Envisioning future landscapes: A data-based visualization model for ecosystems under alternative management scenarios

Highlights

• Ecological data were integrated into 3-D visualization of future landscapes.

• Long-term visual impacts of wildfire, grazing and pine colonization were predicted.

• Mediterranean landscapes were reliably represented using only 4% of the flora.

• The visualization was found to be a statistically valid representation of reality.

• The 3-D model is science-based, integrative and accessible to non-experts.

Abstract

Human-driven landscape changes strongly influence landscape functionality and aesthetics. While landscape planners have access to biophysical data for decision-making, they often do not have the necessary information about social variables, such as aesthetic tastes, feelings, or functions of a place. Visualizing future landscapes under alternative management scenarios could be a valuable tool for aiding land management decisions. Towards these ends, empirical, quantitative ecological data on vegetation composition, pattern, and processes in a Long-Term Ecological Research (LTER) site in Israel were integrated into a computerized, 3-D representation of current and future landscapes.

Our objectives were (1) to visualize landscape-shaping processes, such as wildfire, grazing, and species colonization, that can assist managers, planners, and the public to envision the long-term visual significance of management alternatives and (2) to validate the similarity between the 3-D model and reality. The visual model we developed is based on 30 years of scientific knowledge and ecological data describing vegetation processes in Ramat Hanadiv, a case study of ecological conditions and processes relevant to the Mediterranean and other complex ecosystems worldwide.

Before studying the role of the 3-D model in decision-making, validation was performed by comparing ‘current state’ model representation with real-world photos from the perspective of the observer. The model was found to be a valid representation of reality.

Looking to the future, we suggest that the ability to create future landscapes using scientific data can assist to improve decision-making processes, balancing ecological and social needs.

Introduction

Communicating scientific data to non-experts is a major challenge. This challenge is intensified by the scale and scope of the data in question; we see the increasing collection and storage of big data, in which large databases are built, stored, analyzed, and shared and, potentially, could play a significant role as available information for use by decision-makers. However, this data is inaccessible and thus underused. In ecology, for example, the Long-Term Ecological Research (LTER) Network established in 1980 (similar to NEON in the United States) contains platforms encouraging and enabling the creation of quantitative datasets that describe global environmental change and its effects on ecosystems throughout the world (Mirtl et al., 2018). These databases usually include detailed metadata allowing their use in future syntheses and comparisons.

In the face of its increasing ubiquitousness, some authors argue that collecting vast amounts of environmental data not led by questions and hypotheses threatens the principles of evidence‐based science that supports management and policy (Lindenmayer and Likens, 2018, Collins and Knapp, 2019). Moreover, GIS layers or raster databases are in most cases inaccessible to non-experts like decision-makers and the public that need a more recognizable landscape language (Nassauer, 1995).

Many landscape ecologists aim to use their discipline to guide a management approach that integrates ecological knowledge and social considerations (Liu & Taylor, 2002). However, alongside a movement towards more applied research in this field, as reflected in the professional literature (Wu, 2017) many studies still focus on the spatial distribution of biological elements or make rather little use of modeling approaches accounting for actual ecological processes (Morán-Ordóñez et al., 2019), both do not suffice as a basis for management and planning.

Active management is one of the more contemporary approaches to managing open spaces (Perevolotsky & Shkedy, 2013). In contrast to a more traditional “hands-off” approach, active management advocates intervention in ecological processes to facilitate the provision of multiple benefits from the ecosystem. Such an approach is crucial for addressing a major challenge of managing landscapes that have evolved under frequent anthropogenic disturbances – controlling shrub encroachment and regulating woody vegetation cover and biomass (FAO & Plan Bleu report, 2018). Another concept, the adaptive management strategy (Haney & Power, 1996), is one of the pillars of the Long-Term Ecological Research Network (LTER, Mirtl et al., 2018), and is predicated on the idea that good scientific information will reduce uncertainty and inform future practices through a process that links management experimentation, hypothesis testing and observation of ecosystem responses (Bakker et al., 2018). However, this approach has been criticized for its limited actual application and adaptation to different conditions and scales (e.g., Williams, 2011, Tony, 2020).

