Ecology of Practice

This chapter maps the ecology of practice within which my research is situated and through which it has been produced. Rather than treating the subsequent project chapters as the work of a single, isolated author, I frame them as outcomes of an extended set of communities of practice that span landscape architecture, architecture, engineering, ecology, and environmental science (Lave and Wenger 1991; Wenger 1998). The chapter traces how collaborations with design academics, professional offices, scientists, engineers, and public institutions have provided not only sites and problems, but also methods, tools, and languages that condition what kinds of questions I can ask and what kinds of futures I propose. In doing so, it positions my practice-based PhD as emerging from a hybrid field in which dredging logistics, geomorphic modeling, responsive sensing, territorial design, and adaptive infrastructure are continuously negotiated. This ecology of practice is both empirical and methodological as it documents the networks through which the work travels and it establishes the transdisciplinary ground on which the later chapters that are focused on specific projects and key concepts should be read.

Tracings

When I describe my work as an “ecology of practice,” I extend the language of communities of practice into the relational and multi-scalar realm of territorial design. A community of practice, in Lave and Wenger’s sense, is not just a group of people with similar expertise, it is a configuration of mutual engagement, a joint enterprise, and a shared repertoire of concepts, tools, and stories that evolve through doing things together (Lave and Wenger 1991; Wenger 1998, 2000).

My own ecology of practice spans multiple overlapping communities from design academics experimenting with sediment logistics and technological landscapes to professional offices testing speculative ideas against real projects to scientists and engineers building physical and computational models and to institutions whose regulations and infrastructures quietly frame what can be designed. These are not separate worlds, they are patches in a single, shifting mosaic, an anthropogenic biome of disciplinary cultures, to borrow Erle Ellis’ term for hybrid socio-ecological systems (Ellis and Ramankutty 2008; Ellis 2011).

Writing this ecology autoethnographically means that I do not pretend to stand outside it. I am implicated as collaborator, co-author, consultant, and critic. Autoethnography, understood as the systematic analysis of personal experience to understand cultural experience (Ellis, Adams, and Bochner 2011; Adams, Holman Jones, and Ellis 2015), provides one of the methodological threads of this chapter and my own trajectory, anxieties, and enthusiasms become data about how landscape knowledge is made.

The ADAPTr program (Architecture, Design and Art Practice Research), developed by Blythe and Stamm through the European Commission Marie Curie Initial Training Network, provides a structural parallel to the ecology of practice documented here. Their model treats each practitioner’s comprehensive body of work as a singular case study operating across three orders of knowledge. Specific projects read through research-relevant perspectives constitute the first order. Transformations and shifts visible across the full body of work constitute the second. And the positioning of the individual case study within a field of adjacent, parallel, or contrasting practices constitutes the third (Blythe and Stamm 2017, 56). This three-order structure maps onto what the refraction method produces in this dissertation. Chapter 05’s four phases of tool-making constitute second-order knowledge, transformations legible only across the trajectory. The ecology of practice documented in this chapter constitutes third-order knowledge, the field of adjacent practices within which the work’s claims become testable and its contributions become visible.

Blythe and Stamm’s concept of “transformative triggers,” specific explanatory gaps identified during the research process that provoke new projects designed to close those gaps through a transformed practice (Blythe and Stamm 2017, 59), describes the mechanism by which the tool-making trajectory in Chapter 05 actually advanced. When the Kinect sensor failed to resolve the depositional layer in the Sedimachine experiments at LSU, the failure was a transformative trigger. It did not merely identify a technical limitation but opened a question about what the instrument’s resolution was suppressing, a question that could only be answered by building different tools. Each phase transition in Chapter 05, from seeing to touching, from touching to coding, from coding to letting go, was triggered by a gap that the previous phase’s tools could not close.

What follows is not a neutral survey but a narrative map of how particular collaborations and projects have shaped my methods and concepts, and how, in turn, those methods and concepts participate in broader disciplinary shifts.

Sediment, Technology, and Territorial Interfaces

The Dredge Research Collaborative: Sediment as Medium and Method

One of the most formative communities of practice in this ecology is the Dredge Research Collaborative (DRC), specifically Rob Holmes, Brett Milligan, Sean Burkholder, Brian Davis, and Justine Holzman, among a shifting cast of designers, artists, and advocates. Together, they have spent more than a decade treating dredging not as an invisible engineering routine but as a cultural, political, and designable infrastructure.

