially accessible databases online, which makes it difficult to identify and aggregate data in one location. Next, we look at the frequency of piped or open canals in Cache County, northern Utah, and the number of water features that are considered for National Register status as case studies for a broader discussion of integrating water datasets. We conclude by highlighting challenges that arise when working with geospatial datasets, and we recommend four best practices for working with such data to expand documentation of historic water systems. These include (1) locating relevant data and creating a data organization method for working with spatial data, (2) addressing data integrity, (3) integrating datasets in systematic ways across research cohorts, and (4) improving data accessibility. Given the widespread use of geospatial data in archaeology, and the increasing need to aggregate legacy and new datasets, we hope that our experiences will benefit future projects in academic and professional fields that are documenting complex water systems and their associated historic and archaeological features.

DOCUMENTING HISTORIC WATER SYSTEMS
Documenting historic water systems is part of the WHAP project, which investigates the relationship between communities in Utah and precontact and historic water features. Our approach to water heritage incorporates a suite of tools, methods, and theories from archaeology, anthropology, and geography, and it includes geospatial, archival, and ethnographic (interview and oral history) data. By researching and documenting the history of water management and the use of irrigation, water management infrastructure, and agricultural practices, the WHAP applies a long-term approach to studying water in the western United States and aims to inform future water policy. A long-term multidisciplinary approach to studying waterways offers further understanding of past events and leverages our ability to draw insights and possible solutions for contemporary issues such as aridification and drought in the West (Kaptijn 2018;Williams et al. 2022). The project also creates materials to educate the public about local water histories and provides mentorship opportunities for students to connect with water heritage studies (Cannon et al. 2021). Our current focus is on the Bear River Watershed and adjacent communities in Cache County, northern Utah (see Figure 1), but we are expanding our research to other parts of the state.
Historic water systems are complex social, environmental, and engineered landscapes, consisting of a series of built environment features facilitating water storage (dams and reservoirs), water extraction (wells), and water conveyance (canals, ditches, aqueducts; Hardesty and Little 2009:86;Rynne 2022;Worster 1992). Many of these systems have associated water management components such as levees, hydroelectric generators, headgates, or weirs (McVarish 2008:Chapter 5). These systems manipulate the natural water system across large expanses, and they impact vast watersheds. The nature of cultural resource management projects, limited to evaluations within the Area of Potential Effect, often fail to fully document historic water systems, resulting in fragmented datasets and incomplete documentation. Many of the historic water systems of the West remain in service, complicating documentation of their historic significance further. As water managers strive to meet water demands while experiencing record droughts, modern piping and other engineered improvements to historic systems are often viewed as impacts to the integrity of the historic resource (Walton 2021).
The geospatial portion of our research has focused on aggregating Utah water feature data into a geographic information system (GIS) to document historic irrigation systems, including adaptive technology such as canal piping. Using these spatial data, our key questions about water heritage include the following: (1) What percentage of water features in Utah are considered historically significant, per the NRHP guidelines? (2) How many historic water features in Utah have been evaluated and documented in their entirety? (3) Who owns the water rights and controls the usage of historic water systems?
Documenting adaptive technologies that have been employed through time and across Utah shows the evolution of water management in the state and contributes to our understanding of the historical significance of these systems in the region. It also identifies potential water management obstacles for the future. Through our work with water infrastructure geospatial datasets, we have identified several challenges, including the difficulty of locating and aggregating data and making datasets accessible. Consequently, developing a detailed, consistent methodology is critical for addressing such obstacles. As we discuss below, the WHAP has implemented an organizational system to help reduce   Code 1953); however, the right to use waterways for farming, irrigation, or culinary practices (i.e., drinking water) requires a legal claim or water right. A water right refers to the permission granted to an individual or entity by the Utah DWRi to divert water from a natural source (i.e., groundwater, natural spring) for use in agriculture, industry, and municipalities. Today, Utah water management companies are privately operated, and they finance the maintenance and construction of new infrastructure with payments from shareholders (farmers, landowners, or other entities that are granted a certain amount of water use by the water-right owner) and government-sponsored grants (Smoak 2020;Utah Code 1953). The Utah DWRi provided the WHAP with important water-rights histories of canals, canal companies, and modifications to water rights dating back to the nineteenth century. Attribute information from the DWRi filled critical gaps for creating a timeline of historic irrigation practices that can demonstrate a feature's historical significance by identifying an association with a time, place, or person that may not be possible via archaeological methods. The Utah Division of Natural Resources also has the Open Water Data server, from which spatial datasets can be downloaded and used to show average precipitation, water levels, and drought records for Utah (for an overview of open datasets on water, see https://dwre-utahdnr. opendata.arcgis.com/pages/more-water-data).
Archaeological records for historic-canal archaeological sites are not publicly available, but some of these spatial data were provided by a request to the Utah State Historic Preservation Office (SHPO). This request stipulated that the project principal investigator(s) first apply for and receive Utah Public Land Policy Coordinating Office (UPLPCO) permits and then request the geospatial data from the SHPO records. Archaeological site forms and survey reports for historic canals in Utah came from the Department of Cultural Community and Engagement digital database and, like the request for spatial data, required a permit from the UPLPCO for access. The archaeological site forms and survey reports document field observations for evaluating historic water systems and irrigation features.
A key aspect of the project is to compile information on the NRHP eligibility statuses of water features in order to determine where future documentation and evaluations are needed and to ultimately advise the SHPO on how to proceed with NRHP nominations.
Our geospatial efforts integrate the canal geospatial dataset from the UGRC, historic water rights attribute information from the Utah DWRi, and archaeological information from the Utah SHPO (see Table 1 for a list of databases and accessibility). For example, the Wellsville Canal, located in southern Cache Valley, was originally constructed in 1850 as a 8 km (5-mile) long gravity-fed irrigation canal that diverted water from the Little Bear River (Figures 2 and  3). Information from the Utah DWRi provides the history of management of this system, including determination of water rights, how water was diverted, and for what purposes (Figure 3). Archaeological and historic record information, such as codes used to identify feature attributes (e.g., in service), were also included, as well as any information about the date of construction. Using these datasets, we also created a master geodatabase and smaller databases for addressing research questions about historic water systems in Utah. For example, we are interested in how water systems differ across Utah; consequently, we used the historic canal datasets from the Utah SHPO and the history of water rights from the DWRi to generate comparative analyses between counties with high-density historic water infrastructure investments, such as Cache County (or the Bear River watershed), versus those with sparse investments, such as San Juan County (the southeast Colorado watershed) in southern Utah. This smaller database quickly enables us to visualize the vast differences between the two regions and to collect additional information that addresses the variation, including land ownership and demographic data.
Another important component of the WHAP is to prioritize data management within research, teaching, and collaborative contexts (see discussions in Cohen et al. 2020;Garstki 2022;Kansa and Kansa 2021;Smith 2021). Like other large research projects, the WHAP employs a rotating group of student researchers and collaborators, which necessitates a data management plan that ensures consistency in data generation across research cohorts. Metadata (i.e., a set of data that provides and describes information about other data; Comber et al. 2008) are maintained within a common separated value (CSV) file and then embedded into geospatial data files (also called "shapefiles"). This CSV file contains relevant information: a list of spatial datasets gathered, where we obtained the data, who has manipulated or contributed to the attributes of a dataset, why those data were manipulated, how or what was added to those data, disregarded data irrelevant to our project, as well as where and how those data were used during the project (e.g., maps, ESRI StoryMaps, conferences; Online) and ultimately curated within the USU Digital Commons (https://digitalcommons.usu.edu/water_heritage/). Although the project currently focuses on Cache County in northern Utah, as we add spatial data from other regions of Utah, we will continue to update and maintain our data files for future use.

