Following a self-guided trail within an accredited US campus arboretum: The use of an AI-based app for tree identification and tour enrichment

Abstract

Being in an urban or developed area can adversely affect human well-being. On the other hand, human well-being is supported by recreational activities, which are often carried out outside, particularly in natural areas. Most research on such topics has focused on non-urban/non-developed areas, for which the term ecosystem services describes the direct and indirect benefits that people may receive. In developed regions, limited access to natural features can hinder these benefits. This study explored the specific case of a tree-walking route located within a developed campus in the US. This route is noteworthy for its diverse collection of 40 distinct woody species, which contributes to the campus’s green infrastructure. Two on-site observations were carried out to visually document the trees on the route and to understanding ecological value. An AI-based mobile application, ‘Picture This’, was used to follow the route as a self-guided participant. The results indicate that it is possible to use the application as a guide with approximately 84% accuracy. Its accessibility enhances its potential as a free resource for researchers, students, and nature enthusiasts.

Full Article

Following a Self-guided Trail within an Accredited US Campus Arboretum: The Use of an AI-based App for Tree Identification and Tour Enrichment

Merve Kalayci Kadak 

Being in an urban or developed area can adversely affect human well-being. On the other hand, human well-being is supported by recreational activities, which are often carried out outside, particularly in natural areas. Most research on such topics has focused on non-urban/non-developed areas, for which the term ecosystem services describes the direct and indirect benefits that people may receive. In developed regions, limited access to natural features can hinder these benefits. This study explored the specific case of a tree-walking route located within a developed campus in the US. This route is noteworthy for its diverse collection of 40 distinct woody species, which contributes to the campus’s green infrastructure. Two on-site observations were carried out to visually document the trees on the route and to understanding ecological value. An AI-based mobile application, ‘Picture This’, was used to follow the route as a self-guided participant. The results indicate that it is possible to use the application as a guide with approximately 84% accuracy. Its accessibility enhances its potential as a free resource for researchers, students, and nature enthusiasts.

DOI: 10.15376/biores.20.4.8755-8776

Keywords: Artificial intelligence; Mobile phone applications; Ecosystem services; Campus planning; Climate change; Urban areas

Contact information: Department of Landscape Architecture, Faculty of Engineering and Architecture, Kastamonu University, 37150, Kastamonu, Türkiye; Email: mkalayci@kastamonu.edu.tr

INTRODUCTION

University campuses are established in many locations, with a general goal of increasing the number and value of educated people. One of the crucial missions is to provide a place and environment for the education of qualified people. At the same time, these campus areas are public or semi-public areas, which are organized as urban areas (Huang et al. 2023). Therefore, they also have another function, which is to provide services to the public. Besides, campuses are also significant for controlling the impact of climate change on urban areas. Campuses can include vegetation that is natural or culturally created (Paudel and States 2023; De Montis et al. 2024). Vegetation cover provides protection from climate change effects, especially focusing on landscape structures (De Montis et al. 2024; Kalayci Kadak et al. 2024). In this context, campuses should be designed to provide benefits from these areas to primary users, such as students, faculty and staff, along with the public. Therefore, providing some of the ecosystem services to all users, including the public, who are overwhelmed by the urban structures, can be possible. Most importantly, such services can help to alleviate the adverse effects of climate change (Kim et al. 2024; Gul 2025; Kalayci Kadak 2025; Kaya et al. 2025), such as urban heat islands and extreme precipitation, in urban areas, along with contributing to conservation of human well-being.

Ecosystem services, including those associated with trees on a university campus, are crucial for the sustainability of human’s healthy life (Kalayci Kadak 2025). These services also have vital importance in improving living standards. Ecosystem services are conceptually categorized into four main classes according to their functions (Fig. 1): regulating services, provisioning services, supporting services, and cultural services (Leemans and De Groot 2003; King et al. 2014; Zhou et al. 2024; Kalayci Kadak 2025).