The DRC’s recent book Silt Sand Slurry: Dredging, Sediment, and the Worlds We Are Making consolidates this work, presenting dredging as a “visually rich investigation into where, why, and how sediment is central to the future of America’s coasts,” and as an “unseen infrastructure” that “shapes and enables modern life.” The book is less a monograph than a collective atlas of fieldwork, mappings, and speculative projects, ports, disposal sites, eroding barrier islands, and vaguely defined “placement areas” that quietly organize coastal futures.

Within this collaborative, my role has primarily existed as an invited participant to co-produce within that shared repertoire. My methods working with DRC are research-through-design, developing workshops that explore nascent, responsive infrastructural practices, and staging public events as a primary means of inquiry (Zimmerman, Forlizzi, and Evenson 2007; Lenzholzer, Duchhart, and Koh 2013).

Autoethnographically, the DRC’s workshops and field investigations produced a specific methodological conversion. Before this relationship, I understood dredging as background logistics, something that happened to rivers and harbors to keep them functional. The DRC’s sustained attention to “placement areas,” the sites where dredged material is deposited, often vaguely defined in regulatory documents, often on the urban edge, often in the process of becoming something that no one has yet named made visible a design problem that the academic literature had not framed as one.

The DRC showed me where to look. The geomorphology table research at REAL, and later at UVA, can be understood as an attempt to develop instruments for what I had learned to see at those sites and the choreography of sediment as a medium of territorial formation, not a waste product of navigation maintenance.

The Thresholds installation (2006) had already demonstrated this principle in miniature before the DRC made it legible at territorial scale. People walking through the atrium did not know they were performing landscape. The camera read them as perturbations in the contrast field and the software turned those perturbations into topographic events, momentary ridges, valleys, migrating peaks that appeared and dissolved as each person moved. In this view the occupants become part of the territory. The isolines represent the interaction between datum, occupant and light, and any of those three can be understood as the landscape depending on which is taken as the fixed reference. The occupants were not designed into the system. They entered it and, through the logic of the sensing apparatus, became constituents of the territory the installation produced.

Technological Genealogies: Richard Hindle and Patent Landscapes

In parallel with my sediment work is my ongoing engagement with Richard Hindle’s research into patents and landscape technologies. Hindle mines patent archives to reconstruct the histories of infrastructural devices such as green roofs, irrigation systems, coastal defense, and retaining structures, often revealing how technological imaginaries and environmental control fantasies are encoded in intellectual property. His contributions show how patents form a spatial archive of environmental design, including unbuilt or forgotten artifacts.

Methodologically, Hindle’s work offers a form of design-inflected historiography. Before speculating about new sensing infrastructures or hydro-mechanical interventions, I have learned to look back and trace how similar systems have been framed, which problems they claimed to solve, and what hidden assumptions they carried about nature and control. This historical work complements my field and modeling work, grounding my speculative projects in a longer genealogy of environmental technologies.

In autoethnographic terms, learning from Hindle shifted how I enter technical conversations. I am less likely to accept a new technology at face value, instead, I instinctively ask what is the historical lineage of this device? What failure modes and fantasies are already baked into it?

Territorial Interfaces: Alexander Robinson and Owens Lake

Alexander Robinson’s work at the Landscape Morphologies Lab and the Office of Outdoor Research extends my focus from sediment logistics to broader territorial interfaces. His book The Spoils of Dust: Reinventing the Lake That Made Los Angeles examines Owens Lake, once California’s third-largest inland water body, drained to serve Los Angeles, as a hybrid of infrastructural experiment, environmental justice site, and public spectacle (Robinson 2018).

Robinson’s work changed how I understand the function of maps and models in my own practice. Before this engagement, I treated visualization as a communication tool, a way of explaining design decisions already made. Robinson’s Owens Lake work demonstrated that the map is itself a design act, determining what counts as visible, negotiable, or inevitable. That reorientation runs directly into the Technogeographies argument in Chapter 07.

