Abstract
With more than 120 institutions and more than 300 global members, the annual CELA conference has showcased the most relevant topics related to landscape architecture (LA) research for 100 years. CELA organizes its research by topic into content areas called tracks. Despite a series of overviews of LA-based research trends, systematic examinations of how well the conference tracks align to such topics have never been conducted. This study evaluates tendencies in CELA published abstracts in recent years to determine which topics are most relevant to current LA-based research and assesses how the CELA track system parallels these trends. We gathered CELA abstracts and analyzed them comprehensively and by individual research track. Despite its century-long legacy, thorough data exist only on CELA abstract submissions from 2013–2019, a total of 2,426 abstract entries over a seven-year period. We applied latent Dirichlet allocation models on abstracts to identify the top 20 topics. Trends show that communitybased approaches involving climate change, hydrology, land reuse, and social equity issues are at the core of most recent CELA research while research also continually seeks to embrace new technology and tools. The topics we identified as most relevant to LA research inform recommendations for CELA’s track system; these topics can also be used by other LA-based conference venues to help reshape their organizational structures. The methodology can be replicated and applied to different conference venues.
INTRODUCTION
With more than 120 institutions and more than 300 members globally, the annual Council for Educators in Landscape Architecture (CELA) conference has showcased the most relevant topics related to landscape architecture research for 100 years. CELA traces its beginnings to the 1920s. Originally called the National Conference in Instruction in Landscape Architecture (NCILA) in 1920 (Executive Committee of CELA, 1976), CELA began as a relatively informal gathering of instructors of landscape architecture (LA) to exchange information, hold annual meetings, draft resolutions, and publish research and reports (Hodges & Rutz, 1997). Yearly conferences were initiated in Cambridge, Massachusetts in 1951 and have been continually held since then (with the exceptions of 1962 and 2020). By 1971, the first NCILA publication (a newsletter titled the NCILA News) was disseminated. By 1976, the name was changed from NCILA to CELA, mostly to encourage more international participation.
CELA’s current mission is to encourage, support, and further education in the field of LA specifically related to teaching, research, scholarship, and public service. The CELA track system is the conference’s organizational structure, which is based on current topics relative to LA-based research. The system is managed by the CELA vice president for research and creative scholarship along with a structured assembly of scholarly experts who provide peer review for research and scholarly works focused in content areas representing the discipline of LA. These content areas are called tracks, and each track is intended to be representative of a specific topic relevant to LA research. Each track has a chair or cochairs who help govern the content of research in it. CELA currently has 14 tracks:
Communication and Visualization
Design Education and Pedagogy
History, Theory, and Culture
Landscape Performance
Landscape Planning and Ecology
People–Environment Relationships
Research and Methods
Service Learning and Community Engagement
Urban Design
annual theme track (changes yearly),
Research by Design and Implementation (created in 2019)
Geospatial and Digital Analytics (created in 2020)
Landscape Architecture for Health (created in 2020)
Diversity Equity and Inclusion (created in 2020 for 2021)
LA researchers submit abstracts to the track in which their research best fits in hopes of being accepted and presenting at the annual conference. Unfortunately, authors who do not find a track that matches their research topic have to submit to a less relevant or broader track; those whose submissions fit in more than one track may become confused when submitting. Authors with accepted abstracts are invited to submit full papers that have the potential to be published in Landscape Research Record, the CELA peer-reviewed publication venue for conference submissions. Landscape Research Record began in 2013 as a postconference publication and is published online only.
CELA research tracks are established and maintained through CELA’s Standing Committee for Research and Creative Scholarship. Although any track changes must be proposed and voted on by the CELA Board of Directors, other than the recent track additions, there have been relatively few changes in the CELA track structure since its initial implementation in 2008. Simultaneously, the demand for faculty research, the scope of interdisciplinary research, and the number of research conferences related to LA design (Geodesign Summit, Environmental Design and Research Association, International Federation of Landscape Architecture, etc.) are all increasing.