Historic Canals in Cache County
The history of Utah's water management is complex, with periods of communal management and complicated private-government initiatives (Smoak 2020;Toelken 1991). Early Latter-Day Saints  Arrington and May 1975;Harvey 1989;Houston 2002;Thomas 1920). During the last quarter of the nineteenth century, water development projects occurred mainly as private ventures commodifying water. Water development in the twentieth century evolved into a joint venture between local, state, and federal government agencies (Smoak 2020). Cache County embodies this history through its archaeological and historic water infrastructure, and today there are a few large companies that manage water in the county (e.g., Bear River Canal Company and Cache Highline Water Association) and many small, locally, and family run water management facilities (approximately 50 entities in the county; Cache Water District 2015).
Gathering information on the eligibility status of historic water features helps determine if additional fieldwork or research is needed to document more water features in Cache County, as well as direct preservation practices such as creating a statewide historic context or a multiple property NRHP nomination. According to our growing database in Cache County, approximately 31% of canals (see Figure 2) have been evaluated and recommended as eligible, many of which are canal segments and not the full length of the feature. This is likely because a canal was evaluated within a particular cultural resource management project's Area of Potential Effect rather than in the canal's entirety. As of 2021, only one water feature, Old Newton Dam (see Figure 3), is listed on the NRHP. Most canals in Cache County are unevaluated (∼63%), but from the sample of evaluated canals, very few have been recommended as ineligible (∼6%). Further assessment of an entire canal and other water features, especially for canals assessed in fragments, will aid in the documentation and preservation of historic knowledge that these features offer to researchers and the public.
Piping is a common upgrade to in-service historic canals in the Mountain West because it conserves water by reducing seepage into the arid soils and reduces overall maintenance costs in the long term Shupe 1982  status for listing in the NRHP. Compiling this information will aid archaeologists in determining which historic canals need evaluation, documentation, and revisitation. It is important to document historic features, both in use and abandoned, to ensure a detailed and organized framework for historic context (Hardesty and Little 2009;Morgan 2022). Furthermore, as water becomes increasingly scarce in Utah, there is a need to prioritize water conservation, which could be in the form of piping or other infrastructure changes to canal systems (Anderson 2013; Armstrong and Jackson-Smith 2017; Walton 2021). Given that much of the canal infrastructure has both historic architectural and cultural significance to state heritage, archaeologists must prioritize the documentation of the infrastructure before it is modified or destroyed.

