A Nature First Approach Will Lead to Better Data Centers

Part II of a three-part series

Data center with rooftop solar and a pond, U.S. / istockphoto.com, marcello74

Data center development continues to accelerate. Communities are demanding more oversight and benefits. Government agencies are reconsidering how these facilities are planned and approved. And some technology companies and developers are rethinking how data centers can reduce their environmental impacts. These trends deserve attention because they offer evidence that biodiversity, water, and habitat can become important considerations in data center planning, design, and operations.

Microsoft, for example, recently described a nature-inspired approach for their data centers that began with a pilot in the Netherlands and is now expanding to more than 20 sites in the U.S. and Germany. The company says the approach is being incorporated into all new Microsoft data centers designed in the U.S.. In Mecklenburg County, Virginia, this work includes more than 230 acres protected from development, restored streams, 8 acres of new wetlands, 185 acres of native pollinator habitat, and more than 25,000 trees planted to establish forested buffers. Similar efforts are underway at other Microsoft sites.

Other companies like Vantage Data Centers are using Ramboll's Americas Biodiversity Metric to inform plans for Vantage's Lighthouse Campus in Wisconsin, providing a way to establish baseline biodiversity conditions and evaluate changes associated with development. Google is making tangible commitments to water stewardship, biodiversity, clean energy, and habitat restoration associated with its operations and has developed approaches for evaluating watershed health in communities where its data centers operate.

Google data center with wind turbine and water basin in Eemshaven, The Netherlands / istockphoto.com, HildaWeges

These initiatives vary in scope and application, so it is difficult to determine how consistently they are being applied across the data center industry. Understanding how far these practices extend is complicated by the way data centers are developed and operated. Microsoft, Google, and other hyperscalers own and develop some facilities while also leasing facilities or capacity developed by third parties. Second and third-party developers may acquire and entitle the land, construct the facility, and operate the data center for one or more tenants. Ecological commitments of a technology company do not necessarily tell us what practices are being applied across all their physical infrastructure. This also raises an important question: Are they using ecological and social justice criteria to influence site selection, development, leasing, and procurement decisions throughout the supply chain?

At Biohabitats, we are encouraged with these advances and what landscape architects can learn from them. We are equally interested in the questions that ecological design and mitigation cannot answer.

Understanding the Footprint of Data Centers

A data center project can restore habitat, conserve water, reduce carbon emissions, and improve biodiversity through on-site and off-site measures while questions about location, scale, cumulative impacts, infrastructure, and community remain unresolved. As the physical scale of digital infrastructure grows, these questions become more important because the ecological relationships affected by a data center can extend far beyond the property where buildings and supporting infrastructure are located.

The growing interest in biomimicry makes this distinction notably important. Native plant communities, restored streams and wetlands, habitat connections, and other nature-based approaches can make meaningful ecological contributions. Using vegetation primarily to camouflage a large industrial facility or help it blend into a woodland landscape raises additional issues. If development requires clearing or fragmenting that landscape in the first place, making the resulting facility appear more natural does not address the ecological implications of siting it there. Biomimicry has greater value when it helps us understand and work with the ecological processes of a place, beginning with whether that place is appropriate for this type of development to begin with.

This is where we believe landscape ecology and conservation planning has an important role to play. Before asking how the impacts of a data center can be mitigated, we should understand the landscape well enough to ask whether the development belongs there in the first place. That means looking beyond the proposed site to understand the watershed, ecological connections, wildlife habitat, water availability and quality, climate vulnerability, existing conservation investments, and cumulative development pressures.

It also requires us to understand the larger footprint required to support the data center, including energy generation and transmission, water and transportation infrastructure, and mitigation or restoration that may occur elsewhere in the watershed or region. Together, these relationships provide a more complete picture of what the development requires of a place and whether that place can accommodate it.

The geography of a data center is often considerably larger than the data center property. Energy and water infrastructure can extend across multiple landscapes and communities. Mitigation lands may be adjacent to the facility or located elsewhere within the watershed or region. Understanding the ecological implications of a project therefore requires defining its boundaries through the relationships and systems affected by development, not simply through property ownership.

Site selection therefore needs to be considered along with the amount of land a data center requires. The sprawling, single-story campus has become a common development model, but multistory data centers demonstrate that this form is not inescapable. Previously developed land, brownfields, former industrial properties, and other disturbed sites can accommodate new facilities without converting large areas of forests, wetlands, and greenfields. Underground or partially below-grade facilities may also be possible where geology, groundwater, cooling, access, and other conditions allow.

Before accepting hundreds or thousands of acres of land conversion as an inevitable consequence of digital infrastructure, we should demand that the footprint be reduced, consolidated vertically, located below grade where appropriate, and directed toward land that has already been developed. These choices are fundamental to putting Nature First because they address ecological relationships and land conversion before land planning, design and mitigation begin.

Leveraging a Nature First Framework

This thinking is central to a framework Biohabitats has been developing over the past year - Nature First: A Framework for Data Centers. The framework begins with site selection because many of the most consequential decisions occur before design begins. It then considers communities, living systems, and adaptive futures, providing a way to examine how a proposed data center relates to the ecological and social systems of which it will become a part.