Computational Communities

The ACADIA community (Association for Computer Aided Design in Architecture) forms another crucial node in this ecology, not as a stylistic influence but as the context in which I came to understand computation as environmental practice rather than formal technique. Working alongside Jason Kelly Johnson and Nataly Gattengo at Future Cities Lab, whose responsive installations couple sensing with urban atmospheres, and Dana Cupková, whose material computation foregrounds thermodynamic and climatic consequences, shifted my orientation from what computation can represent to what computation can do within a living system.

The books produced within this ecology of practice deserve a moment’s attention as knowledge contributions rather than merely community artifacts. Digital Drawing for Landscape Architecture (2010, with Wes Michaels) addressed a specific crisis, the profession had adopted digital tools wholesale from adjacent disciplines, but the representational conventions those tools encoded such as vector precision, object clarity, architectural rendering conventions were inadequate to landscape’s constitutive concerns with indeterminacy, atmosphere, and temporal change. The ASLA Award of Excellence jury recognized the contribution, praising the book as a contemporary and accessible treatment of digital drawing for the discipline. What that recognition named was a representational void the book had filled. You cannot design what you cannot represent and the book changed what landscape architects could see.

The drawing techniques in Digital Drawing for Landscape Architecture were calibrated to landscape’s indeterminacy, temporal depth and atmospheric qualities. By choosing tools that make those properties legible, the book shaped what the territory could be known through. The composite drawing became an epistemic commitment. Landscape is heterogeneous, no single representational engine can capture it, and practitioners must work across multiple modes to reveal what any single instrument obscures.

Codify: Parametric and Computational Design in Landscape Architecture (2018, edited with Adam Mekies) made a stronger argument, that landscape architecture is not adopting computational methods from architecture and engineering but is uniquely positioned to define an emerging domain where ecology, urbanization, and technology converge. Richard Hindle’s response, “Codify convincingly argues that Landscape Architecture is uniquely positioned to define this sector of technology, and in the process redefine itself” names the disciplinary ambition. These publications are not the outputs of this ecology of practice. They are interventions within it, reshaping what questions the community can ask.

I write tools to embed my landscape claims in code. Selecting variables, thresholds, and outputs constitutes the same design decisions I make when drawing a planting plan. By coding, I make the decision structure explicit and reproducible. I teach students to write their own tools so they understand the epistemic claims embedded in them. Using a black-box tool abdicates that responsibility. The territory then operates on unknown assumptions. Tool-making therefore becomes a research practice that holds me accountable for the epistemological infrastructure of my work.

In ACADIA and related venues, prototypes, scripts, and installations are explicitly framed as research-through-design artifacts, producing designed objects that embody and generate theoretical insights (Zimmerman, Forlizzi, and Evenson 2007; Zimmerman, Stolterman, and Forlizzi 2010). My own autoethnographic record of conference papers, workshop notes, and moments of failure captures the discomfort and excitement of exposing unfinished tools and ideas in public, and watching them be adopted, modified, or critiqued by others.

Publishing Communities and Collaborative Authorship

A distinct thread within this ecology of practice runs through the books and essays produced with collaborators whose intellectual contributions have been formative rather than merely additive. These are not publications that gathered existing knowledge but relationships through which new knowledge was made.

Wes Michaels and I arrived at Louisiana State University in the same year, hired as assistant professors, having come through the same post-professional MLA cohort at the Harvard Graduate School of Design. The collaboration that produced Digital Drawing for Landscape Architecture (2010) was built on complementary strengths, his grounding in design practice, my orientation toward computation and technology, and on the shared conviction that the profession needed representational tools calibrated to its own material concerns, not borrowed wholesale from adjacent disciplines. The book is now in its second edition and remains a textbook across landscape architecture programs. That continuity matters. It is evidence that the representational void the collaboration identified was real, and that the methods developed to fill it have proven transmissible across generations of practitioners.

Writing the book forced us to articulate techniques that had emerged from practice, turning tacit knowledge into explicit language. This act of making the practice legible generated knowledge that circulated through the discipline, changing how others draw and therefore how they think. The gap between the unspoken studio knowledge and the published articulation became the source of new understanding, demonstrating that the ecology of practice produces epistemic growth through the very act of writing.

Justine Holzman arrived as a student, became a research assistant, and became a co-author. The arc of that relationship spans nearly a decade and its intellectual contribution to this dissertation is specific and significant. Her framing of the Modify chapter in Responsive Landscapes (2016) opened a conceptual space that I had circled without naming. That actively altering an environment in real-time constitutes a knowledge-producing act. Modification as epistemology. The landscape changed, and the change was the data. That insight runs as an undercurrent through everything that follows in this dissertation.

