Discussion
Our approach provides a science-based way to quantify the impacts of the
footprint of physical assets on multiple ecosystem service and
biodiversity metrics. This approach extends environment-related measures
for ESG with ecosystem service metrics. It is scalable from assets to
companies to portfolios, and across sectors, and thus can be used to
inform a range of policy, corporate, and investor decisions. It is able
to differentiate impacts based on assets’ sizes and locations.
This approach accounts for the loss of ecosystem services and
biodiversity attributable to the loss of nature directly from physical
assets, similar to Scope 1 carbon emissions32. For
some sectors, such as mining, these direct impacts may represent a
substantial portion of a company’s impact on nature. For other sectors,
especially those with extensive supply chains, expanding the assessment
to account for supply chain impacts will be critical to capturing the
full extent of a company’s impact. Here, there is great potential to
integrate our approach with methods from supply chain
analysis33 and spatial Life Cycle Assessment
(LCA)19.
Further refinement of our approach could also allow for differentiation
in impacts among and within activity types. For example, an agricultural
field and a parking lot of equal size would have different impacts due
to differences in the maintenance of vegetation, soil permeability,
addition of fertilizers, and so forth. In addition, assets of the same
type and size may differ in their use of best management practices or
nature-based solutions, and thus differ in their impacts. Utilities and
energy companies, for example, are often stewards of large land areas,
and the management of ecosystems in and around solar facilities and
power transmission lines can lead to substantial variation in impact and
even the potential to create gains relative to current
conditions34,35. These differences could be accounted
for within our approach by adjusting the ecosystem service or
biodiversity changes attributed to assets that employ certain
sustainable practices.
Finally, the sustainability field would benefit from the development of
common standards for defining baseline and impact scenarios so that
assessments could attribute impacts in a robust, intercomparable way.
Our approach currently uses potential natural vegetation as the baseline
conditions. Some urban areas have been developed for many decades, if
not centuries or millennia. Attributing full loss of nature at these
sites to the current asset owner could disincentivize activities in
existing urban areas and perversely incentivize new greenfield
development. An alternative approach could consider urban areas fixed
and focus on impacts outside historically developed
areas14, use sites’ restoration potential as a
reference scenario36, or adjust impacts by the current
population density surrounding the site, although each approach would
come with its own methodological challenges and uncertainties.
We focus on accounting for impacts stemming from the loss of ecosystems,
filling an important gap in existing sustainability and ESG approaches.
Although accounting for pollution contributed directly by assets (e.g.,
fertilizer runoff from agriculture, mine tailings, the release of
chemicals or air pollution) is beyond the scope of the current analysis,
this would be a valuable future addition. Integrating estimates from
existing LCA approaches of pollution generated by
assets19,37 with the spatially explicit modeling of
impacts to people in our approach here could advance this aspect.
By using open-source models and drawing on the growing accessibility of
asset location data and high-resolution satellite imagery, we were able
to analyze the impacts of over 2,000 companies and nearly 600,000 assets
without relying on company-reported information on ecosystem service and
biodiversity impacts. This approach can provide corporate ESG metrics
focused on impacts to nature with greater transparency and the potential
for external verification. At the same time, data availability remains a
challenge: asset-level data is available primarily for publicly traded
companies, which represent only a fraction of corporate activities, and
even for these companies, data is incomplete38–40.
Satellite imagery can capture or be used to infer many important
asset-level characteristics41,42 and is continuing to
drive advances in ecosystem service modeling at global and local
levels43. Even so, these approaches cannot be expected
to fully capture all impacts or values, and on-the-ground information
will remain an important complement, especially for understanding local
values. Ultimately, further improvements to the accessibility,
completeness, and standardization of data will be important to extending
these approaches to meet demand from consumers, investors, regulators,
and companies themselves for high quality information on nature-related
impacts.