BACKGROUND

Beginning in the early 2000s, the Department of the Interior created a cooperative conservation project with the vision of managing landscapes “capable of sustaining natural and cultural resources for current and future generations” (https://lccnetwork.org/about/about-lccs). This vision led to the creation of 22 Landscape Conservation Cooperatives (LCCs) that aim to develop best management practices, promote collaborative conservation, and identify gaps within scientific knowledge to promote innovative conservation planning and design in North America. The central goal was to advance landscape-scale conservation for both natural and cultural resources (see LCC 2018; NASEM 2016). The primary mission for each cooperative is to “develop and provide integrated science-based information about the implications of climate change and other stressors for the sustainability of natural and cultural resources” (https://lccnetwork.org/about/about-lccs). The cooperatives were each tasked with defining shared landscape conservation objectives for their region that would be derived from cultural and natural resource conservation priorities. This objective is reflected in a 2016 review of the LCCs by the National Academy of Sciences, in which they state the third science goal of all cooperatives: “Natural and cultural resources are conserved at large landscape and seascape scales, guided by the collaborative application of science, experience, and cultural or traditional ecological knowledge and the generation of new conservation knowledge” (NASEM 2016:4). The cooperatives were designed to uniquely address something no other federal program had—that is, attempting to bridge research and management for both cultural and natural resources (NASEM 2016:4).

Many of the cooperatives advanced their mission through developing spatially situated, landscape-scale conservation designs (LCDs). While there are standard methods and approaches for landscape-scale conservation designs involving natural resources (e.g., Trombulak and Baldwin Reference Trombulak and Baldwin2010), no similar standards exist for integrating cultural resources into landscape designs. Moreover, in practice, cultural resources are more commonly overlaid after initial designs and analyses have been conducted in order to see where potential conflicts to the design may emerge. Building on the mandate of the cooperatives, our study argues that including cultural resources after initial phases of the planning process leads to an emphasis on the regulatory aspects of cultural resource management using existing databases rather than the generation of novel datasets, theoretical advancements, or new methodologies for cultural resources. This is a critical discrepancy given the fact that the cooperatives were tasked with integrating cultural resources as an integral component of conservation science—a task that is not accomplished through regulatory compliance alone.

The differences between practices for natural and cultural resource conservation planning are far-reaching. For example, in preparation for the Appalachian landscape design, natural resource stakeholders established shared habitat priorities and modeled indicator species through a series of iterative workshops. Similar efforts were not conducted for cultural resources. In addition, given the lack of available datasets for karst landscapes across the region, collaborative efforts were made to establish new datasets, while similar incomplete or missing datasets for cultural resources were not addressed. This gap was acknowledged in the 2016 report (NASEM 2016:21), which stated that although all the cooperatives had adequately focused on natural resources, their plans to address cultural resource conservation remained underdeveloped. If the cooperatives are the only federal program addressing both the natural and cultural aspects of landscape-scale conservation, the limitations of their efforts only serve to further highlight the gap between cultural and natural resource conservation science in North America as a whole. Our project is designed to fill those gaps.

Mandating the integration of cultural resources at the early stages of design and planning, the cooperatives initiated a new era in landscape design and planning. In practice, implementation of this mandate proved challenging, and subsequent efforts ended up heavily emphasizing natural resources. With the support of the National Park Service, the Appalachian Cooperative, and the Wildlife Management Institute, we designed a pilot study in January 2016 to develop processes for integrating cultural resource information in the initial stages of landscape conservation design, similar to the work accomplished for natural resources (see Baldwin et al. Reference Baldwin, Trombulak, Leonard, Noss, Hilty, Possingham, Scarlett and Anderson2018; Leonard et al. Reference Leonard, Baldwin and Daniel Hanks2017). Our pilot was initially limited to Pennsylvania in order to establish foundational data standards. We inventoried key cultural resource conservation themes, along with direct and indirect geospatial indicators of those themes (Brown and Murtha Reference Brown and Murtha2019; Goldberg et al. Reference Goldberg, Murtha, Orland, Gobster and Smardon2018; Mazurczyk et al. Reference Mazurczyk, Goldberg, Murtha, Gobster and Smardon2018; Murtha Reference Murtha2017). For example, direct indicators included statewide inventories of archaeological and historic sites, while indirect indicators linked land-use and land-cover data to document visual resources (Goldberg et al. Reference Goldberg, Murtha, Orland, Gobster and Smardon2018).