Fig. 1. Ecosystem Services (Millenium Ecosystem Assessment, 2005)

The first class, regulating services, has some benefits within ecosystem processes, such as regulating climate change effects, controlling epidemics or other types of diseases, and water purification. The provisioning services provide much help, including food security and supplying fresh water for all creatures. The supporting services provide critical benefits for soil formation and nutrient cycling. Finally, the fourth class, cultural services, has several benefits. Supporting services offer non-material benefits, such as recreational activities, educational issues, aesthetic gains, and engagement in community life from ecosystems (Castro et al. 2016; McInnes and Everard 2017; Cianchi et al. 2024). The perspective of ecosystem services has become progressively crucial in regulation, protection, and management of environmental rules due to providing an assessment frame for the ecosystem’s benefits (Li et al. 2022; Liu et al. 2022; Lenhardt 2023; Gül and Esen 2024). Incorporating stakeholders in the assessment process is important regarding determining primary demands and needs, which can be obtained from ecosystem services by local administrations (Beaumont et al. 2017; Friedrich et al. 2020; Li et al. 2022). There has been an ever-increasing importance of mapping by using technology-based techniques for the assessment of ecosystem services’ benefits to ease policy-making processes (Daily et al. 2009; Brunina et al. 2016; Friedrich et al. 2020).

Nowadays, there are a lot of advancements in geographic information systems (GIS) and remote sensing (RS) tools, which are used for mapping methodologies by many researchers. These improvements allow researchers to understand ecosystem services and benefits related to ecosystem services by analyzing relationships between natural and cultural systems (Burkhard et al. 2012; Bagstad et al. 2013; Vargas et al. 2018; del Río-Mena et al. 2020; Zhou et al. 2024). Advancing technology accompanies the development of artificial intelligence (AI) technology, which is used in all assessment areas, including innovative or scientific, for the commonly used applications in decision-making processes (Koumetio Tekouabou et al. 2023; Prodanovic et al. 2024). Accordingly, advancing technology tools have become considerably important in the utilization rate of ecosystem services in urban or semi-urban areas (Chee 2004; Morya and Punia 2022). Integrating state-of-the-art tools, particularly machine learning algorithms, into advancing innovative methodologies is essential for benefits that are used by people for sustainability of Earth. Thus, managing natural resource-based benefits can be more perceivable and sustainable (Raihan 2023; Kalayci Kadak et al. 2024). These approaches will eventually be used more widely in urban planning.

Among the places to understand observing ecosystem services is the arboretum. Arboretums have many goals for conservation of natural resources and educational purposes. Crucial objectives for the ecosystem benefits include biodiversity protection and sustainable use of natural values. Key values such as ecological approaches, educational needs, and community engagement affect the objectives and managing/designing natural resource value of the arboretum (Roman et al. 2017, 2022). In this context, some campuses have been declared as arboretums as part of the Arboretum Accreditation Program (https://arbnet.org). The Storrs campus, the main campus of the University of Connecticut, was certified as a Level II arboretum in July 2024 (Reitz 2024). The arboretum character of semi-public or public areas, such as campuses, helps urban landscapes be in harmony with sustainable nature (Lee 2021; Li et al. 2022). Accreditation by national or international organizations supports not only natural resources conservation but also encourages community engagement via educational goals and recreational activities. In this regard, the design of a campus, which is among the semi-public areas, is a multifaceted effort to involve natural values in functional community solutions (McDonald et al. 2018; Boyd 2022). Also, university campuses are essential for promoting social connections by the public (Moreno and Franquesa 2023). In this context, some specific routes, which can be divided into two categories as self-guided and guided, have been created within campuses by the university administrations (Melo et al. 2020). These routes are significant for responsible usage and participatory approach on campuses. It allows users to explore natural or artificial values through the campus area (Al-hagla 2010; Moore et al. 2012; MacLeod and Hayes 2013; MacLeod 2016; Janeczko et al. 2021).