The CELA track system was developed over a decade ago, and to date there has been no comprehensive analysis to determine how these tracks perform. Such an analysis could be an initial step for developing suggestions on how the track structure could be modified to help CELA continue to align its research topics with the current needs of the LA profession. Simply put, not enough is known about how applicable or active the contemporary CELA research tracks currently are. An evaluation of the performance of CELA tracks would allow for necessary adjustments to the current track structure to (1) broaden its appeal to educators, students, and professionals; (2) increase the relevance of CELA by improving its abilities to capture contemporary and cutting-edge topics in a rapidly changing research climate; and (3) enhance CELA’s capabilities for the inclusion of new and interdisciplinary fields in the purview of LA.
Although there are a small number of narrative overviews of LA research trends in the current literature, systematic and longitudinal examinations of topics researched at CELA have not been conducted or linked to these existing overviews. This article evaluates tendencies in CELA tracks based on abstract submissions over seven years to determine the topics most popular in current LA-based research, evaluates how these research trends have changed, and presents a new set of contemporary LA-based research topics and a replicable approach for other conference venues to evaluate their systems. The research determines how the CELA track structure performs when examining data generated by existing research. We use this evidence to suggest future adjustments to the track system. To achieve this, we gather CELA abstract submissions and apply a latent Dirichlet allocation (LDA) approach to conduct cross-comparisons of topics intensified by the LDA with the current CELA track (or topic) system. Our approach and framing is useful because it shines light on what is most critical to LA-based research and then compares these findings with the current literature and the current CELA track structure.
A REVIEW OF LA RESEARCH TOPICS
LA has long been researched through a variety of lenses. Recent literature has increasingly emphasized the multifaceted nature of LA and the need to democratize the process of the profession (Hester, 2006; Linn, 2007; Francis, 2016). The interdisciplinary nature of LA inherently supports topics and methods in a broad range of natural science, social science, and humanities-related disciplines (Deming & Swaffield, 2011; Bruns et al., 2016). In addition, the recent advancement of the notion of research by design indicates more comprehensive participatory processes with fewer barriers between experts and the public and more integrated design-based research (Deming & Swaffield, 2011; Nijhuis & Bobbink, 2012; Nijhuis & de Vries, 2019). These circumstances suggest a need to reexamine important LA-based research topics—ones traditionally considered important and those of new interest to the contemporary LA community.
Since about 2010, an emerging body of literature has investigated the topic areas of published LA research over time. As one of the first studies to formulate topical areas, Powers and Walker (2009) reviewed articles submitted to Landscape Journal from 1981 to 2005 to identify the primary research themes of LA, classifying the subjects into 11 categories (derived from the CELA conference tracks during 2007–2008): History and Culture, Landscape Planning and Ecology, Human and Environment Relationships, Design Theory, Urban Design, Landscape Design Implementation, Communication and Visualization, Methods of Inquiry, Sustainability, Landscape Architecture as Profession, and Design Education and Pedagogy. Among them, the first three categories constituted more than half of all papers. This framework was referenced in multiple later studies.
Since Powers and Walker (2009), six other articles examined the main topics of LA research and identified theme areas based on diverse scholarly sources and professional documents (Table 1). Three of these six articles used citation or content analysis to summarize topics included in published articles in LA-related journals. For example, Gobster, Nassauer, and Nadenicek (2010) reviewed articles published in Landscape Journal and compared their strengths against competing journals. They proposed a framework of nine different themes and noted that Landscape Journal was strongly positioned in Landscape History, Design Case Studies, and LA Education and Profession but was weaker in such areas as Environmental Management, Policy, and Perception compared with other journals. New topics such as Landscape Change and Measurement and Climate Change were also revealed. Cushing and Renata (2015) expanded on this approach and included articles published in Landscape Journal, Landscape Research, and the Journal of Landscape Architecture. The topics identified largely overlapped Powers and Walker, with a few unique categories, such as Parks and Open Space and Gardens. Vicenzotti et al. (2016) reviewed articles in the European-based journal Landscape Research. They demonstrated clear patterns of topic diversification in LA research between 1976 and 2014. Before 1980, two topics, Landscape Planning and Ecology and Sustainability and Perception, dominated the published topics (81%), and the majority of topics had less than 2% representation. By 2011–2014, although the primary topics stayed strong, other topics such as History and Culture (15%) and Human and Environment Relationships (28%) displayed rapid growth.