Broadly speaking, our long-term aims for this research are (1) to identify and describe conservation priorities related to cultural resources, (2) to evaluate data requirements and determine procedures for landscape-scale analysis, (3) to establish quantitative and qualitative measures for defining cultural resource conservation priorities, and (4) to investigate existing and potential land-use changes as well as coupled human and natural spatiotemporal dynamics for establishing cultural resource conservation principles.

For our first step, we assembled available data and developed a cultural resources conceptual framework that defines a series of procedural phases used to evaluate quantitative and qualitative aspects of cultural resources (Figures 1a and 1b). We developed this general framework following procedures similar to those used for natural resources (e.g., Trombulak and Baldwin Reference Trombulak and Baldwin2010) and allied decision support systems models, such as the Intrinsic Landscape Aesthetic Resource Information System developed by Jones & Jones (Jones Reference Jones2007). The Jones & Jones GIS-based decision support tool offers guidance for practitioners and landscape planners when evaluating the intrinsic value of natural resources of landscape regions for sustainable future design. Our theoretical framework applies several overlay analyses to explore spatial patterns of resources using both direct and indirect cultural resource data sources. The overall model is assembled around “tangible” and “intangible” resources (Figures 1a and 1b), which are broken into 11 discrete submodels. Tangible resources aim to inventory the built environment and physical landscape features (e.g., archaeological sites), while intangible resources aim to inventory aesthetic and visual qualities. The submodels are then inventoried and parameterized using available geospatial data. Archaeological data occupies a central role in our model development.

FIGURE 1. The conceptual model of cultural resource planning.

Early on in our study, we recognized the potential value of archaeological information and statewide archaeological databases as foundational resources for informing landscape conservation design. We also recognized a parallel need to compare statewide databases of these geospatial indicators because of the anticipated variability in availability and quality among states. Clearly, this data variability between states is not exclusive to archaeological information, but the specific preservation focus and secured nature of archaeological databases reinforces this important early comparative step if archaeological information is to be integrated in landscape conservation design. In reality, all of the cultural resource databases in our framework exhibited some variability across state boundaries. We concluded that archaeological data offered a unique opportunity to overcome that interstate variability because of emerging trends in practice (Alstchul Reference Altschul2016; Doelle et al. Reference Doelle, Barker, Cushman, Heilen, Herhahn and Rieth2016; McManamon et al. Reference McManamon, Doershuk, Lipe, McCulloch, Polglase, Schlanger, Sebastian and Sullivan2016) and deep traditions in landscape archaeology (Crumley and Marquardt Reference Crumley, Marquardt, Allen, Green and Zubrow1990).

To investigate the variability among state archaeological databases, we compare the databases of the Pennsylvania Cultural Resources GIS (PACRGIS) and the Florida Master Site File (FMSF). The Pennsylvania database was chosen for this study because it was already being utilized in the Appalachian cooperative work. Florida was chosen as a comparative measure because it provides a distinct context (archaeologically, historically, and in contemporary times) from the states surrounding Pennsylvania. For example, whereas Pennsylvania was a part of one of the largest multistate cooperatives, Peninsular Florida was its own cooperative. Moreover, each state has its own distinct planning and development pressures. Pennsylvania has no coastal resources, while Florida is dominated by them. Beginning this assessment with a comparison of Florida and Pennsylvania offers the opportunity to compare two states that have significantly different historical and cultural traditions but that operate within the same national cultural resource management framework. By highlighting these differences, pathways for coordinating across states, regardless of the similarities of their archaeological resources, can be established.

PENNSYLVANIA AND FLORIDA DATABASE COMPARISON

In this article, we focus on the challenges and opportunities encountered while specifically working with archaeological data from two statewide GIS databases. We describe and compare archaeological information from Florida and Pennsylvania in order to consider how each could be leveraged in landscape planning practices. Some of the challenges and opportunities are unique to each state, while others are shared across state boundaries. Perhaps most importantly, it is clear that databases of archaeological resources can offer important information for use in landscape-scale conservation design and planning, especially early on, when conservation priorities are being established.

