Arctic Changes: Where my polar bear go?

1 Introduction

Given the drastic, rapid, and concurrent changes in the high latitudes and their impacts on global processes and peoples of the North, the Arctic represents a complex system that warrants urgent integration of research efforts. The necessity for integrative approaches addressing cumulative and compound effects of multiple drivers of changes has been highlighted in recent reports (AMAP, 2021; Arctic Council, 2016) emphasizing the need for the Arctic research community to step away from the relative comfort of disciplinary silos and move toward the development of holistic system views and novel paradigms.

Arctic research demands convergence science.

How can convergence science be construed? In contrast to the plain meaning of its etymon (the Latin convergere) to “incline together,” convergence research as a novel type of scientific endeavor encompasses not just passive integration of knowledge or a cascade of boundary conditions from one disciplinary group to another. On the contrary, it calls for the identification of fruitful research areas of opportunity to foster the emergence of new views, scientific principles, and even disciplines—the process in which diverse participants operate with a common language and reference points (Sharp & Langer, 2011; Thompson et al., 2023). In theory, a true application of the convergence science approach, individual disciplines and traditional concepts intersect, fuse, and cross-pollinate to gain novel insights and to understand emergent complexity, while accelerating solutions to big, complex problems (Stokols et al., 2008).

How can convergence science approach be implemented, given the various cognitive and social barriers (Wagner et al., 2011) associated with the number of and relative separation among the disciplines? In practice, convergence science enhances research through interdisciplinary teams of scientists and stakeholders working together to push the scope of scientific inquiry beyond the typical boundaries of their respective fields, to foster mutual learning and novel collaborations, and develop a transdisciplinary language and knowledge consolidation to solve specific problems and respond to demands from society. For the purposes of this paper, we use the definition of transdisciplinary research offered by Rosenfield (1992) (“researchers work jointly using shared conceptual framework drawing together disciplinary-specific theories, concepts, and approaches to address common problem”) and view it as a required element of convergence science (Thompson et al., 2023) that focuses on big, complex socially-relevant problems. The implementation is not without challenges and tensions: identifying disciplines needed to address complex system-level problems, selecting compelling questions that will emerge as research foci, and integration and application of diverse methodologies require “engineering of communication” among experts and distillation of integrative perspectives.

The central objective of this paper is to illustrate finely grained objectives and processes of convergence science, moving away from the “generalist,” broad contemplation level, to the application of the concept to specific Arctic stressors and mechanisms they imply. We seek to showcase how this approach allows the identification of linkages critical to the development of system-level understanding of stressor impact propagation. In the process of developing that understanding, we uncover knowledge gaps falling within the scope of both interdisciplinary and discipline-specific research. Concurrently, this paper also aims to demonstrate how a diverse group of author-scientists, who were trained largely within the niche of their single discipline, can through integrative thinking advance questions and understandings in ways that cannot be achieved with studies in their “host” disciplines alone.

To provide an intuitive application of the concept, we use the Yamal peninsula in the Western Siberia (Russia) as a case study region for this synthesis, as evidence indicates increasingly intertwined processes caused by multiple stressors on the abiotic, biotic, and socioeconomic systems over the past four decades. Combining responses to these multiple drivers of change, Yamal is a vivid illustration of the need for convergent scientific understanding of Arctic change. Three representative convergence science threads are developed in this paper.

2 Facets of Convergence Science of Arctic Change

The Arctic is subject to several types of novel stressors that evince multiple levels of interaction with its environment and inhabitants. Here, we focus on two specific stressors that likely incorporate the larger fraction of unknowns and concerns across scientific and stakeholder groups and therefore constitute immediate research needs: climate change and industrialization. We explore how convergence science can trace the effects of stressor impacts across systems and may support genuine synthesis and shared understanding across disciplines.

3 Arctic Microcosm—Yamal, Western Siberia

We regard the Yamal Peninsula as a natural laboratory with Pan-Arctic explanatory relevance, because of the concentration of a large variety of pertinent components in the earth and social-ecological systems. These characteristics of the Arctic environment include various types of the permafrost, abundance of water, sharp changes in vegetation forms, snow and ice seasonalities, migrating animals (reindeer and birds), the high sensitivity of natural systems to climate change as well as the presence of relevant socio-economic-cultural aspects such as strong indigenous culture and livelihood, industrial development, a large non-indigenous population, and affluence of the region because of the natural resource extraction industries. These same features can be found in other regions of the Arctic such as Alaska, Arctic Canada and Fennoscandia. However, only in Yamal do they occur in such a high density, which makes this region exemplary and suitable for the development of convergence science frameworks.

