Earth’s human-sustaining environment is broadly the result of biogeochemical cycles (interactions of living things, minerals, and chemicals in cycles) that provide the essential elements required for interactions between the environment’s biotic and abiotic components. Biogeochemical cycles are driven in complex ways by interactions in the soil, as the soil is where water, air, geological minerals, organic matter, microbes, plants, and other organisms interact chemically and physically, driven by solar, gravitational, and geo-thermal energy to produce the systems that make up the natural environment. When we refer to sustainability, we are referring to the fundamental ability of a soil-based biosphere to sustain humans. The implication is that, for a BLSS that utilizes biological functions to be sustainable and thus provide the necessary services required for human survival, the sustainability of those biological functions can only be based on the natural environment and the “root” soil basis from which the biological elements of the BLSS derive.

The Millennium Ecosystem Assessment (2003) proposed a framework to assess how the natural environment enables human life through ecosystem services — a framework now used by scientists, policymakers and government agencies around the world to assess resource use and conservation (Millennium Ecosystem Assessment 2005; Daily and Matson 2008; Fisher et al., 2009; Primmer and Furman 2012; FAO 2021). The framework proposed four categories of ecosystem services, starting with supporting services (e.g., biogeochemical cycles, biodiversity, soil formation, photosynthesis) that then support regulating services (e.g., biological control, carbon storage sequestration, air quality, climate regulation) that then regulate provisioning services that relate to human needs (e.g., clean air to breathe) and cultural services that relate to the development of society (e.g., education, recreation, custom, commerce, governance). Without the basis of supporting services, in which soil is a critical part, ecosystems would not exist. With an estimated 95% of the food consumed by humans being produced from the soil (FAO 2015), humans are dependent on the sustainability of a soil-basis for survival. Furthermore, ecological system theory defines the sustainability of Earth’s biosphere in terms of the stability of natural cycles and processes under varying load feedbacks and disturbances (Benedetti-Cecchi 2003; García Molinos and Donohue 2011; Moreno-Mateos et al., 2017). To assess the sustainability of BLSS, we need to explore ecosystem stability in the context of a system sustaining humans (providing provisioning services) while experiencing loads and disturbances.

Just as provisioning services provided by Earth’s environment are described in terms of human needs, loads on provisioning services are specific to human activities. All other activities that naturally occur within the environment (e.g., animals feeding on plants and on other animals), are part of the environment and are not considered to be loads in the context of sustainability. Human activities include not just the loads of the humans themselves (e.g., eating food and breathing) but also the loads of the infrastructure and industry that humans build and operate (e.g., urbanization, intensive agriculture, and the resulting increase in greenhouse gas emissions that load the Earth biosphere and contribute to climate change). Infrastructure and industry are built up and expanded using supply chains. Supply chains act as artificial extensions of provisioning services to multiply the use of consumable resources and extend the use of the resources to humans in locations where the resources would not naturally be available. Thus, supply chains must then be considered as both part of the environmental processes and part of the loads in the determination of sustainability of the environment.

In ecology, a disturbance is any event “that disrupts the structure of an ecosystem, community, or population, and changes resource availability or the physical environment” (White and Pickett 1985). Disturbances can act on any parts of a system’s natural cycles (e.g., Müller et al., 2002; Tylianakis et al., 2008; Jentsch and White 2019) and its supply chains (e.g., Davis 1993; Christopher and Lee 2004; Bhatnagar and Sohal 2005; Hobbs 2020), not just the points of the natural cycles and supply chains where human consumption normally occurs. Disruptions can greatly reduce or eliminate the availability of provisioning services and the resources they provide. As artificial extensions of provisioning services and resources, supply chains introduce additional potential types and points of disturbances.