Making invisible processes visible changes how communities experience their territory. When perception shifts, political agency follows, not automatically, but because what was previously unnoticed becomes available for contestation. The distance between what a community senses about its environment and what policy acknowledges about that environment is where the design of responsive landscapes does its most consequential work.

Together, Holzman and I organized the Adaptive Devices workshop at DredgeFest Louisiana (2014), where we tested the modify-as-knowledge intuition before the language existed to name it. The workshop staged responsive sediment infrastructures at tabletop scale before the argument had been formalized, before we knew what we were demonstrating. The KERB essay “After Modify” (2015), co-authored with Holzman, was where the formalization happened. The workshop’s intuitions became articulated claims. And her independent chapter in Codify (2018), published in the volume Adam Mekies and I edited, extended the argument further, hinting at the landscape itself as a model, a claim the dissertation later develops more fully. This is what a community of practice looks like when it is generative rather than merely collegial. An idea moves through a workshop, into a co-authored essay, and returns transformed in an independent voice.

Adam Mekies represents a different but equally sustained collaboration. Our relationship begins at Sherwood Design Engineers, where theoretical frameworks developed in academic contexts are tested against engineering constraints and institutional realities of professional practice. But it extends into the intellectual register through Codify, which we co-edited and opened together. The Dredgefest encounter and the Codify essay are two moments in a longer conversation about what it means to think computationally about landscape rather than merely use computation in landscape work.

Student Communities

No account of this ecology of practice is complete without acknowledging the communities of students whose work has been generative rather than assistive. Three registers of relationship deserve recognition, each producing a different kind of knowledge contribution.

Research assistants, working at the edges of funded projects and laboratory investigations, have extended the capacity to prototype, test, and document. Their contributions are embedded in the technical infrastructure of the work, in the sensing systems, the software scripts, the image archives, often without appearing in citations or credits.

Students in studios, seminars, and thesis work have pushed ideas into territory the laboratory could not anticipate. The student projects cited throughout this dissertation were not illustrations of pre-existing frameworks but inquiries that opened new directions. Studio work operates through a productive asymmetry. The instructor sets conditions, but the outcomes exceed the conditions, and the excess is where the knowledge is.

Foundation studios are typically understood as sites of pedagogical transmission. Prototyping the Bay inverted this structure in important ways. The Chesapeake Bay’s coastal landscapes in conditions of active sea-level rise and ecological transition are not well enough understood for the instructor to have answers to the design questions the site poses. The instructor has frameworks for inquiry but not answers, because the answers depend on ecological dynamics that are still unfolding and on design propositions that the studio would develop rather than transmit. The studio was a research site as much as a pedagogical one, and the students were researchers as much as learners. This is what the doctoral inquiry recognizes as the ecology of practice in its pedagogical form. A community of inquiry in which the instructor and students are learning together from the site, from each other, and from the tension between the design frameworks the studio provides and the site conditions that challenge those frameworks.

PhD advisees have constituted the most reciprocal relationships, collaborations in the fullest sense, where advising and being advised are not cleanly separable. Marantha Dawkins, whose research on climate and landscape culminated in a co-authored paper in Prospectives (the Bartlett BPro online journal), brought a sustained focus on environmental justice and the political dimensions of responsive systems that sharpened the dissertation’s own justice arguments. Xun Liu, co-collaborator on the UVA geomorphology lab research and the Venice Biennale’s Indeterminate Futures (2021), has been a consistent intellectual interlocutor in the transition from responsivity to autonomy, and co-author with Zihao Zhang and myself on a chapter in the Routledge Handbook on Artificial Intelligence in Architecture. Zihao Zhang, co-author on that same chapter and on the essay “Cultivated Wildness: Technodiversity and Wildness in Machines” published in Landscape Architecture Frontiers (2021), is the collaborator with whom the concept of the Third Intelligence was first formally framed, a concept that now anchors Chapter 11. That the dissertation’s central claim about non-human agency in design emerged through collaborative inquiry with students-become-peers is not incidental. It enacts the argument.