4 Convergence Science Method

The research questions considered here are examples of multi-disciplinary convergence science questions, which resulted from a multi-stage brainstorming process undertaken by the author group. We pose it as an example of “engineering” an approach to facilitate effective and productive communications and generate a convergence science approach to research. This method, while not without its shortcomings, illustrates the value of an explicit structure for interactions within a heterogeneous expert group to yield integrative thinking—a prerequisite for the development of convergence science. We believe this method may be applicable to any group seeking to generate broad, important questions that rest on pillars of disciplinary knowledge. We describe it below, starting with an outline of our strategically structured workshop and “rules” of discussions that led to drafting convergence science questions, some of which are presented in this article.

In March of 2020, the rapidly evolving COVID-19 epidemic resulted in travel restrictions that disrupted plans of our team for an in-person meeting. We thus organized an online workshop. We set the goal of identifying high impact convergence science questions that spanned the breadth of disciplines represented by the group in social, natural, and built-environment systems. Specifically, we consisted of social scientists (4), engineers (4), earth system scientists (7), and ecologists (12) from across Russia (11), Europe (4), Middle East (1), and the US (11) and one logistical issue limiting our day-to-day interactions was that the team members were separated by as many as 11 time zones. Beyond discipline and space boundaries, we were scholars of various cultural identities and, described broadly, the team consisted of North Americans of European, Russian, and Asian descent, Western Europeans, Asians, and Russians. While the workshop team did not include non-scholar participants, several co-authors had conducted long-term studies (for more than a decade) in the communities of Yamal and developed broad social networks that included indigenous reindeer herders, public organizations of indigenous peoples of Yamal, leaders of reindeer herding communities as well as various Yamal government departments. These stakeholder groups therefore furnished information input for the exemplary research questions of this manuscript—as mediated by our co-author experts who provided the necessary competence for translating the knowledge and needs of the non-scholar stakeholders into a culturally appropriate discourse during the convergence science process. The topics encompassed climate change, reindeer herding management, burgeoning industrial development, and others.

Our key challenge was to find the grounds for science questions that emerge far beyond the disciplinary expertise of any single group member or subset of the larger group. This required systematic efforts that would push workshop discussions out of disciplinary silos and concentrate them on the development of views and questions in which no single expert group could claim dominant expertise. As a result, we structured the workshop to start within our disciplinary comfort zones, then to gradually integrate disciplines in a hierarchical manner, to culminate in a drawn diagram of connections between elements of the Arctic system from which we could distill high-impact convergence science questions (Figure 1). Our goal at each phase of the workshop was to identify critical elements within multiple disciplines (such as reindeer, arctic fox, herders, permafrost, shrubs, etc.), and then determine connections among these elements defined by explicit mechanisms—even if they remained elusive or unknown due to missing expertise in the team. The focus on elements connected by mechanisms kept all discussions grounded in practical rather than abstract terms. This facilitated discussions in which element-mechanism interconnections would be converted into concrete research objectives driven by testable science hypotheses and questions. The workshop participants also regulated discussions to prevent their swaying toward non-actionable science (i.e., too remote from team's expertise or too uncertain due to current inability to observe or measure relevant processes). This prevention emerged as a social norm during workshop discussions by the participants, rather than imposed as a “rule of the conduct” (which would have probably limited brainstorming and original thinking by individuals to some extent).