Previous Studies That Identified Top Topics/Themes in LA Research
Expanding the scope from published articles to other sources, two studies synthesized the findings from Powers and Walker (2009) along with other academic and professional frameworks. Deming and Swaffield (2011) reviewed the IFLA definition of LA professional activity, LA Body of Knowledge (LABOK) topics for professional development, and the Le Notre key areas of knowledge and expertise and linked these findings to CELA member surveys. Five main topics with subcategories for LA research were identified: Sustainability and Biophysical Processes, Social and Cultural Processes, Health and Well-Being, Urban Regeneration, and Tools and Technologies. Future research directions, such as theorization of the profession in the late modern global city paradigm, were identified. Langley, Corry, and Brown (2018) used similar approaches and included the Fein Report (Fein, 1972), as well as topics of the Council of Landscape Architectural Registration Boards required courses.
Another study (Meijering et al., 2015) used the Delphi method with an expert survey to acquire deeper understandings of LA research topics based on Deming and Swaffield’s (2011) categorization. A diverse panel of experts was selected from around the world representing both scholars and professionals. The findings showed some patterns that diverged from studies that used review methods. For example, the most important topics identified were Human Dimensions of Planning and Design, Green Urban Development, Built Environment and Infrastructure, and Global Landscape Issues.
If we synthesize the findings from literature, two patterns become notable. Reviews that summarize refereed journal articles report relatively consistent topic areas, but reviews that include other sources tend to identify new topics of interest to the educational and practical community (Langley, Corry, & Brown, 2018). For example, Milburn and Brown (2016) showed an increased productivity of LA researchers but also pointed out the differences in the topics rated as important by LA faculty versus practicing professionals. In addition, the literature shows a broadening of the topics covered in LA-related research and thus a critical need to identify emerging and contemporary trends. Clearly, the primary topics related to LA shift through time yet continually yield a diversity of thought, methods, and approaches. Continually updating CELA tracks based on this pattern is therefore necessary to effectively study and practice LA.
RESEARCH OBJECTIVES
Current CELA tracks mostly equate to topics already identified in the published articles analyzing LA research topics, with a few missing or only tangentially covered topics. The 2020 CELA track structure appears to be primarily based on the topics identified by Powers and Walker (2009). A systematic review of published conference abstracts that focuses exclusively on CELA trends of the most recent few years can contribute to the discussion of shifts in priorities in LA-based research. While journal articles represent high-quality published materials, the print times are usually longer than conference abstracts. Furthermore, not all LA-based research gets published as journal articles due to its heavy reliance on design. As such, conference abstracts from LA-focused meetings may be a better proxy than published articles in landscape-focused journals when evaluating such a subject. Conference presentations are usually the first step in developing and eventually publishing a research article. Feedback from such experiences provides the author(s) with valuable feedback for continuing the research (Newman, Ozdil, & Li, 2018). A review of publishing venues that offer faster peer review (such as conference abstracts or proceedings) would allow for a larger volume of submissions to be evaluated, would capture a broader range of research topics, and would be complementary in generating the emerging patterns of research that may be published as articles in the future. Also, since Powers and Walker (2009) adopted the 11-category framework in classifying LA research topics, many subsequent reviews used this categorization as the code guide to form their own categories (Cushing & Renata, 2015; Vicenzotti et al., 2016). However, emerging topics may differ from the established patterns of the previous decade. Therefore, topics that are developed entirely from data without propositions or thematic guides can offer new insights.