Transforming REAL into a pedagogical platform required making the method transmissible. Teaching forced the explicit articulation of knowledge that had remained tacit within the lab’s own practice, protocols that had developed through accumulated experience needed to be named, sequenced, and documented before they could be shared. That act of codification was itself a form of knowledge production, and it reshaped who could access and influence the territory’s design by opening the method beyond the original research team.

Theoretical Interlocutors

Beyond the Waldheim and Raxworthy framing that closes the design-academic world, the ecology of practice that produced this dissertation has been shaped by a wider set of theoretical interlocutors whose ideas and relationships have been equally formative.

Christophe Girot’s work sits at the intersection of terrain perception and digital computation. His more recent work with point clouds and digital terrain at ETH Zürich argues that new sensing technologies can recover attentiveness to the specificity of place that homogenized ecological planning methods have suppressed. The Codify foreword he contributed names this directly. Coding, at its best, is a return to the need to inscribe essential meaning in our daily lives on the ground we tread upon, a counter to the flattening of landscape into two-dimensional scientific abstraction. Girot’s insistence that computational methods must remain answerable to terrain’s material and cultural specificity has been a persistent check on the more totalizing ambitions of responsive systems thinking.

Elizabeth Meyer’s work on aesthetics and performance in landscape architecture grounds this dissertation’s understanding of what design is for. Her argument, developed across “The Expanded Field of Landscape Architecture,” “Sustaining Beauty,” and related essays, is that experiential and ecological dimensions of landscape are not separable. Aesthetic experience is not decoration applied to ecological performance but the medium through which ecological value becomes legible and meaningful to those who inhabit a place. This insistence on the constitutive role of experience runs beneath this dissertation’s concern for cultivated wildness and multi-species co-authorship. If the landscape’s autonomy is invisible or inert to human perception, it cannot function as a site of ethical or epistemological encounter.

Robert Pietrusko and I overlapped at Harvard during the years of the REAL lab’s development. His argument, that data does not neutrally record the world but produces the categories through which the world becomes known and designable, provided theoretical grounding for the sensing work at REAL and UVA. Together we developed a proposal for an MDeS program at Harvard focused on Data Ecologies, the design of data acquisition, processing, visualization, and application as a coherent disciplinary practice. The proposal was not realized, but the conversation it required crystallized the argument that the design of sensing infrastructure is a form of territorial design, a claim that runs through the Technogeographies work in Chapter 7.

Nicholas de Monchaux and I were both fellows at the American Academy in Rome, and it was there that he articulated something that reoriented how I understood the dissertation’s central claim. That the landscape is the medium. Not the model of the landscape, not a representation of it, not a surrogate, but the landscape itself as the medium through which knowledge is produced and tested. His Local Code, which uses parametric tools to locate distributed urban remnants and design them as collective ecological infrastructure, enacts this principle at urban scale, treating the city’s actual terrain as the computational and design medium rather than abstracting it into a model.

Theoretical Bookends

Two theorists quietly bookend this design-academic world with Charles Waldheim and Julian Raxworthy.

Waldheim’s essay “Strategies of Indeterminacy in Recent Landscape Practice,” written in the context of Downsview Park, articulates a mode of distanced authorship in which landscape architects design frameworks and processes rather than fixed forms (Waldheim 2001). This concept gives a name to design with sediment logistics, AI-mediated management, or responsive infrastructures and I am setting up conditions, rules, and feedback loops that will enact, over time, a design that no one fully controls.

Raxworthy’s book Overgrown: Practices Between Landscape Architecture and Gardening insists on gardening, continuous, embodied maintenance, as central, not peripheral, to landscape architectural practice (Raxworthy 2018). He argues that design cannot be separated from the temporal, vegetal, and labor-intensive realities of care. This insight haunts my collaborations and when I work with geomorphological models, view a sediment test plot, or explore a heavily engineered coastal project, I see them through Raxworthy’s lens as complex gardens requiring long-term tending. The technologies we deploy should not replace gardening practices, instead they are part of an expanded repertoire of Wetware and Reflexive Stewardship. And this is the place where this dissertation departs from Raxworthy toward its own theoretical territory. The cultivant, as developed in Chapter 11, is my extension of the viridic. Living matter is a growing medium, but the cultivant names something further, the ongoing relationship between designed intention and biological agency, enacted through cycles of territorial maintenance, constituting a form of communication that accumulates knowledge over time (Raxworthy 2018; see Chapter 11).