In comparing the Pennsylvania and Florida databases, we identified several key themes: (1) issues surrounding the efficacy of interagency data sharing and the long-term durability of the data being shared; (2) the importance of developing web-based GIS programs of some scale that can be accessed, used, and understood by design and planning professionals; (3) the imperative for the development of publicly accessible archaeological information at scales that do not undermine site preservation; and (4) the imperative of regional coordination of data. While the statewide databases are similar in that they are complex efforts to archive archaeological information, they are fundamentally different in structure and accessibility. Both databases effectively archive archaeological site information for purposes of management and conservation. Both can be used for the development of statewide predictive and significance modeling. And both can be used more effectively in emerging archaeological practice, which contributes to landscape-scale conservation design.

Pennsylvania Cultural Resources GIS (PACRGIS)

PACRGIS is organized as a relational database that records multiple types of GIS data. This includes individual shapefiles for field surveys, archaeological sites, historical structures, historical bridges, historical cemeteries, National Register sites, and resource groups. There are 139,539 features of a variety of types in the overall PACRGIS. There are 116,275 historical features, whereas there are 23,164 features associated with archaeology (Figure 2). Over 50% of the historic features are referenced as buildings, while over 30% are not identified by a category. Interestingly, over 40% (47,825) of the historic features are located within 10 km of urban areas, and 26% (30,903) are located within 100 m of a stream. There is a clear spatial bias for historic features, with the majority near or within urban areas. This is a key factor when considering the application of these data for design and planning. Simply, if landscape-scale conservation priorities are derived solely from the distribution of documented resources, entire cultural landscapes could be undervalued and not fully represented in conservation plans.

FIGURE 2. The 23,164 archaeological features in the PACRGIS summarized by small watershed.

As mentioned above, there is spatial location data for 23,164 archaeological features (Figure 2). The 23,164 sites provided by the Pennsylvania Historic Museum Commission have basic categorical division such as listed National Register sites, demolished/100% destroyed, SHPO (State Historic Preservation Office)-eligible and -ineligible sites, KEEPER-eligible and -ineligible sites, insufficient-information provided-sites, and non-categorized sites. In total, 42 tables with over 125 attributes can be linked through the use of the relational database in Pennsylvania. The majority of the categories are focused on key characteristics for managing sites. Detailed cultural information is accessible, but largely through grey literature—that is, linked reports or site files when they have been scanned or digitally acquired.

The Florida Master Site File GIS Database (FMSF)

FMSF records multiple types of GIS data, and is structured as multiple shapefiles. This includes individual shapefiles for field surveys, archaeological sites, historical structures, historical bridges, historical cemeteries, National Register sites, and resource groups (Figures 3 and 4). The archaeological sites’ shapefile includes spatial location data, along with site boundaries, for 35,257 archaeological sites throughout the state of Florida, each represented by an individual polygon. FMSF records multiple attributes for each site with a clear emphasis on codifying and characterizing cultural and chronological information. The attribute fields include “SITEID,” “SITENAME,” “SITETYPE,” “CULTURE,” “SURVEVAL,” “SURVEYNUM,” “D_NRLISTED,” "SHPOEVAL,” “PLOTTYPE,” and “HUMANREMNS.”

FIGURE 3. The different features within the FMSF: (a) 35,357 archaeological features of the archaeological resources shapefile, (b) 169,441 features of the historical structures shapefile, (c) 1,333 features of the historical bridges shapefile, and (d) 1,313 features of the historical cemeteries shapefile.

FIGURE 4. Further illustration of attributes of features in the FMSF: (a) 1,770 National Register–listed sites in the National Register shapefile, (b) 18,648 features in the field survey shapefile, and (c) 2,650 features in the resource groups shapefile.

Whereas the SITEID and SITENAME fields record the essential and traditional site number and name, the SITETYPE field records important information about site function and environmental location. This field is split into six separate attribute fields, SITETYPE1–SITETYPE6, which is an important structural component of the database. The purpose of dividing this attribute into six separate fields is to account for multiple site components as well as the multiple functions of a site, or multiple functional areas a site might exhibit. There is a total of 101 different descriptors applied to the SITETYPE fields.