This study evaluates tendencies in CELA abstract submissions between 2013 and 2019 to determine (1) what topics have been relevant to current LA-based research and (2) how these research trends link to and compare with the current CELA track structure. Because the 2021 CELA track structure has no abstracts currently submitted, we use the structure from the 2020 conference, which does not include the Diversity, Equity, and Inclusion track, includes the Sustainability track, and separates the Research by Design and Design Implementation tracks into two distinct tracks (instead of the 2021 Research by Design and Implementation track). Full papers were excluded from the analysis because some data were missing during this time frame, the abstracts in each paper would have duplicated existing abstract data, and because full papers allow for an additional peer-review process, making the process different from the abstract submission process (which has only one double-blind peer review opportunity). It is important to note that the time frame of this research is limited to 2013 and after because this is the when CELA began using an online submission system; thorough digital data before this time frame are nonexistent. While the tracks were still in use, the data were not archived in an open online shareable database.
METHODS
CELA Abstract Extraction
We gathered the published abstracts from all available CELA abstract publications (https://thecela.org/past-conferences/). As noted, despite its century-long legacy, digital abstract content data on CELA annual conference abstract submissions only exist annually from 2013. After removing duplicates, there were a total of 2,426 abstract entries over the seven-year period (Table 2). The tracks to which the authors choose to submit their abstract may reflect their assessments of the most closely aligned area from the list of the CELA tracks. We used abstracts as a proxy to CELA conference presentations and papers because abstracts better represent research topics of interests to the broader LA community.
Number of Abstracts by Year and Track
Abstract Processing Using Natural Language Processing
Natural language processing techniques were used to process the abstracts and transform them into bigram representations of the text corpus using R packages. Figure 1 shows the analysis process. The abstracts (including title and keywords) were first cleaned by removing punctuation, numbers, and special characters, and common abbreviations were replaced with the original words. Tokenization was performed to extract words from every abstract. After removing common stopwords, lemmatization was performed to remove the inflectional endings of words and return the dictionary form of a word. The resulting word frequencies were reviewed reiteratively by the researchers, and domain-specific stopwords were created from the most frequently appeared 1,000 words. These stopwords belonged to three types: commonly appeared in all studies (e.g., methods, results, conclusions, and research); procedure verbs without significant substance (e.g., propose, explore, include, develop); and words that appeared in all LA research (e.g., landscape, architecture, designer). After this process, documents were transformed into a word-frequency format for modeling.
Abstract preprocessing workflow.
Topic Modeling using LDA
Topic modeling is a probabilistic model used to identify latent topics that occur in text. It has emerged as a valid approach to distinguishing the underlying structures in text (Wei & Croft, 2006). Among methods of topic modeling (e.g., latent semantic indexing, probabilistic latent semantic analysis, and correlated topic model), LDA proves to be the most widely used in text mining for scientific and popular media (Edison & Carcel, 2020). Since 2015, LDA has been introduced as an effective technique for literature reviews (Mo, Kontonatsios, & Ananiadou, 2015) and has shown great potential in revealing topics and trends of research across subdomains and over time (Sun & Yin, 2017). This approach has been applied to identify trends using data from annual conference publications in transportation (Das, Sun, & Dutta, 2016).
LDA uses an unsupervised modeling approach to reconstruct the writing process by generating text on a given topic (Figure 2). It employs a Markov chain Monte Carlo method to sample and approximate the underlying distribution (Gibbs sampling) and runs iteratively. The model represents each document as a mixture of K latent topics. Each topic displays a multinomial distribution of a vocabulary that consists of W words. For each of Nj words in document j, the model uses a generative process:
Choose a topic zij ~ Multinomial (θj), where i is the word index, j is document (abstract) index, and θj is the topic distribution of abstract j.
Choose a word wij ~ Multinomial (φk), where k is topic index, φk is the word distribution of topic k; the words in each topic and topic in each document follow Dirichlet priors (Blei, Ng, & Jordan, 2003; Blei & Lafferty, 2006).
Latent Dirichlet allocation model.