Design Practice Tension

Process-Driven Urban and Territorial Design

In professional practice, a sustained collaboration with Stoss, led by Chris Reed in Boston, has been where responsive strategies first encountered the institutional resistance that academic contexts rarely produce. My role has been as an embedded researcher and consultant in early concept phases, introducing sensing methods and adaptive frameworks that must survive contact with engineers, contractors, and permitting agencies. What Stoss taught me was which ideas survive that contact and which do not, and that the difference is rarely about the quality of the idea. It is about whether the idea can be translated into a deliverable that an institution can maintain. That lesson shaped the dissertation’s insistence that adaptive epistemology must account for institutional capacity, not just ecological complexity.

This is a form of research through practice and iterative theorization where projects become testbeds for ideas about responsive landscapes, infrastructural ecologies, and adaptive urbanism. These projects weave inquiry, strategy, and design in ways that produce knowledge about how certain strategies have agency in the world (Deming and Swaffield 2011; Swaffield and Deming 2011).

Territorial Application of Theory

More recently, my collaboration with Sherwood Design Engineers and in particular with Adam Mekies in their New York office, has become a crucial part of this ecology of practice. Sherwood describes itself as an engineering firm working “at the forefront of ecological infrastructure and urban systems design,” specializing in climate-responsive planning, smart-city infrastructure, and performance-driven modeling. Mekies, an associate principal at Sherwood and my co-author on Codify, is a landscape architect whose work focuses on the role of computation and construction in environmental and ecological design.

With Sherwood, the scale and stakes of my speculative ideas intensify. Projects range from large-scale territorial landscapes that address coastal protection, regional infrastructures, and urban districts, to focused studies of specific technologies. Questions I have explored in academic and design contexts regarding sediment logistics, AI-mediated management, or infrastructural ecologies are translated into deliverables that must survive engineering review, public scrutiny, and financial analysis.

The work with Sherwood has allowed me to interface with the building of multi-scalar models that integrate hydrology, energy, and land use with spatial design scenarios focused on adaptation. It has also developed computational workflows in the form of scripts, parametric models, and data pipelines that allow engineers and designers to test variants under different environmental scenarios. In some of the most enlightening aspects of this collaboration, scenario planning engages governmental agencies and stakeholders by using speculations and models to facilitate discourse around infrastructural futures.

In these projects Sherwood’s team of civil engineers, computational designers, and landscape architects puts my theoretical developments under stress. An idea like “cultivated wildness” must be translated into performance metrics and management regimes and “Wetware” becomes technical specifications justified by ecological performance.

The NEOM consultation (2022–25) is the sharpest instance of this theoretical stress-testing. Hydrological modeling using GeoHECRAS across multiple storm frequencies revealed that conventional channelization of the wadis surrounding The Line would require infrastructure widths exceeding 200 meters with extensive hardened concrete at velocities that made ecological function impossible. The finding was not a design preference but an engineering limit. The alternative was a hybrid hydrological approach that engaged managed complexity rather than minimized it, routing water through reconceived wadi systems as holding areas, recharging aquifers, sustaining mangroves through fluctuating isohaline zones an outcome demanded by the project’s constraints rather than imposed by theory.

This is the methodological contribution of professional collaboration that academic practice cannot replicate. The concept of “managed complexity” does not mean the same thing in a research seminar and in a GeoHECRAS output showing 200-meter infrastructure widths. Working with Sherwood sharpened the argument by forcing it to survive contact with the conditions it claimed to address. The friction was productive precisely because it was uncomfortable and it required translating “cultivated wildness” into a language that engineers could evaluate, which in turn required clarifying what the concept actually claimed.

Autoethnographically, my experience with Sherwood oscillates between exhilaration and discomfort. It is heartening to see speculative concepts about sediment as infrastructure, AI as co-designer, or computational wildness circulate in engineering drawings and policy briefs. It is also uncomfortable to see how quickly those concepts can be simplified, instrumentalized, or resisted. But this friction is methodologically generative as it forces a refinement of language, clarifies assumptions, and confronts the institutional realities that theoretical frameworks must navigate.