We began with recorded 15-min expert presentations (Figure 1, “EP” elements at the base of the workshop pyramid) by select members to present key disciplinary questions to non-experts of the discipline (i.e., representatives of other disciplines). Presentations were grouped into four “themes” combining similar fields: Theme I—“Arctic climate, weather, hydrology, permafrost, landscape vegetation,” Theme II—“Herbivory, predator-prey interactions, birds,” Theme III—“Plants, productivity, nutrients, dendrochronology, paleo-botany, ontology, ecology,” and Theme IV—“Social anthropology: human-environment interactions, reindeer herding in Yamal.” These expert summaries were foundational for building a collective language of communication among the participant scientists. Overall, we had 11 expert presentations within the context of the four Themes. Indeed, the act of summarizing key disciplinary questions in a language that non-experts could understand forced each expert to break down those questions to their most fundamental and important elements. This allowed each to step into the shoes of the other participants–to consider the question, “why is this important to others?” That in itself turned out to be an important key to our convergence science approach, one challenged to integrate ideas from vastly different disciplines.

The first challenge for each participant was to identify one mechanistic linkage between two elements of high scientific interest drawn from different presentations within each theme. By initially bridging related disciplines, these linkages formed the foundation for the gradual building of cross-disciplinary networks. Two such example linkages could be: shrubs (element 1) elevate winter soil temperature (element 2) by trapping snow (mechanism); linear infrastructure (element 1) increases fox abundance (element 2) via human-discarded food subsidies (mechanism). It was critical to keep linkages transparent and formulaic, avoiding abstraction or grouping of concepts. Thereby, when linkages were later connected into networks crossing disciplines, they could represent manageably sized research questions containing measurable elements and mechanisms.

Within the subsequent breakout groups corresponding to geographic regions of the team (Figure 1, “B” elements), each group discussed the individual linkage choices, and attempted to modify and distill two top linkages that crossed disciplines (within themes) and appeared to present opportunities for novel science. For example, while the effect of shrubs on snow retention has been well studied in the plant and biophysical sciences, the effect of shrub distributions (element 1) on ptarmigan population structure (element 2) via landscape snow redistribution (mechanism) is more likely approaching novel scientific territory by bridging related biological and biophysical disciplines. Most groups also opted to create thematic network diagrams to gain a more holistic understanding of the emerging science in preparation for later integration across themes. In a desire to avoid established questions of disciplinary interest and facilitate our forging into novel convergence science territory—in which no group member could claim expertise—we chose to have each breakout session led by a non-expert of the theme, while thematic experts were assigned rapporteur roles and were directed to express their opinions last.

Our next challenge was to consolidate the regional consensuses on linkages that were identified to have potentially novel science. We worked through two full-group meetings (Figure 1, “G” elements) to debate and edit the 12 linkages from two Themes identified by the breakout groups, and ultimately integrate them into two “spaghetti diagrams” (an example of one is in Figure 2a). This step required the identification of missing links, especially between science themes, thereby expanding the emerging convergence science pathways. During this integration process, the team also contextualized the developed spaghetti diagrams from a regional perspective: we explicitly considered the question of what makes Yamal a scientifically appealing place to study questions of interest (Figure S1 in Supporting Information S1). At this stage we also filtered out what was viewed as apparently non-actionable science (still, this was done in a conservative fashion to minimize a disregard for high-risk, high-yield research areas).

After synthesizing pairs of Themes, we divided again into regional breakout groups tasked to “distill” the two spaghetti diagrams (Figure 1, “D” elements). Distillation entailed applying “Occam's razor” to diagrams overpopulated with elements, and highlighting key linkages leading to apparently novel convergence science (Figure 2a). We again made a concerted effort to avoid the tendency for abstraction, and maintain networks made of observable elements connected by explicitly hypothesized mechanisms. At this stage we began to formulate higher-level science questions that could be informed by an integrated study across a trans-disciplinary chain of elements.

Our final challenge via full-group discussions was to integrate the distilled spaghetti diagrams into a single unified network diagram and identify high-impact and actionable convergence science questions that arise from that network (Figure 2b)—a process that we referred to as “The Great Spaghetti Cook-Off” (Figure 1, the top of the workshop pyramid). Prior to this meeting, the distilled spaghetti diagrams were integrated into a single network containing all of the elements and their connecting arrows (as presented in group distillations), with elements color-coded by themes, ready for live editing during the full-group meeting. First, the full group debated and refined the network structure and labels. Then, we divided into three sub-groups, each of which was tasked over 45 min to develop two science Threads—a sub-network of five elements connected by mechanistic linkages within the integrated spaghetti diagram. Each Thread was to represent elements and processes of high scientific interest, with perceived strong mechanistic interactions between elements from different disciplines. We then reconvened as a full group to discuss and refine the Threads, ensure their mutual distinction, and critically re-evaluate the elements and mechanisms that had not been assigned to any Thread. Our final product was a cohesive network diagram with six color-coded Threads (Figure 2b). At the workshop closure, by tracing mechanistic pathways in each of the Threads, our team thus drafted tractable convergence-science questions of highest interest that were best informed by cohesive inputs from multiple disciplinary studies. We felt that these questions, by the nature of their construction through our workshop design, necessarily forged into novel scientific territory, successfully producing convergence science objectives.