Despite the availability of good scientific data and high public trust in science in general (Wellcome report, 2018), most decisions regarding the management of natural landscapes are challenged by additional considerations, including trade-offs, aesthetic values, and cost-benefit analyses subject to individual and group interpretation, often influenced by underlying values and perceptions, e.g., what is the desired landscape and who should decide what is desired? How is the landscape perceived by people from diverse backgrounds and what are its important visual qualities, aesthetics, and functionality vis-à-vis diverse human uses?

By combining empirical scientific data and subjective considerations regarding diverse priorities, desires, and perceptions, landscape management has the task of merging what ecosystem services an area can provide, what people want, and how the area can be designed and managed to achieve what people want (Oliver et al., 2013). To meet the goal of integrating these components, both scholars and practitioners of land management must develop, test, and apply decision-support tools that can merge the public's needs with the area features towards management strategies that are both ecologically sound and socially acceptable (Robinson et al., 2019).

Our knowledge of forest ecosystems and management impacts has expanded significantly in recent years (Schweier et al., 2018, Leal et al., 2019). This expanded understanding is expected to support decision-making according to the sustainable forest management approach (Osem et al., 2008, Machar, 2020). However, forestry is a highly complex field that needs to integrate information from different disciplines and make it comprehensible to people of different backgrounds (Meitner et al., 2005, Kaspar et al., 2018). One underlying problem is that alongside the spatio-temporal and biophysical complexity of forest ecosystems, there is a high degree of social complexity with diverse goals, values, and visions of future forests.

Many studies have examined the impact of landscape management on ecosystem functionality and diversity (e.g., Turner, 1989), but far fewer have addressed the actual impact of such management on aesthetic preferences (although see Gundersen et al., 2017; and a review by Gobster, 1999). Forest and landscape management operations such as clear cutting and thinning or removal of undergrowth have impacts on the aesthetics and the amenity value of the landscape. People have diverse and often strong opinions regarding such management, particularly near urban areas (Depietri & Orenstein, 2020), and land managers need to deal not only with changes in the landscape, but also with changes in the public's perceptions of the landscape (Ryan, 2005, Depietri and Orenstein, 2020) and their implications on the acceptability of different management plans.

These challenges require developing new decision-support tools that can make empirical data and scientific knowledge more accessible and relevant to stakeholder-driven processes and balance the tradeoffs between ecological and social management goals (Kaspar et al., 2018). This can be done, for example, by the illustrative demonstration of the consequences of different decisions in an interactive process between researchers and stakeholders (Haberl et al., 2006, Grêt-Regamey et al., 2013, Bennett et al., 2017).

In this paper we present a visualization tool developed specifically for communicating ecological data and scientific knowledge on complex ecosystems to various audiences.

The need for effective communication in landscape management and planning has resulted in a considerable increase in the use of two and 3D visualizations (Edler, Kühne, & Jenal, 2020, Lewis et al., 2012). Many authors have described and reviewed human reliance on visual information to process information and distinguish between different situations (e.g., Bruce et al., 2003, Sheppard, 2001). Visualizations are considered a common language that uses our innate abilities to understand visual information.

Computerized, evidence-based visualization models can integrate social, economic, and ecological parameters and enable interdisciplinary analyses. 2D visualizations, although commonly used in fields like forestry (as maps, GIS layers, or spatial modeling outputs), are often too abstract and cannot fully represent landscape complexity and aesthetic qualities. Photographs may provide a valid representation of current landscape conditions, but their inability to represent future or hypothetical conditions limits their utility in public participation contexts on forestry issues (Lange, 2001, Meitner et al., 2005). To compensate for the shortcomings of other visual data, 3D models can extrapolate upon data from plot-level monitoring with a vast number of information layers and can be used as an empirical basis for constructing visual representations of future scenarios at large scales. Dynamic 3D visualizations have the power and flexibility to present alternative future landscapes side-by-side, within the same setting, and over time, and therefore offer a powerful comparative tool to engage people in environmental issues and problem-solving. Such models have been used, for example, to visually present the possible consequences of climate change, thereby educating stakeholders, raising community awareness, and setting a common ground (i.e., boundary object) between diverse demographic groups. This visual presentation thus catalyzes stakeholder-informed policy formulation (Schroth et al., 2015; Sheppard, 2012).