INTEGRATING WATER MANAGEMENT AND ARCHAEOLOGICAL DATA FOR HISTORIC WATER SYSTEMS: CHALLENGES AND SOLUTIONS
Creating a historic context for canal infrastructure in Utah requires data from various sources-such as the Utah SHPO, the UGRC, and the Utah Division of Natural Resources-to create a comprehensive picture of water infrastructure changes in Cache County. Integrating datasets from multiple sources into a GIS can be complicated because individual datasets often have unique sets of codes and identifiers for attributes with limited or no associated metadata. As discussed above, water management datasets may not contain the most current information regarding water infrastructure, such as recent piping and dam expansions. Below, we highlight some of challenges and proposed solutions when integrating water management datasets with archaeological data, including how to locate appropriate data with integrity, how to integrate the data, and how to make them accessible.

Locating Relevant Water Management Data
There are a limited number of publicly available datasets that contain recently updated information about canals and water rights. However, finding such data can be a laborious process. One initial step might include requesting information from the individuals or agencies that originally collected the spatial data regarding the codes, attributes, or other coding decisions they made when creating the data. This may be an essential step because these data were invariably collected in different ways, resulting in a lack of standardization and frequent use of discipline-specific jargon, acronyms, and coded information and making analysis difficult without first translating the data into usable language for archaeologists. During the initial stages of the WHAP, this supplemental research involved contacting the entities that collected the data and requesting a codex, guide, or DESC Description Any additional historical information; note the purpose of the geospatial data (i.e., for use in the historic waterways project at USU).
X, Y XY Locations. This may be modified to protect sensitive locations. Any modifications will be noted in the master CSV file (separate fields).

MOD Modified locations.
YES or NO: if YES, an explanation of modification will follow, e.g., county centroid. This will inform future users that the location of the original data has been modified to protect the location of protected sites or features.
Note: This codex was used to organize data created or obtained from outside sources with the WHAP geodatabase. Maintaining this information allows for quick knowledge transfer between multiple collaborators, and it prevents the loss of vital information about the origin of data and the manipulation of datasets.
other document describing the data. Sometimes this was feasible, such as with SHPO-restricted data; but in other cases, contacting large organizations with open data servers, such as the Utah SGID, was an exercise in futility. The COVID-19 pandemic also exacerbated our difficulties with communication because canal companies and federal and state agencies were difficult to reach.
One of our solutions was to locate similar research and to try to define codes within the context of like datasets. For example, we located informational web pages from the Utah Division of Water Resources (Utah Division of Water Rights 2018) that explained some of the codes, such as "MaxCFS," a common code used in water research to quantify a canal's flow capacity, maximum cubic feet per second. Deciphering acronyms and jargon are not new issues, and spatial datasets with undocumented codices or data dictionaries are common in open data sources (Barry and Bannister 2014;Comber et al. 2008). The expansive nature of water studies exacerbates this issue, where there are a variety of publishers from different disciplines contributing to these servers, each with different geospatial goals, disciplinary language, and access provided to datasets. Solutions and topical focus in water-related geospatial datasets are inconsistent and tend to shift from the provision of metadata to the ease of access and manipulation (Corsar and Edwards 2017).