Much like the SITETYPE attributes, the CULTURE field is split into multiple attribute fields (CULTURE1–CULTURE8). The CULTURE fields record information regarding temporal associations and cultural affiliations. There is a total of 114 distinct descriptors applied to the CULTURE fields in the Florida database, and they cover both prehistoric and historic affiliations. The purpose of splitting the CULTURE field into eight distinct fields is to account for multicomponent sites. This is a common characteristic of sites throughout Florida. There are 14,537 sites (41.23% of all sites) in the database with at least two occupational components recorded. Furthermore, there are 6,512 (18.46%) with three components, 3,290 (9.33%) with four components, 1,897 (5.38%) with five components, 1,156 (3.27%) with six components, 739 (2.09%) with seven components, and 492 (1.39%) with eight components. Because of the number of multicomponent sites throughout the state, the splitting of the CULTURE field in such a way is a necessary structural aspect of the Florida database.

The majority of the affiliations are designations based on the archaeological culture regions outlined by Milanich (Reference Milanich1994). Each of these archaeological cultures has a distinct chronology, and the majority of these are post–500 BC affiliations (Milanich Reference Milanich1994; Milanich and Fairbanks Reference Milanich and Fairbanks1980; Wallis and Randall Reference Wallis and Randall2014). The exceptions to this are the Mount Taylor culture that begins during the Middle Archaic, and the emergent regional cultures of the Late Archaic period, including the Orange, Norwood, and Elliott's Point Complex cultures (Endonino Reference Endonino2008, Reference Endonino2010; Milanich Reference Milanich1994; Randall et al. Reference Randall, Sassaman, Gilmore, Blessing, O'Donoughue, Wallis and Randall2014; Sassaman Reference Sassaman2010). Furthermore, within a single culture region, the designations are split among the specific cultural chronologies for that region (e.g., Belle Glade I, Belle Glade II, Belle Glade III, and Belle Glade IV). In addition to this cultural information within the statewide database, there are bibliographic references to site surveys and field reports, if available.

Relating to assessment, there are three attribute fields devoted to National Register of Historic Places (NRHP) eligibility determinations: “D_NRLISTED,” “SHPOEVAL,” and “SURVEVAL.” The SURVEVAL field documents the surveyor's assessment of a site, whereas the SHPOEVAL field records the evaluation of the SHPO regarding NRHP eligibility. The attributes listed in these fields often do not correspond to one another due to differences in opinion between a field surveyor and the SHPO. The D_NRLISTED field is only used for sites that are already listed on the NRHP, and it records the date the site was listed. Additionally, the PLOTTYPE field records whether the polygon was mapped as a “GV” (geographic visualization) or through “NORM” (normal) methods (e.g., created through GPS coordinates of positive shovel test locations). The “HUMANREMNS” field records whether or not human remains were encountered at a site.

Comparing the GIS Databases

In trying to integrate these data into landscape conservation designs, several key comparisons are critical. First, the structures of the databases are completely different. Pennsylvania's data is structured as a multitable relational database, where attributes are grouped into several related tables. Because it is a relational database, the complex arrangement of joins and relates necessitates GIS expertise in addition to knowledge and experience with the archaeological resources. The database structure is accessible through a web-based GIS (PHMC 2018), which is useful for users who do not have experience in GIS, but this limits interoperability and any analytical processes. Florida's database, in contrast, is set up as a series of individual shapefiles, with multiple columns in a single attribute table. This structure allows for use with only basic GIS experience, and it is easily portable to the end user's personal device. Currently, there is no web-based portal to query or use these data, so end users must have software installed to analyze these data.

Second, for planning and design, site assessment of both eligibility and evaluation of preservation state is critical, especially when considering standardization. There are substantial differences in this content of the databases. First, over 70% of the sites contained in the Pennsylvania database are listed as having “Insufficient Information” for National Register of Historic Preservation eligibility, whereas only 8% are identified as having “Insufficient Information” in the Florida database (see Table 1). The differences here suggest not simply a difference in the completedness of the assessment of resources but a fundamentally different conceptualization of the determination of eligibility. For instance, if the classification of “Not Evaluated by SHPO” is included for the Florida data, 64% of features are either not evaluated or have insufficient information, which is similar to the 70% frequency for the Pennsylvania. While this variability is intimately known and discussed in state offices, planning professionals and private firms rely on expert assessment data in order to develop conservation designs. For example, if a designer wants to develop a conservation plan targeting impacted streams in Pennsylvania or West Virginia, there are standards to quickly combine with and overlay these on other design-relevant information, such as percent impervious surface. Standardizing the assessments will offer the ability to integrate these data into designs more completely.

TABLE 1. Comparison of Florida and Pennsylvania Cultural Resource Databases.

¹Cross-listed in the NRHP Eligible category; not used in the calculation of totals.