The work with Sherwood Design Engineers and Stoss is where theory meets productive tension and where ideas formed in research are either flattened by cost and risk or adapted and strengthened.

Co-Authoring Ecologies

AI-Mediated “Wildness”

My collaborations with Erle Ellis, a landscape ecologist known for his work on anthropogenic biomes, move my practice into new disciplinary territory. Ellis’s classification of the terrestrial biosphere into human-modified biomes makes visible the extent to which ecological patterns are already shaped by long-term human land use (Ellis and Ramankutty 2008; Ellis 2011).

In joint projects and writing, with environmental historian Laura Jane Martin and technologist David Klein, we ask what it would mean to design AI-mediated systems that manage these anthropogenic biomes toward forms of perceived wildness that are orchestrating disturbance, succession, and access patterns in ways that privilege ecological complexity and more-than-human claims, while remaining legible to human governance. These are speculative exercises that test how an “agent” might govern land management decisions according to different objectives.

The Designing Autonomy paper, published in Trends in Ecology & Evolution, a leading journal in ecological science, not a landscape architecture venue, is itself an artifact of this ecology of practice. That publication required the Ellis collaboration, the Martin collaboration, and the willingness to place landscape architecture’s claims in a scientific journal that hadn’t encountered them before. One of the dissertation’s arguments is that landscape architecture is uniquely positioned to define an emerging domain at the intersection of ecology, technology, and territorial governance and the fact that this claim could be published and received in Trends in Ecology & Evolution is evidence that the ecology of practice that produced it extends beyond landscape architecture’s own disciplinary boundaries.

The projects are instances of research through design operating at the scale of intractable planetary issues and design scenarios and models are not only visualizations but narratives to interrogate how “wildness,” control, and equity are encoded in machine intelligence.

Rivers in the Laboratory

With civil engineer Clint Wilson at Louisiana State University, my research has evolved through the friction that emerges when speculation and engineering come into contact. Our work overlapped within the Louisiana State University Coastal Sustainability Studio, where a range of projects from design competitions to historical preservation assessments allowed me to test beginning intuitions about modeling alongside one of the leading scholars in civil engineering.

From Wilson’s standpoint, physical models are aimed at understanding and predicting geomorphic behavior under different flow regimes and engineered configurations. From my point of view, they are generators of design heuristics through tangible testing of river cross-sections, levee alignments, cut-off channels, and sediment diversions, metaphorically sketching with water and sand.

The Sedimachine experiments at LSU (2012) crystallized this distinction. When the Microsoft Kinect depth sensor failed to resolve the depositional layer as the sediment was too thin for the instrument’s resolution to capture meaningful topographic variation, Wilson and I drew different conclusions. From a hydraulic engineering standpoint, the failure indicated a technical limitation requiring either better equipment or a different experimental setup. From my standpoint, it indicated something else in that the plexiglass substrate and controlled flow conditions were producing phenomena at a spatial scale that exceeded the instrument’s resolution, which meant that the sediment choreography must rely on the expression of the outcome (morphological results) as pattern rather than particle. The failure redirected subsequent research toward the EmRiver geomorphology table and toward ultrasonic range-finding and image analysis as complementary sensing modes.

This interpretive divergence is not a miscommunication between disciplines, it is how communities of practice with different purposes produce different knowledge from shared material. The ecology of practice depends on this divergence remaining in productive tension rather than resolving toward either purely scientific or purely design purposes. The geomorphology table research at REAL and UVA has maintained this tension deliberately, the table is not a hydraulic engineering instrument, but it draws on hydraulic engineering knowledge to generate landforms that are read as design material.

The methodology aligns with research through designing, except that designing is delegated to physical processes (Lenzholzer, Duchhart, and Koh 2013). The models produce time-based traces in the form of photos, videos, and point clouds that are aligned with scientific data to produce design material. Reading them together, Wilson is looking for predictive patterns and scaling laws while I am looking for emergent landforms that might host futures for settlement, habitat, or public access.

Geomorphology and Design Labs

With geomorphologist Ajay Limaye at the University of Virginia, I co-developed research that pairs geomorphic modeling with design experimentation. Limaye’s work uses numerical and physical models to understand channel networks, delta morphodynamics, and river planforms.