Addressing Data Integrity
Another major challenge is that available public datasets are often obsolete. In the case of Cache County (as of Summer 2020), most spatial information on canal status is from 2013, and since that time, there have been numerous canal piping projects in the region (e.g., Schill 2021;Webb 2021). This means that we had to update the obsolete data ourselves, and that required additional time to call canal companies or visit some of the canal segments. The lack of public access to the current status of water features creates challenges for researchers establishing historic contexts or submitting NRHP nominations who do not work for the data managing entity. Additionally, much of the data available on open data servers are raw or in an original data collection format, meaning that they are neither fully processed nor organized. Data access may be quicker, but it is not as informative (Wang et. al. 2018).
One way that we addressed data integrity is by combining incomplete or obsolete data with compatible supplemental data that were gathered by archaeological or historic preservation projects. The WHAP has been conducting reviews of past NRHP evaluations and surveys of canal infrastructure in northern Utah and incorporating information from those evaluations into our master historic canal dataset, which shows the current canal infrastructure. These data include the eligibility status of historic water features and modifications to canals after 2013 (i.e., piping or lining). Updating the current status of a historic feature is important for determining documentation priority for undocumented historic features that are marked for upcoming modifications or upgrades.

Integrating Data
Datasets collected from multiple sources required manipulation and updates. For example, we gathered public data containing information on water infrastructure, but these data often do not contain information about historic status or current standing. Another challenge is that the spatial data utilized for this project are repetitive and require manipulation, including closing polygons and smoothing or appending linear features. Within the Utah canals dataset, we discovered that many canals were documented in segments. Some of these segments are named using codes (segment A, segment B, etc.) that may represent different historic periods of construction or use, whereas others were assigned with the name of the canal. This becomes an issue when creating a more detailed visual representation of the canals (i.e., with individual canals labeled on maps for presentations or other outreach products), where preservation of some but not all of the coded attribute information is necessary and requires more work designing, rewriting, and reimagining legends and labels-and sometimes recreating the feature datasets to be continuous polylines (see Figures 4 and 5 for examples).
Retaining and creating metadata and a geodatabase are critical for preserving historic contextual information from legacy datasets and for documenting geospatial and attribute manipulations (Comber et al. 2008). In addition to basic metadata fields regarding the source, age, and ownership of data, the metadata catalog should detail who has worked with or manipulated a dataset, and how that data may have changed over time. This is especially useful within long-term projects that work with multiple undergraduate and graduate students throughout the duration of the research. Given that students and other project collaborators may change regularly, the metadata catalog provides both a manual for future additions and a record of past modifications. Storing the metadata within the geodatabase will ensure that spatial data is organized, secure, and easily accessed by multiple users. The metadata also includes deciphered codes from original datasets and their meanings, as well as the codes created by the WHAP for supplemental data added from our ongoing research. Finally, we recommend working with a file geodatabase that stores and queries spatial data because the structure allows for a smooth transfer, upload, and dissemination of spatial data through multiple sources. A geodatabase will benefit not only data organization but also functionality and shareability; the size of a file geodatabase is limited only to the hard drive's available disk space and can be compressed for easy transfer between users.