Supply chain sustainability to load and disturbance is built upon performance factors affecting, for example, a supply chain’s production flexibility, lead time, demand variability response, inventory, and overall product and service quality (Davis 1993; Levy 1995; Bhatnagar and Sohal 2005; Sreedevi and Saranga 2017). Location-specific factors include access to resources, cost, infrastructure, and geographic distance to the end-customer (Levy 1995; Bhatnagar and Sohal 2005). Relational factors are the various activities along the value chain of a product or service and the associated costs to link each into the supply chain (Levy 1995). Supply chain performance analysis has shown that the greater the quantity of performance factors, the more susceptible the supply chain is to natural disturbances such as weather events, and human disturbances such as distorted information, border closings, strikes, and socioeconomic fluctuations (e.g., Lee et al., 1997; Kleindorfer and Saad 2005; FAO et al., 2020; Hobbs 2020). Thus, supply chains with long paths and the need to reduce lead times commonly have increased risk of disturbance to meeting supply, process, and demand. With a smaller margin of error, the cost to maintain or recover efficiency in the short term increases and sustainability degrades in the long term.

The 2020 COVID-19 Pandemic almost immediately revealed how unstable our long-but-efficient supply chains are to sudden disturbances. The uncharacteristic freezing temperatures in Texas in 2021 also revealed how even localized supply chains (e.g., Texas electric grid) with minimal relational factors and shorter geographical distances can be disrupted by disturbances. The Texas cold wave of 2021 also demonstrated the catastrophic levels of failure that result when risk mitigation steps and adaptive management plans are not in place to recover or create pre-disturbance sustainability in supply chains (Blackmon 2021; Calma 2021; Krauss et al., 2021).

Engineered life support systems derive their sustainability from long, artificial supply chains that extend across Earth and out into space as the systems rely on resupply from the provisioning services of Earth to maintain operation (e.g., the ISS). An alternative to using Earth as a basis for sustainability that we will mention here (but leave detailed discussion to another paper) is to use another planetary body, such as Mars, as a basis (e.g., Kading and Straub 2015; Irons 2018; Berliner et al., 2021). For example, Irons (2018) proposes a “quasi-closed agroecological system” that utilizes ecological buffer zones, in situ resources, and ecosystem service reservoirs to establish natural cycles independent of supply chains from Earth.

It is noteworthy to consider the assumption that BLSS is inherently more sustainable than ECLSS due to the presence of bioregenerative components. For any given functional objective and the best available ECLSS and BLSS technologies that can be selected to meet the objective, sustainability will be affected by the engineered resilience of the non-regenerative parts of the system (Haimes 2009; Matelli and Goebel 2018) and the particular supply chains needed for the given technology to provision maintenance materials, repair parts, and consumable materials. The expectation is that sustainability is heavily weighted by both ECLSS/BLSS design and supply chain performance under load, and the risk mitigations, and adaptive management plans prepared for the common and unique disturbance points.

A theoretical advantage in sustainability of a BLSS over an ECLSS emerges as internal, non-regenerative components of the design of a BLSS are replaced with bioregenerative components, eliminating potential non-regenerative part failure points and reducing the need for supply chains for replacement parts. A theoretical advantage in sustainability of a BLSS over an ECLSS also emerges when the bioregenerative function is applied in a natural cycle that is more adaptive to loads and disturbances, rather than a forced cycle that has fixed steady state modes that are non-adaptive (Figure 1). When operated closer to a natural cycle, the implication is that it becomes more dependent on local, naturally renewable resources and buffering capacities and less dependent on supply chains, making it more adaptable to recovery from unplanned disruptive events (Irons 2018). Based upon this implication, operating a BLSS within a natural cycle with few or no supply chains would make it inherently more sustainable.

An area of research and ecosystem management here on Earth that is highly relevant to the question of sustainability of a BLSS is that of fragmented ecosystems. Fragmented ecosystems, such as isolated forests resulting from clear-cutting practices and segmented agricultural fields, experience constant edge effects such as wind, temperature fluctuations, and pest disturbances, that negatively affect the populations and communities that are openly exposed to those elements. The stability and quality of such systems is constantly under threat of habitat degradation, loss of connectivity to biogeochemical cycle inputs and outputs, and the loss of biodiversity and genetic movement through the system (Fahrig 2003; Hanski 2011; Wilson et al., 2016). A BLSS operated with need for a supply chain from Earth is effectively a fragmented ecosystem, considering the biological elements within the engineered system will have a large “edge” comprised of the supply chain between it and the Earth-basis ecosystem from which is derived the resources needed by the BLSS.