Our convergence science questions were further refined in post-workshop activities led by team sub-groups, whose composition continued to represent diverse areas of expertise. For each Thread and its corresponding science question, the sub-groups were tasked to refine and detail hypothesized conceptual models of mechanistic pathways connecting Arctic stressors to their direct impacts. They were also called to consider effects exerted on elements of natural, social, and built-environment systems, as well as triggered responses and adaptive strategies. Three examples of such convergence science questions are provided in the next section.

Overall, the described method proved successful as a foundation for the convergence science questions presented here and for a subsequently funded grant proposal (through the National Science Foundation “Navigating the New Arctic” program). In addition, the group members found the process highly stimulating, enhancing cross-cultural and cross-disciplinary connectivity, broadening each of our knowledge bases, and improving understanding of how each of our research foci fits into a broader picture of Arctic processes.

5 Convergence Science Threads

During a workshop in 2020 and subsequent synthesis activities, the authors co-developed several convergence science questions whose causal mechanisms were perceived to be of high priority to understand. The objective for the selected three research threads below is to illustrate how a given question relating a certain stressor (i.e., climate change and industrialization) and an Arctic system agent(s) (e.g., reindeer, herders, tall shrubs) can be addressed via the development of a holistic view of system interactions and their spatial and temporal scales. Specifically, we illustrate how the integration of disciplinary knowledge of processes and the relevant linkages can lead to conceptual models of interactions in a network of interlinked elements. Such models can then serve to identify specific causal connections that can be addressed in research to further our understanding of overarching mechanisms. We also formulate Emerging Questions (denoted as EQ) that are important for the overarching research thread question and currently represent knowledge gaps.

The growing presence of an indigenous population (with their reindeer) and newcomers (with their infrastructure) has complex impacts on local animal species (Figure 3). For example, food subsidies grow following an increased number of reindeer (Ehrich et al., 2017; Sokolov, Sokolova, et al., 2016). Additionally, more people in tundra potentially means an increase in food waste and anthropogenic food subsidies. On the other hand, domestic dogs can prevent endemic mammalian scavengers from accessing subsidies in settlements of hydrocarbon extraction fields.

Infrastructure can reduce and fragment natural habitats, but at the same time can create new habitats for some species. This may benefit generalist predators, such as corvids and foxes and alter predator-prey relationships, potentially leading to an increased predation pressure on wild prey species, such as ground-nesting birds and rodents. This can also lead to a discussion of the context of arctic fox in relation to reindeer (Terekhina et al., 2021). Effects are further complicated by climate-related phenomena, such as prolonged winters (late springs) leading to delay in arrival of migratory birds, another important fox food source. This highlights the complexity of relationships between infrastructure and increased human presence on the one hand, and wildlife on the other.

The Obskaya-Bovanenkovo railroad (Figure S1b in Supporting Information S1) presents an example of how infrastructure development impacts wildlife on Yamal (Sokolov et al., 2017). The railroad extends from the forest tundra in the south to the high Arctic in the north and crosses numerous rivers. Bridges associated with the railroad have allowed the expansion of new species into the Arctic zone. Ravens and gyrfalcons did not breed in Yamal outside of forested areas and rocky cliffs prior to 2011, when the railroad was constructed. Bridges associated with the railroad provide elevated nesting sites previously unavailable to ravens. Gyrfalcons followed the ravens northward, utilized their nests, and flourished. This led to the first documented gyrfalcon breeding in Yamal, where the flat landscape does not provide sufficiently high natural rock cliffs or tall trees for this species to breed. Expansion of this top predator along the Yamal railroad potentially has an impact on prey populations too, especially ptarmigans. The raven population, which preys on nests of numerous birds, could likewise have a negative effect on avian species, such as grouse and waders (Henden et al., 2021; Rød-Eriksen et al., 2020).