²Based on the number of site components identified in the CULTURE and RTIME fields.

³There is no spatial data for Florida classified strictly as “Wild” or “Natural” areas. The managed areas of Florida include national parks, national forests, national wildlife refuges, state forests, state parks, ecological restoration areas, wildlife management areas, and more.

Third, the two databases are somewhat similar in terms of the chronology of sites included. For example, both state databases have roughly 20% historic period features in the archaeological database (Table 1). Additionally, both databases have about 30% of features broadly listed as prehistoric. This designation is used where no specific chronological information is available, which presents unique challenges in regard to research contexts. The Florida data, however, contains unique chronological information for post-Archaic sites that is based specifically on the archaeological cultures of Florida. Because of this, comparison of the two databases requires creating an additional data field—which we labeled “RTime,” for regional time—that converts the Florida-specific post-Archaic chronology to the general chronology used for the broader southeastern United States (Anderson and Sassaman Reference Anderson and Sassaman2012). The chronological comparison in Table 1 is based on this field. From a landscape conservation design perspective, having comparative information about past cultural information tied to broad regional chronologies could offer transformative potential for integrating what we are learning about past coupled natural and human systems into future designs. Having these cues in databases offers designers new models for considering the shape and structure of landscape conservation designs.

Fourth, spatial representation is critical for landscape conservation design. Significant variation in the spatial distribution of features is an important point to consider here. In Pennsylvania, historic resources that are not classified as archaeological resources are predominantly found near or adjacent to urban areas. Fifty percent of historic sites are found within 10 km of urban areas in Pennsylvania. By contrast, only 20% of archaeological features are found within 10 km of urban areas in Pennsylvania (Table 1). Archaeological features in Pennsylvania are also highly correlated with streams and inland water, with 39% of the features found within 100 m of a stream (using the USGS National Hydrography Dataset). Florida, on the other hand, is dominated by resources within 10 km of urban areas. Importantly, nearly 30% of these resources are found within protected or managed areas, which differs drastically from Pennsylvania where only 3.6% of archaeological resources are within protected state lands. Understanding differences in land tenure, administrative boundaries, and socioecological contexts that encompass important archaeological sites and resources is critical to advancing their conservation and interpretation. The spatial distribution of resources when compared to modern settlements can influence conservation design. The cultural, topographic, hydrological, and ecological landscapes of Florida and Pennsylvania are fundamentally different, and this observed spatial variation is critically important for policies and decision-making as related to landscape conservation design and planning.

CONCLUDING REMARKS

Several years ago we started a project to develop processes for integrating cultural resources in landscape conservation design. Early on, it was clear that archaeological information and state databases offered the critical leverage to develop these processes. Although our work is ongoing, there are several conclusions that emerged when we applied our research to Florida and Pennsylvania site file GIS databases.

First and foremost, archaeological databases offer tremendous potential for use in design and planning, but there is much work to be done. We believe that this need for adapting cultural resources databases for use in emerging arenas such as landscape planning can provide the impetus for obtaining grant funds for both evaluating the information in the databases and restructuring them. Framing the need for restructuring in terms of its ability to increase the efficiency of landscape-scale planning may provide the medium for locating funds outside of the traditional sources for archaeological research (i.e., funding from state planning organizations, etc.). Building on trends in Pennsylvania and Florida, such as web-based access to GIS data by planning professionals in Pennsylvania and the Florida Public Archaeology Network's efforts to develop public databases and awareness of the threats to coastal resources, new investment could bring about new products that provide information during the design process—not after construction plans have been developed.

Second, there is an opportunity to open new doors of archaeological practice in design and planning. In the foreword to Paul Brace's Archaeological Resources and Land Development, O'Donnell wrote, “The fields of landscape architecture and archeology share a common ground—land with a history of human use” (Brace Reference Brace1984:v). In the technical publication, Brace (Reference Brace1984) describes how landscape architecture could more effectively integrate archaeological resources in site analysis and states that archaeological resources are our keys to understanding cultural landscapes. Brace's (Reference Brace1984) technical document has not been updated in 35 years. With the emergence of important landscape-scale conservation design and planning programs, such as the Landscape Conservation Cooperatives, there are diverse opportunities for integrating archaeological information in design and planning. Statewide databases hold exceptional promise as first steps in creatively bridging landscape planning and archaeology in practice.