Analyzing Topic Distribution
We used LDA models to extract the top 20 topics from 2013 to 2019 from all tracks, following LDA modeling conducted in the literature (Blei, Carin, & Dunson, 2010; Das, Sun, & Dutta, 2016). Based on the LDA model, we obtained the posterior topic distribution of each abstract and the word distribution of each topic. For each topic, we displayed the top 20 words in word clouds with the highest probability, because the word cloud approach prioritizes higher probability keywords and offers a visual way to compare the most important ideas across different topics. Because LDA is an unsupervised learning approach, the top 20 keywords with the highest posterior probability were interpreted by two authors independently, and topic titles were developed by consensus.
As we aimed to identify primary research topics and salient current trends across the different CELA tracks, we extracted the per-document-per-topic probabilities and classified each abstract based on the topic with the highest probability. The Film, Research by Design, and the yearly theme tracks were excluded from the subsequent analysis, as they existed for less than three years and were not comparable with the other stable tracks. We generated crosstabs and plotted topics by tracks and years as a chord diagram and line charts. Finally, we examined the track-level patterns and identified three key topics that emerged from each track.
Validation of the model results was performed by manually coding the abstract topics. We took a stratified random sample of 500 abstracts from all abstracts with the tracks as the strata. The 20 topics identified using the LDA model were used as the initial codebook. Trained researchers conducted the coding by reading the abstracts and determining which topic each abstract was most closely aligned with. We calculated the confusion matrix to show the agreement between the LDA and manual approaches for all abstracts and each topic.
RESULTS
Analyzing Current Topics Most Relevant to LA-Based Research
The LDA results identify the most common latent topics of research in CELA abstracts. These 20 topics are: Historical and Cultural Landscapes (Topic 1), Parks and Public Space (Topic 2), Climate Change (Topic 3), Design Pedagogy and Curriculum (Topic 4), Ecosystem Services and Management (Topic 5), Garden and Planting Design (Topic 6), Landscape Performance (Topic 7), Material and Construction (Topic 8), Land Development and Reuse (Topic 9), Sustainable Agriculture (Topic 10), Student Learning and Studio Process (Topic 11), Technology and Digital Tools (Topic 12), Design Theory (Topic 13), Urbanism and Urban Design (Topic 14), Hydrology and Stormwater (Topic 15), Visual Communications (Topic 16), Human Preference and Behavior (Topic 17), Environment and Public Health (Topic 18), Community Engagement (Topic 19), and International and Vernacular Landscapes (Topic 20). The keywords of each topic are visualized using word clouds (Figure 3), with the relative size of the word representing the probability of the word belonging to a topic. Because of visual biases in word clouds related to different lengths of the words (Tufte, 2001), we provide the actual keywords in Table 3. For example, the most commonly appeared keywords associated with Climate Change (Topic 3), included change, climate, scale, coastal, environmental, resilience, flood, infrastructure, future, and sea. Other words (such as adaptation, waterfront, risk, intervention, and scenario) also emerged as high-probability words under this topic. Together they suggest that research has focused on impacts and implications related to water. Most water-related research tends to focus on climate change, adaptation, and intervention strategies for flood resilience. For Landscape Performance (Topic 7), a pattern emerged with the keywords performance, assessment, environmental, economic, criteria, and quantitative that describe the goals of using quantitative metrics to assess environmental and economic effects of design. Keywords like LAF and Foundation demonstrate the role of the Landscape Architecture Foundation in leading and facilitating this mission. Specific evaluation items. such as heat, island, thermal, wind, and microclimate. show a focus on the microclimate-related effects of landscape planning and design.
20 topics of LA research (2013–2019).
List of Most Frequent Keywords of the 20 Topics
Results from manual validation showed that on average, 92% of the LDA classifications agreed with the manual coding, suggesting that the models produced results that are consistent with how landscape researchers would interpret the main topics of abstracts. Topics related to Climate Change (Topic 3), Garden and Planting Design (Topic 6), Hydrology and Stormwater (Topic 15), and Material and Construction (Topic 8) showed 100% agreement between results classified by LDA and manual coding. Topics that showed lower agreements (84–86%) included Community Engagement (Topic 19), Historical and Cultural Landscape (Topic 1), Landscape Performance (Topic 7), and Urbanism and Urban Design (Topic 14). The validation confusion matrix is displayed in Table 4.