EQ (Emerging Question): How does human activity modify top-down control of trophic interactions in tundra food webs? How does climate change impact the effects of anthropogenic disturbances on ecosystem dynamics?

Changing seasonality of winters affects both reindeer herding and systems that provide services to communities in the region. We hypothesize that warm spells throughout winter make mobility and transportation on Yamal more difficult. The start of the winter season with snow and ice determines the pace of reindeer herders' migration and camp movement to meat processing facilities where herders get their main yearly income. These migratory patterns, where timing is key, are such that some rivers need to be solidly frozen to be crossed with camps and herds. Research question 2 explores how warmer winters and seasonal shifts affect both reindeer herding and mechanized transportation (Figure 4).

Unpredictable winter seasonality results in bottlenecks for larger households during autumn and spring Nenets migration schedules. These delays create herbivory pressure along transit routes on both sides of the major water bodies, where nomads are “stuck” waiting for the ice. This, in turn, can lead to conflicts. Smaller private herders that rely on these pastures for their winter grazing close to villages (Yar-Sale, Ports-Yakha, Salemal, Panaevsk) complain that big herds that are supposed to go to the other side of the wide Ob’ River destroy their winter grazing areas. Now, more Nenets aim to leave the peninsula for winter, as they are not satisfied with tundra pastures during winter. This leads to a higher concentration of herds waiting on the banks to cross rivers, resulting in high pressure on pastures. EQ: What is the role of concentrated urine and defecation resulting from concentration of herds? What are the resulting changes in biomass of plant communities and how are they altered, potentially reducing, or improving overall pasture quality?

On the other hand, herders report that some of this grazing pressure will be “taken back by nature” (Russian: priroda voz'met svoyo), when large iced-covered patches of pastures caused by rain-on-snow (ROS) events become inaccessible for longer periods. EQ: To what extent does ROS allow the pastures in those areas to recover, by reducing the density of animals grazing and moderating the negative impact of trampling?

The most obvious disturbance for herding migration patterns and danger for nomadism as a livelihood as well as for herding as an economy is the impact of the increased frequency of ROS events and thaws (Serreze et al., 2021; Stammler & Ivanova, 2020). Recent publications have already shown these tremendous impacts (Forbes et al., 2016), and how herders respond to them (Golovnev, 2017; Stammler & Ivanova, 2020). These events do not necessarily increase competition among herders for scarce resources. A smart Nenets strategy relies on the animals' autonomous survival skills in the times of crises (Stépanoff et al., 2017). In Yamal, this is known as free grazing, where herders release their herds to roam freely, refrain from any herd control, and hope that the animals find pastures somewhere, even if it results in mixing with other herds. When a herd moves on their own, the owners can avoid direct confrontation with other herders by arguing that the herds are not driven by the people on purpose. EQ: Given increasing controversy about supplemental feeding in Fennoscandia (Horstkotte et al., 2021; Pekkarinen et al., 2021), how might a switch to intensive supplemental feeding change dependencies on state subsidies and affect reindeer health on Yamal? Would supplemental feed help vegetation recovery or reduce reindeer mortality? What combination of meteorological events leads to an ice crust critical for winter reindeer grazing?

Early winter road melting in spring means the danger of winter roads collapsing under big trucks that carry heavy loads. In extreme years, some winter roads do not open. For example, in 2019–20, the winter road from Salekhard, the district center, to Yar-Sale, the administrative center of the Yamal Peninsula, remained closed leading to increasing prices of all goods in the village. Reindeer herders responded to these pressures by purchasing more in Nadym, a town in the forest zone in the area of winter campsites and transporting goods on snowmobiles to the Peninsula. Some even took it as a business opportunity and hauled barrels of petrol from Nadym for sale in Yar-Sale. Not every family has the opportunity for such plastic responses to weather.

EQ: What is the effect of such events on equality of access to mobility and goods for the residents of remote villages lacking the formally established communications? How much warming will it take to cause a shift in the kind of transportation that is needed in winter along the Ob’ River?