During our work together, the collaboration aimed to build a stable workflow where design students engage directly with geomorphic models, modifying boundary conditions, exploring interventions, and interpreting outputs as design prompts. This led to the creation of physical models and computational tools that unpack assumptions about settlement patterns, infrastructure, and ecological restoration in terms of sediment budgets, flow regimes, and morphological thresholds.

The collaboration was, in effect, a designed community of practice, with student research assistants at both undergraduate and graduate levels and where geomorphologists, landscape architects, and students learn to wield cross-disciplinary tools and methods. There is an insistence in practice-based methods that landscape research should operate across the case study, lab experiment, and design research (Deming and Swaffield 2011), and this collaboration is a clear instantiation of that insistence.

Boundary Encounters

My engagement with institutions like the U.S. Army Corps of Engineers, The Nature Conservancy, and the Environmental Protection Agency is more episodic but no less formative. These agencies manage and regulate many of the landscapes where my work has relevance. Working with them means reading dredge management plans, environmental impact statements, and design guidelines not as neutral technical texts but as design and governance documents whose assumptions shape what can be built. It means participating in workshops, stakeholder meetings, and site visits where community concerns, engineering constraints, and design ambitions collide. And it means producing what Star and Griesemer (1989) call boundary objects, maps, scenarios, and diagrams robust enough to maintain common identity across professional cultures yet plastic enough to serve each discipline’s purposes. These are emotionally charged spaces where power, frustration, and institutional memory are palpable. My sense of caution, hope, or skepticism in these rooms is part of the research. It shapes the questions I ask and the compromises I am willing to entertain (Luz 2000). These institutions function as boundary partners rather than core communities, places where the repertoires developed with the DRC, ACADIA, Stoss, Sherwood, Ellis, Wilson, and Limaye are tested against formal governance systems (Wenger-Trayner and Wenger-Trayner 2014).

The ADAPTr model also reframes how communities of practice function in practice-based research. Rather than treating them as static disciplinary backgrounds, Blythe and Stamm describe them as “highly dynamic, and varied to the point where the researcher may attribute variable communities of practice to different individual projects within the body of work” (Blythe and Stamm 2017, 58). The community of practice that shaped the Sedimachine at LSU, consisting of hydraulic engineers, sediment scientists, and geomorphologists, is not the same community that shaped REAL at the Harvard GSD, which drew on computational designers, ecologists, and interaction designers. Each project within the ecology of practice documented here carries its own community, and the practitioner’s movement across these communities is itself a form of knowledge production. What Blythe and Stamm call the practitioner’s capacity to “constellate,” to operate in public modes beyond the stereotype of the solitary creative individual, is what makes the ecology of practice an epistemological structure rather than a biographical one (Blythe and Stamm 2017, 60).

Situating the Ecology of Practice

The ecology of practice documented in this chapter participates in several disciplinary shifts simultaneously. Classical civil engineering has been shaped by an ethos of control, levees designed to historical flood stages, rivers straightened to maximize conveyance. My collaborations with Wilson, Limaye, Sherwood, and the Corps participate in a gradual reorientation toward adaptive and process-based infrastructures, one in which physical models are read as both experiments and design sketches, and territorial propositions are evaluated on socio-ecological criteria alongside hydraulic metrics (Pickett et al. 2001, 2011). Within landscape architecture, the work extends Bélanger’s argument that landscape is a synthetic medium capable of organizing flows of water, waste, energy, and capital (Bélanger 2009, 2016) by treating sediment itself as infrastructure and infrastructure as landscape. It extends Waldheim’s strategies of indeterminacy (2001) into adaptive sediment regimes and responsive coastal systems. And it extends Allen’s concept of field conditions (1999) into the literal fields of sensor arrays, diversion grids, and robotic infrastructures that the later chapters document. Ecologically, the work is grounded in Ellis’s anthropogenic biomes (Ellis and Ramankutty 2008) and Pickett’s patch dynamics (Pickett and White 1985), frameworks that treat nearly all of the landscapes engaged here as hybrid socio-ecological mosaics at different stages of disturbance and succession. The ecology of practice is an autoethnography of working inside those mosaics, using design to steer trajectories rather than to reset baselines.

If the practice is the research instrument, then what exactly has the practice done? What is the accumulated body of work, and how does each project build on the one before it? What were the instruments, the provocations, the failures that redirected the inquiry? And what theory has the practice been enacting all along?