Making Data Available and Accessible
A final challenge relates to how to make water heritage datasets more accessible to other researchers and the public. Visualizing cultural heritage via accessible, printed maps is an important part because it allows us to illustrate a connection between artifacts, other material remains, and the landscape (Warner-Smith 2020). Although spatial data can be collected, stored, processed, and displayed using GIS platforms such as ESRI products, QGIS, or GRASS (Hacigüzeller 2017), more thoughtfully drafted accessible maps are an important tool for engaging communities and for providing a straightforward presentation for the public. The vast number of canals in northern Utah alone makes it difficult to create comprehensive maps for the state, so we have produced only general regional maps thus far. Labeling and classifying large amounts of spatial data within one map presents several visibility issues. Labeling too many canals within the GIS at an inappropriate scale means that more spatially extensive canals are labeled, whereas smaller, less extensive canals are not automatically labeled. Fragmented data creates a related set of concerns, causing excessive labeling that is difficult to mitigate without the time-consuming task of creating new feature datasets. An automated legend is also problematic because there are too many segments to create a comprehensible inscription.
Mapping the Cache County canal systems presents examples of these challenges. Our efforts to visualize a comprehensive map of all Cache County canals encountered aesthetic drawbacks, including overcrowding and low resolution. We remedied this by producing a series of regional maps of Cache Valley and surrounding areas focusing on fewer canals at a time, as well as using additional graphic design software-such as Adobe Illustrator-to create legends and labels outside of the ArcGIS software (see Figures 4 and 5).
An approach that helped us to visualize a crowded region was to compare Cache County with a sparsely populated county elsewhere in the state. For example, San Juan County in southern Utah has few active canals and was easily mapped using labeled canals as well as magnified data frames for close-ups of the canals. Visualizing this contrast led us to further divide Cache into subregions, which helped to highlight the extensive canal infrastructure. The legend lists the names of the canals with a corresponding number, and that number was placed within the map surface to indicate the appointed canal feature so as to eliminate cluttering of canal names across the map image ( Figure 5). Although this required multiple steps, we succeeded in creating a comprehensible map for the canals in Cache County.
To facilitate greater public access, we created a series of interactive visual representations of spatial data via ESRI StoryMaps called "Water Heritage in the West" (see https://arcg.is/iHzS00). A StoryMap is a web-based application from ESRI that has been effectively used to present spatial research by incorporating interactive components such as maps, text narratives, photos, and videos (ESRI 2013;Howland et al. 2020 (Baculi et al. 2021), and the "Road to Promontory: Archaeology along the Central Pacific Railroad in Utah" (Cannon and Keith 2020) have successfully used spatial data to expand awareness about local history and environmental change. Our goal for the StoryMaps is to communicate and share our research with the public, most recently with the Native American Summer Mentorship Program and the Summer Citizens programs at Utah State University (Cannon et al. 2021).
The "Water Heritage in the West" StoryMap collection was created by WHAP student researchers, and it includes information such as the definitions of water features and archival photos. The collection incorporates information from all dimensions of the WHAP-including archival, ethnographic, and spatial datasetsand presents the information in a visually appealing and concise way. Ideally, these presentations can be used to educate the public about water heritage in Utah.
In archaeological research, the inclusion of community stakeholders supports, enhances, and ensures a positive atmosphere and maintains a sense of identity and ownership for community members (e.g., Atalay 2006;Brighton 2011;Kuwanwisiwma 2008). A primary goal of collaborating with community members for research is to correct the power imbalance between the researcher and the researched by ensuring that community members have a say in how their own heritage is documented and disseminated (Atalay 2006;La Salle 2010). Because heritage is often understood as tangible features of a physical landscape, but also intangible aspects, both should be considered when trying to understand and protect cultural heritage. Addressing heritage concerns of a community ensures that archaeologists maintain integrity and serve the needs of the community; active collaboration with community members "presents a more culturally sensitive, community responsive, ethically aware, and socially just" archaeology (Nicholas et al. 2011).
The WHAP has collaborated with both private and public entities to gather spatial data and information on Utah's waterways and water heritage, including the Utah SHPO, other USU departments, and private canal companies, such as the Bear River Canal Company. Through these partnerships, we have fostered longterm connections with groups and entities that are continuing their own research and development on the canals in northern Utah, and as a collective, we are able to share data and pursue holistic research. Collaboration with the community in Cache Valley acknowledges ownership of heritage and allows community members to have a say in how their histories are portrayed both publicly and academically. The WHAP works alongside community-facing institutions such as the USU Museum of Anthropology to develop outreach programs. In 2021, the WHAP developed content for the Exploring Water Heritage AnthroPak, a backpack filled with engaging cultural education and craft supplies, which was distributed to families during local STEM fairs. Going forward, the WHAP hopes to engage and collaborate more with canal companies throughout Utah; provide opportunities to foster community voices through the collection of interviews, oral histories, and public presentations on our research; and to work with Native communities in Utah. Collaborative research not only facilitates communication but also creates a sense of community with fellow researchers and the public.

CONCLUSION
The use of legacy and living datasets can enrich our understanding of past water use and management, which are particularly important to discussions about sustainability and consumption in the western US. The current state of some of these datasets, however, requires updating and manipulation for use in archaeological or anthropological research. The WHAP has used geospatial datasets from many online and publicly available sources, such as the UGRC and the Utah DWRi, but we also acquired semi-open-access data via special request from the Utah SHPO. These datasets provided basic information about the water management systems in Utah-mainly canals and other built systems-but they were found to be fragmentary and outdated. We used gray literature, via archaeological survey reports and site forms, to obtain more information about the historic status of existing canals from the Department of Cultural Community and Engagement digital database, which is not publicly available to those without specific credentials (Ellis 2008(Ellis , 2015(Ellis , 2017. We used these data to create several maps pertaining to the current status of piped versus open canals and displaying the eligibility of historic canals for listing on the NRHP in Cache County, Utah. Through the development of these maps, we have identified a few best practices for organizing and manipulating these types of geospatial datasets, as well as suggestions for improving methods for our research team and other researchers. A long-term spatial data management plan, updating outdated or fragmentary data with additional information through supplemental research, and collaborating with local communities will ultimately enrich similar archaeological research.