Increased heat content in the atmosphere during snow-free summers in particular has significant impacts on reindeer due to the higher rate of biotic and human activities during this season (Figure 5). We hypothesize that the increased summer heat alters reindeer migratory patterns via impacts on landscape fragmentation by novel processes (shrubification, permafrost wasting, fire, and shifts in vegetation composition) that affect foraging, disease, insect harassment, and mobility. We additionally hypothesize that landscape constraints interact with socially negotiated access to migration routes of herding groups, leading to reduced flexibility in adaptation choices.

The direct implications of weather-scale change in the atmospheric heat content are well-characterized (Figure 5, “Reindeer heat stress”): reindeer have poor tolerance for high ambient temperatures and they avoid overheating through reduced metabolism and foraging only during night hours (Klokov & Mikhailov, 2019), often leading to animal weight loss. Furthermore, according to the Nenets, extreme summer heat leads to calves' lung disease (pers. comm.). Less apparent effects of heatwaves are due to the acceleration of frozen ground thaw (Figure 5, “Soil thaw; wet substrate”), which may create conditions of foraging in high wetness conditions. In combination with high temperatures and reduced mobility, this may force the herd to stay in trampled, boggy grazing pastures, that is, favorable conditions for hoof bacterial infections via scratches or wounds (Riseth et al., 2020). Heat waves are also associated with weather patterns during which low, calm winds can promote animal stress due to mosquito, warble flies, and nose bot flies harassment that inhibit reindeer foraging (“Low wind conditions”). Hagemoen and Reimers (2002) argued that the resultant decrease in feeding and resting and increase in demanding activities may compromise reindeer physical fitness, with possible consequences for winter survival. Excess summer heat therefore affects reindeer and nomads by reducing their mobility, with at least three negative effects: inhibiting efficient foraging to develop fat stores needed for winter survival (Nilssen et al., 1984); by increasing the spread of pathogens among dense, stationary reindeer; and by reducing the survival rate of calves in winter.

EQ: How will the continued increase in summer heat affect reindeer health, body conditions and their winter survival? Is there a particular migration or pasturing pattern that is more conducive to hoof infections (e.g., a rapid south-north migration on partially thawed surface vs. summer pasturing phase on warmer but wet soil)? Can these local, landscape-niche level effects of thaw feedback onto nomad migration patterns?

Furthermore, hot weather increases flammability of vegetation, making tundra (summer pastures) and boreal forest (winter pastures) more susceptible to ignition (Figure 5, “Tundra and forest fire”). Tundra fire has been observed to increase in the Arctic (Witze, 2020), likely due to the combination of lightning activity and frequent dry tundra conditions (He et al., 2022). Studies in Siberia are infrequent but indicate massive fires in the forest areas of mixed genesis, citing anthropogenic impacts (Moskovchenko et al., 2020). Fires can increase ecosystem productivity in tundra ecosystems, and may promote biodiversity (Heim et al., 2022) and active recruitment of shrub species (Myers-Smith et al., 2011) that can partly offset the fire depletion of biomass. However, fires also lead to the loss of slowly growing, energy-rich lichen—the preferred, if not dominant, winter nutritional element of reindeer (Turunen et al., 2009), highlighting specifically the vulnerability of winter reindeer pastures.

EQ: Can the loss of lichen due to the changing fire regime in the forest and tundra place additional pressures on the mobility of nomads and reindeer and social negotiations among herders, during both summer and winter periods? Can fire patchiness in the landscape impose violations of traditional reindeer foraging boundaries? Can the patchiness of fires impact the size of grazing herds as larger herds need larger pastures not impacted by fires?

Wind is one of the key determinants of mosquito relief (Skarin et al., 2004) and heat stress relief and higher coastal winds in Yamal are a strong enough feature of attraction for the migrating nomads. Coastal areas also have herbaceous plants of higher nutritious content that are forage for the reindeer. EQ: Will productivity of forbs in northern coastal areas increase with summer warming?