The effectiveness of self-guided trail routes can be advanced by utilizing new technologies, such as a mobile AI application. Participating in the community adds and enriches local memories and cultural values into the visitors’ experience (D’Antonio et al. 2022; Marion 2023). These benefits also contribute to the cultural services, which is a type of ecosystem services (Koo et al. 2013). Additionally, using self-guided trails will support environmental literacy among visitors by attaching individual technological development, such as AI applications, on mobile phones. These trails can help individuals observe ecological issues (Janeczko et al. 2021; Marion 2023). Therefore, taking advantage of AI for trails is rational to ensure sustainability while planning a modern campus. Designers and researchers should be careful while integrating ecological factors into the strategic planning of campus needs. Thus, a balance can be achieved between protecting natural resources and usage goals (Orenstein et al. 2019). At present, this integration can be possible with technological support (Yang et al. 2020; Li et al. 2023). Overall, it is important to espouse innovative approaches, such as utilizing AI (Mashhood et al. 2023).

This research focused on the main campus of the University of Connecticut (UConn), located in Connecticut State, USA. The primary aim was to understand the opportunity to track the self-guided tree route, organized by the university administration, through an AI application. Additionally, the benefits obtained by self-guided routes from ecosystem services were considered. In this study, where striking results were derived, the value and importance of urban or semi-urban routes within the framework of different dynamics were also analyzed. Several previous studies (Kothencz et al. 2017; Gould et al. 2019; Zhou et al. 2022) concluded that outcomes of recreational use of ecosystem services, especially in urban or semi-urban areas, affect life quality favorably. The main goal of this paper is to reveal the possibility of assessing urban ecosystem services’ benefits on life quality, by harnessing an innovative perspective by AI. The results of the study aimed to represent an approach, especially for decision-makers, in view of current technological improvements for the sustainability of the world. Additionally, the results of the article potentially will serve sustainable Development Goal 11 (SDG 11) of the United Nations. Because cultural ecosystem services can contribute to most of the SDGs (Xu and Peng 2024). These objectives can be summarized as follows: The urban areas must be sustainable, available for everyone, resilient, and safe. In summary, the study contributes to more liveable urban lifestyles.

EXPERIMENTAL

This study focused on the tree route located in the main campus of the University of Connecticut (UConn), which houses 40 special tree species (see Appendix) (Kask 2008; Reitz 2024). The AI application, called “Picture This,” which can be downloaded to mobile phones (available on the Apple App Store and Google Play Store), was used in the tracking of this tree route.

Framework Definition and Study Area

UConn’s main campus, called the Storrs campus, is located in Mansfield Township, which is east of Hartford, in the Connecticut’s Capital Planning Region, located near the coordinates 41° 48′ 26″ N and 72° 15′ 9″ W (Fig. 2).

The study area, the University of Connecticut’s main campus, was awarded prestigious arboretum accreditation in 2024. The application for Level II Arboretum Accreditation, approved by ArbNet Arboretum Accreditation Program, is administered by the Morton Arboretum.

To be accredited as a Level II arboretum, the applicant must have the requirements listed below:

• There should be at least 100 species on the campus.
• There should be at least one person as the agent of maintaining the facility.
• The applicant must be able to show a reliable inventory of the collection.
• Educational and public benefit programs should be offered.

Fig. 2. Location of the tree route in the study area, CT (Original, 2025)

In this context, the campus of UConn has more than 425 species within 90 unique genera. It presents a system that is available for campus users and provides free and informative labelling for campus trees. This system is GIS-based and is being updated periodically.

Moreover, the campus was named Tree Campus, which is the first in the Connecticut state, by Arbor Day Foundation. There are many ongoing actions, including the following, to preserve the qualities of the campus by the Office of Sustainability (Office of Sustainability):

• Organizing an advisor committee for campus trees
• Managing campus tree-care plan
• Observing Arbor Day
• Connection with community beyond campus borders.

A self-guided tree walking route (Kask 2008) has been formed considering these goals, (Appendix: Supplementary Material (SM)-I).

Methodology

The study was carried out in four phases: Definition of framework, data and tool collection, study area research, and office work. The workflow chart in Fig. 3 shows the steps of the methodology.

Fig. 3. Methodology workflow diagram

In the first phase, the concepts of the study subject were focused comprehensively to assess the study’s outputs in terms of the knowledge gaps. In the second phase, the data and application, which are the main materials of the study, were obtained. The document related to the tree route (Appendix: SM-I) was examined and elaborated. Picture This™ (Glority Global Group) AI-aided application was downloaded to a mobile phone.