List of Most Frequent Keywords of the 20 Topics (Overall Agreement = 92%)
Current Topics by CELA Track
Figure 4 shows the CELA tracks under which abstracts of the 20 topics were submitted, where the interconnections between topics and tracks are displayed in Bézier curves. The upper half of the ring shows the CELA tracks, and the lower half shows the topics. The thickness of the curves linking the tracks and topics illustrates the strength of the interconnections: the frequencies of each track contributing to each topic. Some topics in themselves form the majority of abstracts under a single track (as represented by a thick curve from a topic to a track), while other topics are drawn from multiple tracks (represented by several dispersed thin curves going out from a topic). For example, 59.7% of the papers on Historical and Cultural Landscapes (Topic 1) were contributed by the History, Theory and Culture track. The Design Education and Pedagogy track contributed 59.9% of the papers under Design Pedagogy and Curriculum (Topic 4) and 82.7% of the papers under Student Learning and Studio Process (Topic 11). Similarly, the People–Environment Relationships track served as the major contributor to two topics: Human Preference and Behavior (Topic 17, 47.1%) and Environment and Human Health (Topic 18, 66.7%). Other strong linkages exist between Ecosystem Services and Management (Topic 5) and the Landscape Planning and Ecology track (43.8 %); Landscape Performance (Topic 7) and the Landscape Performance track (30.0%); Design Theory (Topic 13) and the History, Theory, and Culture track (32.5%); Visual Communications (Topic 16) and the Communication and Visualization track (37.0%); and Community Engagement (Topic 19) and the Service Learning and Community Engagement track (35.4%). Some topics are relatively equally distributed across the different tracks. For example, Garden and Planting Design (Topic 6) draws considerable numbers of papers from the Design Implementation, History, Theory, and Culture, Landscape Performance, Landscape Planning and Ecology, Research Methods, and Sustainability tracks. Such topics also include Land Development and Reuse (Topic 9), Sustainable Agriculture (Topic 10), and Technology and Digital Tools (Topic 12).
Interrelationships between topics and tracks.
With respect to the topic distributions of each track, some tracks are centered on a few topics, while others see a more balanced distribution across many different topics. For example, Design Curriculum (Topic 4) and Student Learning (Topic 11) dominated abstract presentations under the Design Education and Pedagogy track, and topics related to Environmental Behavior (Topic 17) and Environmental Health (Topic 18) constituted the majority of submissions under the People–Environment Relationships track. However, for tracks such as Design Implementation, History, Theory, and Culture, Landscape Planning and Ecology, Research Methods, Sustainability, and Urban Design, submissions cover a much broader range of topics. The Research Methods track particularly shows balanced linkages to all topics, suggesting that methodological explorations and innovations have a significant role regardless of the topic area of LA research.
Research Trends over Time
To examine the trends of topic distribution over time, we calculated and plotted the change of topic coverage normalized by the total number of papers each year. Ideally, trends should be identified with a starting period and an ending period, each averaging across several years to remove the influence of irregularity that occurs in a single year. Because we only had seven years’ worth of data, we used the percentage increase and decrease from 2013 to 2019 and identified the topics that showed clearly increased or decreased trend and change in share of at least 1% (Figure 5). Some topics show an increasing trend. These topics include Climate Change (Topic 3), Technology and Digital Tools (Topic 12), Urbanism and Urban Design (Topic 14), and Community Engagement (Topic 19). The shares of Topic 19, Topic 12, and Topic 3 nearly doubled in the span analyzed. Certain topics displayed greater fluctuations than others, for example, International and Vernacular Landscapes (Topic 20), as the 2017 CELA Conference was hosted by Tsinghua University, Beijing Forestry University and Peking University, and thus attracted a greater amount of research with an international focus. Topics that show a decreased trend from 2013 include Landscape Performance (Topic 7), Material and Construction (Topic 8), Land Development and Reuse (Topic 9), and Sustainable Agriculture (Topic 10). However, among these topics, yearly variations and substantial periodic increases also occurred. For example, Topics 8 and 9 reached low points in 2017 but have since seen some growth. The other topics, such as Parks and Public Space (Topic 2) and Design Pedagogy and Curriculum (Topic 4), remained relatively consistent.