The effects of long-term, climate-scale changes in the atmospheric heat content are less understood, as they may initiate and convey processes that evolve over similarly long time scales (Ims, Jepsen, et al., 2013). Warmer summers, a deepening of the soil active layer due to the increased permafrost thaw, and the lengthening of the phenological period of photosynthesis have facilitated the growth of tall, woody shrub vegetation in tundra, though the main areas of growth cluster around water tracks and more severe active layer erosion (Elmendorf et al., 2012; Myers-Smith et al., 2011) (Figure 5, “Tall shrub expansion”). Studies in Fennoscandia showed that reindeer foraging can hold back shrubification of tundra (Horstkotte et al., 2017) but once sufficiently mature, shrubs may create localized impacts on herding practices. Herders typically avoid guiding reindeer through tall shrub thickets, so as to avoid reindeer feet and hoof injuries and excessive insect harassment, but they may also choose winter campsites in proximity to these thickets, which are a source of firewood and auxiliary subsistence material.

EQ: Can choices of pasture areas feed back onto the regional-scale vegetation cover by limiting shrub spread or mediating their distribution, for example, due to increased animal nitrogen enrichment via excreta? Overall, will the direct impacts of high ambient temperatures on reindeer and the indirect implications stemming from heat-induced changes in landscape lead to shifts in nomadic reindeer herding practices?

6 Discussion and Conclusions

The objective of this paper is to illustrate how convergence science—an approach that has seen increasing in interest over the past decade across a spectrum of disciplines (Thompson et al., 2023)—can be applied to the terrestrial Arctic system. While there has been general understanding that a convergence science approach bringing together earth systems scientists, social scientists, ecologists, and the affected stakeholders is needed, and science funding bodies over the world have recognized this and committed enormous resources to it (e.g., the US National Science Foundation's Navigating the New Arctic initiative), there are shockingly few publications on how to “do” convergence science, and none focused on studying the terrestrial Arctic system.

This paper presents a framework for breaking down disciplinary silos in bringing together social scientists, natural scientists, and engineers, and integrating the knowledge of non-scholar stakeholders. Our team, representing vastly different disciplines, countries, and cultural identities, came together to focus attention on climate change and industrialization and their impacts on the Yamal Peninsula of Arctic Siberia. This paper relays our process of design of convergence science questions and lessons learned during a hierarchically designed workshop in March 2020 that began with disciplinary perceptions of importance of Arctic system elements, and progressively integrated these views via specifically structured discussions into a unified network model. The latter served as a basis for the development of several convergence science questions, three of which were used as examples in this paper, aiming to showcase the identification of linkages critical to generating system-level understandings of stressor impact propagation. Overall, the discussed approach provides a foundation of value to a very wide audience—not just researchers and stakeholders interested in Arctic climate change or even climate change in general, but many interested in the convergence science thinking.

The exemplified convergent science approach has identified research questions that are broad in nature. As illustrated in their discussions, they do not have direct and immediate answers and cannot be fully addressed within a single disciplinary study because of numerous linkages conveyed as feed-forward connections between stressors, processes, and effects, and feedback mechanisms that are expressed as adaptive adjustments or strategies exhibited by Arctic elements. Nonetheless, because these questions were the result of integration, they can also be “differentiated,” that is, parsed into material links and tangible inquiries of disciplinary and interdisciplinary nature. While we present Yamal as a case-study region, many of the elements and their mechanistic connections presented here are broadly representative of the Arctic, and we believe our methodology for the discovery of convergence science can be applied to integrate disciplines in other systems of high complexity.

Even though not demonstrated in this paper explicitly, the process of mapping tangible questions onto a space of practical implementation is the next vital stage of the convergent science process. The illustrated development of question formulation is already tremendously useful for setting priorities among study elements and contemplating about the issues of research feasibility. But it is in the implementation stage that the research team may appreciate their gaps in expertise, methodology, and instrumentation. The team may further feel compelled to formulate additional—what we call here “emerging questions” (EQs), which can enrich and expand the scope of the overarching research thread. EQs are current knowledge gaps and stimulate a recursive (and iterative) assessment of thread linkages and content, particularly as new data and analyses start coming in to provide novel insights. It is also at that stage of the convergence science process that the relative importance of thread elements and their interactions and relevant mechanisms are re-assessed and convergence science questions are further scrutinized. Overall, a traversal of the entire convergent science pathway may characterize it not as an ultimately terminal endeavor, but as a learning process with ever-expanding dimensions of scientific inquiry.