The third phase was the most critical stage of the study. This phase required high focus and precision. The study area was visited two times at different seasons (summer-autumn) due to foliation and defoliation periods. The visit was scheduled to examine the site survey along the route and take photos of trees in different hours of the day as bright as possible. Seasonal changes are essential for the machine learning of AI-based applications and visual documentation processing. This enables the application to be more productive. For these reasons, having high focus and precision was crucial.

In the fourth phase, visual documents, which were obtained by site survey, and the photographs, used to identify by AI, were organized. All tree data were tabulated. Afterwards, the information provided by AI and the tree walking tour guide information, prepared by the arboretum committee, were compared proportionally and graphed out. Thus, the results of the AI were verified with a cross-validation approach. By this means, all data were promoted to be more understandable for readers and researchers. Lastly, the possibility of using AI in self-guided or even unguided routes was interpreted based on the accuracy of AI identification. It is probable that these results and perspectives will support educational, recreational, aesthetic, and spiritual cultural services.

RESULTS

This section presents and discusses the main results of the research. These results were obtained by visiting the study site twice, in July 2024 and October 2024. The goal was to make the assessment more reliable by surveying and examining the plants in two different seasons (Primack et al. 2023). Seasonal changes affect the appearance of the plant habitus and foliation (Kuper 2013; Xu et al. 2022). The ‘Picture This’ AI-based application was used while visiting the study site as mentioned in the methodology section.

Table 1. Campus Trees of Walking Tour and their Status

One of the most striking results is that the AI-based application named ‘Picture This’ showed the capability of detecting almost all plants on the tree walking route. Previous studies have shown that plants can be described using remote sensing technologies (Huang and Asner 2009; Cerrejón et al. 2021). On the other hand, the ecosystem services, especially cultural ecosystem services, were assessed using social media photos (Egarter Vigl et al. 2021; Tulek 2023). However, identification of plants using a free mobile AI-based application has not been commonly used so far, according to the current literature. AI-based applications or similar AI technologies employ machine learning procedures. Effective plant identification can be attributed to the fact that a machine learning process supplies access to more data (Willcock et al. 2018) and it assesses rarer values (Cerrejón et al. 2021). Thus, it provides more reliable results after assessment by using a bigger data repository. This substantial result should be compared with previous research that focus on identifying woody plants. Results of the statistical analysis showed that the application, ‘Picture This’ detected approximately 84% of all standing trees on the route. This result is a promising finding in terms of the use of AI-based applications on a seasonal basis.

Table 2. Misidentified Plants

In the study, the plant identification application was used to determine the effectiveness of AI while using self-guided routes, which provide benefits from cultural ecosystem services to people. While following the route that houses forty species, it was found that four trees; Sassafras albidum, Betula davurica, Styrax obassia, and Hovenia dulcis, were not present in their intended locations (Table 1).

Therefore, the assessment of AI-based application, which is the primary material of the study, was performed with 36 species.

The application misidentified only four out of 36 species: Sciadopitys verticillata, Koelreuteria paniculata, Pinus parviflora, and Kalopanax septemlobus. The visual identification documents related to undetectable species are given in Table 2 (IDs in Table 2 refer to Table 1).

Furthermore, there were two species unidentified on the second tour, despite being identified correctly on the first tour: Pinus rigida and Sequoiadendron giganteum (Table 3).

Table 3. Misidentified Plants on the Second Tour

In summary, ‘Picture This’ AI-based application successfully identified 30 out of 36 species in the route (Table 4).

The ecosystem’s qualifications, such as aesthetics (Kaya and Corbaci 2025), cultural heritage values, and habitat provision, are variable (Sari and Karasah 2023; Tulek 2023). Besides, these depend on whether the woody plants are native or non-native (Sari et al. 2020).

At the end of the research, it was determined that 17 species are native, two species are hybrid, and 21 species are non-native. Besides, four out of 40 species (one native and three non-native) were not present at their location, as mentioned earlier.

The origins of the woody plants on the campus route were examined, and they were classified as native, non-native, or hybrid (Table 5).

Table 4. Identified Plants