Trends of topics (2013–2019).
Topics Most Relevant to Each Track
Figure 6 displays the example top topics based on submitted abstract content for each track. Most of the topics appear to be normative and easily relatable to each track based on the content necessary for each track’s description. For example, the Communication and Visualization track tends to concentrate on representation, technology, and digital/virtual tools. Rather than discuss topics easily relatable to each track, this section concentrates on trends by track that appear to be outliers. This will help us better identify anomalies in current submission trends to inform future CELA track directions.
Topics by track.
Interpretation of the results reveals four key findings related to similar content and primary topics that appear to be routinely submitted across multiple tracks. These findings are: (1) topics related to water and hydrology (e.g., low-impact development, stormwater management) are scattered across five different tracks; (2) social-related topics (e.g., environmental justice, diversity) overlap four tracks; (3) climate change–related topics (e.g., temperature, heat, sea level rise) are embedded in four tracks; and (4) urban (across three tracks) and regeneration (across three tracks) make up a total of six different tracks. Thus, there is much overlap in presented content topics across tracks. For example, performance assessment and environment–user analysis appeared as repeated topics not just under Landscape Performance or People–Environment Relationship but also under the Research Methods track, suggesting that it may be confusing to authors whether to submit based on the thematic area of the research or the methodology. In a similar way, sustainability-related topics are present across multiple CELA tracks, not just the Sustainability track; this suggests that the sustainability topic may be too broad to be a single track. These findings suggest that there may be a need to reconsider the CELA track structure slightly to optimize the track–topic relationships, reduce confusion, and create new tracks that better focus on contemporary research trends, recognizing that there may be cases where the current tracks are still appropriate.
DISCUSSION AND CONCLUSION
This study evaluated tendencies in CELA abstract submissions between 2013 and 2019 to determine which research topics are most relevant in current LA-based research and evaluate how these research trends have changed. The findings were used to compare current topics in LA research to the current CELA track structure. Results show that the primary LA-based research topics (compared with existing content analyses of the same subject) have some continual core topics but tend to vary through time.
This time-based fluctuation indicates that CELA authors may not always be choosing the most appropriate research track based on their abstract contents, either by track selection error or by missing tracks/foci from the CELA. Our findings sorted the top LA research topics based on topic probability and these are ranked from most to least probable in Figure 7 (Current Study column). Among topics with high probabilities, Climate Change (Topic 3), Technology and Digital Tools (Topic 12), Urbanism and Urban Design (Topic 14), and Community Engagement (Topic 19) are gaining popularity over the seven-year period.
A comparison of CELA track structure, topics identified 1981-2005 (Powers & Walker, 2009), and 2013-2019 (this study).
These topics can be used by other LA-based conference venues to help reshape their organizational structures, or the methodology can be replicated and applied to each venue. We recommend that CELA immediately implement track content related to Climate Change, Hydrology, and Urban Regeneration. We suggest that Climate Change can become its own unique track, while Hydrology can be emphasized in the Landscape Planning and Ecology track description, as this is where the majority of hydrology-related research is currently submitted. Further, Urban Regeneration should be included as a primary research topic in the Urban Design track. It may prove beneficial to list a set of categorized topics under each track to better specify what content should be submitted to each particular track. We also suggest that (1) CELA conduct such a longitudinal analysis every five years to stay current and allow the track system to adapt to needs as they surface and research topics change, and (2) other LA-based conference venues do the same. This will allow such findings to be compared across venues to develop more comprehensive topic lists and identify specializations across organizations. Once a body of data is developed over time, longitudinal studies on the assessments can be conducted to better identify topic changes through time.