Nature First: A Framework for Data Centers / Biohabitats

Location is perhaps the most important dimension of responsible decision making for data centers. Communities need enough information, early enough, to understand what is being proposed and participate meaningfully in decisions that may alter their landscapes, infrastructure, economies, water, and quality of life.

Information asymmetry, non-disclosure agreements, and limited early engagement have been major concerns for communities confronted with data center developments. The Bipartisan Policy Center, following discussions with more than 150 stakeholders, identified transparency and earlier engagement as key, while the World Resources Institute has emphasized early, transparent, and continuous engagement in developing community benefit agreements.

Landscape architects have a history of engaging with these relationships among ecology, community, and place. Randolph "Randy" Hester, FASLA, explored them throughout much of his career. In the book Design for Ecological Democracy, he asked practitioners to consider who participates in decisions about place, what knowledge counts, whose interests are represented, and how the living world is considered in the choices we make. His work is part of a larger body of landscape architecture thought upon which we can continue to build as the profession confronts a novel generation of infrastructure.

For data centers, that means communities need access to information about land, water, energy, infrastructure, tax incentives, ecological impacts, and long-term commitments, along with independent expertise capable of evaluating that information. Our understanding of community should also encompass the living systems affected by these decisions. Forests, wetlands, streams, aquifers, plants, animals, and other living communities have needs and relationships that can be represented through ecological science, direct observation, monitoring, Indigenous knowledge, Ancestral Wisdom, and people who have developed deep relationships with a certain place.

Supporting Nature's Intelligence

This larger understanding is leading us to explore what we call Nature Intelligence (NI). Nature Intelligence recognizes the knowledge embedded in landscapes and living systems. Watersheds reveal relationships among rainfall, geology, soils, vegetation, groundwater, streams, and rivers. Species movement reveals connections across landscapes, while forests, wetlands, grasslands, and other living communities continually respond to disturbance and changing conditions. Ecological science helps us understand these patterns, while direct and deep observation keeps that understanding grounded in place. Indigenous knowledge and Ancestral Wisdom contribute ways of knowing rooted in long relationships with land and living communities, while Two-Eyed Seeing, introduced by the Mi'kmaw as Etuaptmumk, offers an approach for bringing Indigenous and Western knowledge systems into relationship while respecting the integrity of each.

Bringing these ways of knowing together can deepen our understanding of what a landscape can support and flourish, which relationships need to remain intact and nurtured, and where ecological limits may be approached or exceeded. This has relevance for an industry whose rapid expansion is being driven in large part by artificial intelligence.

As AI becomes increasingly integrated into planning, design, engineering, and construction we can begin by understanding the intelligence already present in nature and the accumulated knowledge derived from sustained relationships with place. Within Biohabitats, we are exploring this as NI, where AI is subservient to NI and is used only when it advances nature's intelligence.

The ecological innovations emerging within the data center industry provide important opportunities for learning. Restoring streams and wetlands, reestablishing native plant communities, protecting and conserving habitat, reducing water consumption, measuring changes in biodiversity, and engaging communities early on can contribute to better outcomes. Native vegetation used to screen a facility may also provide ecological benefits, but those benefits need to be considered alongside the land converted by the facility and opportunities to reduce its physical footprint. Brownfield redevelopment, adaptive reuse, multistory construction, and other approaches that reduce greenfield conversion should be considered along with restoration and mitigation.

As the data center industry becomes increasingly interconnected, responsibility for ecological outcomes should be shared across developers, operators, hyperscalers, utilities, investors, designers, engineers, regulators, and communities. This raises questions that extend beyond individual data center developers:

• If a technology company leases capacity in a facility developed and operated by someone else, do its ecological obligations influence where that facility is located and how it is designed?
• Do biodiversity, water, community engagement, and landscape connectivity become part of the procurement or leasing criteria?
• Do those commitments encompass the generation and transmission infrastructure supporting the facility and the lands used to mitigate or compensate for impacts?

Nature First is intended to help bring these questions into the process early enough to influence fundamental decisions and positive change. It can help identify previously developed and industrial lands that may be more compatible with data center development. It can also look beyond the development parcel to understand how a data center fits within the larger landscape and community.

When Biohabitats began developing the framework, we did not intend to create another rating system or certification program. We wanted a practical way to ask better questions and bring nature into decisions that are often made before the design, engineering, and public have meaningful opportunities to participate. The framework will continue to evolve as we learn from communities, practitioners, researchers, Indigenous knowledge holders, conservation organizations, governments, utilities, developers, and others working through these issues.

The advances being made by data center developers and others are part of this learning curve, as are the experiences of communities developing their own approaches to influence where and how data centers are built, how decisions are made, and how benefits are addressed. Bringing these experiences together can help expand the conversation from improving individual facilities to understanding how rapidly growing digital infrastructure can fit within the ecological limits and relationships of place.

As technologies change and experience grows, applying a Nature First framework will go a long way in mitigating climate change, reducing the loss of biodiversity, and redressing environmental injustices.