Advances in computer processing power and graphic software have substantially improved the precision and accuracy of environmental visualizations (Downes and Lange, 2015, Edler, Kühne, & Jenal, 2020). Further, electronic communications and computer networks enable efficient and economical distribution of visualizations to expanding audiences. Consequently, the use of visualization in landscape assessment research and practice is gradually increasing (Lovett et al., 2015). Yet, alongside their benefits, visualizations also pose challenges for both the modelers and the users (Deussen et al., 1998, Sheppard, 2001, Nassauer, 2015).

One basic assumption behind the use of visualizations is that they reflect valid representations based on accurate perceptions and sound judgments made in response to direct experience with the landscape (Daniel and Meitner, 2001, Downes and Lange, 2015). However, the need for abstraction and simplification may conflict with the desire to produce a highly realistic visualization. There is an open and ongoing discussion about what should be considered a valid representation of the landscape and what level of realism is sufficient for engaging the public (Lange, 2001, Appleton and Lovett, 2003, Billger et al., 2016). Some researchers argue in favor of maximizing realism. Highly realistic visualizations of forest landscapes were found to be more valid (Daniel and Meitner, 2001, Lange, 2001, Ribe et al., 2018) and to improve communication, while simplified representations were harder to communicate even to experts without the addition of verbal information (Barrett et al., 2007).

Facing these challenges, our goal was to develop a valid tool based on ecological data and in-depth scientific research which can be used for the communication of complex landscapes to various audiences. The visual products (images, panoramic tools, and short films) will later be tested in stakeholder workshops to assess their efficacy in communicating to decision makers and the public future landscape possibilities and the science and management strategies that may shape those futures.

The Mediterranean landscape provides an excellent opportunity to explore this approach. Lacking a “natural landscape” archetype, Mediterranean landscapes have been described as multi-scale mosaics of different vegetation types and structures, associated with high resilience and rich ecological diversity, co-evolving with social systems through an ongoing history of human intervention (Blondel, 2006). In the face of increasing human pressures on natural ecosystems and their high biodiversity, such complexity is important from a conservation perspective (Myers et al., 2000), and requires the establishment of management strategies at the landscape scale (Scarascia-Mugnozza et al., 2000). We believe that our findings will also be relevant to a wide range of dynamic and highly complex ecosystems such as tropical forests, managed commercial forests, forest-savanna transition zones, and more.

Given the high complexity of these landscapes, our main challenge was to find the optimal balance between abstraction and realism and to identify the minimal set of landscape variables that will provide a valid representation of an extremely diverse plant community in the eyes of the beholder.

A significant portion of the literature reviewing visualization deals with improving communication of environmental data by combining different data sources or translating numbers into symbolic or figurative representation or images (Edler, Kühne, & Jenal, 2020, Metze, 2020) . Visualizations are often used to illustrate the visual impact of adding elements such as wind turbines or solar panels to the landscape (Maehr et al., 2015, Ribe et al., 2018), or to envision possible large-scale impacts of climate change (Schroth et al., 2015, Sheppard, 2012). Yet, very few of these visualizations express the science of dynamic ecosystem processes, such as grazing or fire that have complex effects on ecosystems.

In this study, we present a state-of-the-art 3D computerized landscape model and assess the quality of visualizations produced by the model and their potential relevance for management decision-making. The model is based on long-term quantitative ecological data, expert knowledge, and findings from in-depth scientific research. These sources are integrated to visualize the predicted appearance of future landscapes under alternative management scenarios.

Two questions were posed in this study: (1) how can quantitative scientific data describing vegetation composition, structure, and spatial pattern, be translated into a three-dimensional computerized visual model of current and future landscapes? And (2) is the model a valid representation of reality? i.e., does the visualization reflect the same perceptions and judgments that would have been made in response to direct experience with the landscape?

Our overall objective is to develop and validate our 3-D model, both regarding its degree of perceived visual accuracy (the current study), and its utility in stakeholder-driven management processes (follow-on research).