To form a meaningful comparison with the CELA track structure, we grouped our 20 topics based on the CELA track structure (Figure 7). Because Powers and Walker (2009) used a framework derived from the original CELA tracks, we pooled in the their topic shares to help form a quantitative comparison. The topic ranks from 1981 to 2005 showed History, Theory, and Culture; Landscape Planning and Ecology; and People–Environment Relationships as the top domains. In 2013–2019, the quantity of History, Theory, and Culture–related submissions have decreased. Tracks such as Landscape Planning and Ecology, People–Environment Relationships, Design Implementation, and Communication and Visualization have experienced less than 10% change. However, many of these topics now encompass more specialized branches of knowledge, demonstrating the topic diversification process that was pointed out in the literature (Vicenzotti et al., 2016).
Consistent with the conclusions of Gobster et al. (2010), our findings suggest that attention should be given to emerging topics related to Climate Change, Hydrology, and Health as individual research domains. Design Education and Pedagogy emerged among the most popular topics, indicating a resurgent interest in researching the process, affirming the effectiveness of curriculum and student learning (Newman et al., 2016, 2018), and reflecting CELA’s role in promoting LA education. Topics such as Community Engagement and Technology and Digital Tools represent new research domains that are highlighted in recent abstracts. Three other categories— Methods of Inquiry, Sustainability, and LA as a Profession—did not emerge as recent research topics, as the definitions of these terms have evolved and they often serve as overarching concepts. For example, based on the 2030 Agenda for Sustainable Development Goals, sustainability is considered a broad umbrella term, covering the majority of the 20 topics identified. Relatedly, Gobster, Nassauer, and Nadenicek (2010) showed a significant need to broaden the scope of Landscape Journal publications and improve submissions to the weaker-performing areas of the journal. Broadening the scope of CELA will allow for submitted papers covering less typical topics related to LA to be peer reviewed and published in Landscape Research Record. These papers may be more difficult to include in Landscape Journal due to their scope or fit with the journal’s goals. Previous research has shown increased numbers of LA faculty publications but a diminished number of outlets to publish unique or exploratory research (Deming & Swaffield, 2011). Therefore, it makes some sense to omit less popular tracks and create new tracks to balance the number of abstracts.
Similar trends are also reflected in the CELA submission numbers, which show that the Landscape Planning and Ecology, Design Education and Pedagogy, and People–Environment Relationships tracks are increasing. However, the Design Implementation and Sustainability tracks are decreasing considerably, as specific topics under their broader umbrella appear more frequently under other tracks. While all other CELA tracks remain relatively stable and cover most of the identified increasing topics listed above, Climate Change, Hydrology, and Land Reuse do not yet have specific CELA tracks but routinely occur across submitted track content and in historic literature. Research on Social Issues is also submitted across multiple tracks. Simultaneously, there is much overlap in the content of CELA submissions across tracks, which may oversaturate content across multiple tracks, which can result in fewer submissions to some specific tracks. The addition of the Geospatial and Digital Analytics and the Landscape Architecture for Health tracks created for the 2020 CELA conference seem to fit the current trends, as digital tools represent a fast-growing topic. Also, the existing People–Environment Relationship track clearly diversifies into a behavior and perception-related direction and a health-related direction. Finally, the 2021 track created on Diversity, Equity, and Inclusion fills responds to the critical need for research related to social issues. While research related to Diversity, Equity and Inclusion can cross all tracks, the creation of a specific stand-alone track related to such issues is as an important step in highlighting its importance.
Following the suggestions presented by this research could strengthen CELA’s role at the frontier of LA research and education, catalyze cutting-edge research, encourage scholarship that discusses pressing societal issues, and allow for more focused sessions at future events. Although minor changes can continue to be made based on these findings, we suggest that major changes only take place after another time period analysis is complete. More research should be conducted to determine how well such topics align with the research needs of the LA profession. The LDA method can continue to provide insights into the natural clustering of topics and may be used to assist in the process of populating the sessions of CELA conferences to ensure papers most relevant to each other are presented in the same session.
AUTHOR CONTRIBUTION
The authors are solely responsible for developing and presenting all aspects of this article.
PEER REVIEW STATEMENT
This article was subjected to a blind review by at least two external peer reviewers. The editors gratefully acknowledge the reviewers’comments on the article.