One example of a BLSS that could require a supply chain is a hydroponics system. In addition to the edge effect of the supply chain, hydroponics systems create edge environments along plant roots that would have otherwise been protected and buffered by a natural soil environment microbiome on Earth. An edge effect the plant may experience is exposure to pathogens that would have otherwise been kept at bay by a soil’s microbiome. A field-based agricultural system in an artificial habitation module, even if it uses natural Earth soil, will also have an edge where plantings end along the engineered boundary and would have a separate edge associated with a supply chain if fertilizer additions are required to support the load of crop production. Edge environments are more susceptible to disturbances and load feedbacks and can be a place to measure the most critical factors that will limit the sustainability of a BLSS.

Maintaining a human presence in space over the long-term using BLSS implies the system must have stability under nominal and potentially abnormal load conditions, as well as both expected and unplanned disturbance events. Stability of forced cycle engineered systems has commonly been defined as an asymptotic measure of whether a system maintains an equilibrium state under load and how quickly it returns to its equilibrium state following a disturbance or shock event (Holling 1996). However, the goal of a BLSS is to support or produce a natural cycle. Stability of natural-cycle systems over a long term will not necessarily be restricted and asymptotic. In comparison, naturally cycling ecosystems are complex networks of individual species and groups of species that interact with the biotic and abiotic environment, creating unique and redundant functions and processes (Gray et al., 2021) made up of consumer-resource (i.e. producer) interactions that can vary over time and not necessarily around a mean (Benedetti-Cecchi 2003; Inouye 2005). Perturbations of varying magnitude, duration, and frequency over different spatial and temporal scales will elicit a range of responses that may affect the ecosystem immediately or may have long term effects that are not measurable on small time scales (Sousa 1984; Müller et al., 2002; Tylianakis et al., 2008; Pincebourde et al., 2012; Donohue et al., 2016; Jentsch and White 2019).

Considering the variance of the large number of consumer-resource interactions comprising the complex network of an ecosystem, quantifying stability as a binary measure of single dimension (i.e., output is nominal or off-nominal) as is done with forced-cycle engineered systems is inaccurate and non-predictive for measuring the sustainability of a natural cycle system and can lead to ineffective management (Donohue et al., 2016; Pimm et al., 2019). The marine resources industry is a key example of how management of marine ecosystems is shifting away from focusing on single species stability and single ecosystem service management to ecosystem-based management—a management strategy that incorporates all ecosystem interactions, including those of humans, into the research and adaptive management of the system (Pikitch et al., 2004; Tam et al., 2017). The multi-dimensional approach of such a management strategy should be applied to assessing the stability of a BLSS. However, the large number of potential factors represented by all the consumer-resource, send-receive, force-action interactions that make up the operation of a BLSS can be daunting. An approach is needed to select a subset of critical factors out of all interactions.

Ecosystem control points provide such an approach. In Bernhardt et al. (2017), ecosystem control points are defined as “areas of the landscape that exert disproportionate influence on the biogeochemical behavior of the ecosystem under study.” For soil, these ecosystem control points are places where the movement of water and gases result in the transport, accumulation and export, and delivery of reactants to support permanent and activated biogeochemical activity. The concept of ecosystem control points can be applied to the process of an ecosystem service being delivered to humans by an Earth biogeochemical soil basis, a network of supply chains, and a BLSS comprised of non-regenerative parts combined with bioregenerative elements. One identifies the ecosystem control pointsto select a subset of all interactions and factors. This requires a careful analysis to identify the measurable, disproportionately influencing factors of the biogeochemical cycles and environmental processes, supply chains, non-regenerative subsystems of the BLSS, and bioregenerative subsystems of the BLSS. These disproportionately influencing factors are called critical factors.

Finally, the myriad possible ways that a system can be impacted by load feedbacks and disturbances necessitates limiting them to classes of load feedbacks and disturbances (Haimes 2009). For the BLSS schema, we recommend the following classes: non-regenerative part failure, biotic feed stock loss, pathogen and toxin introduction, mutation of biotic elements, unplanned load/leak, bioregenerative overgrowth. Classes of load feedbacks and disturbances for the ecosystem services on Earth and the associated supply chains that support the BLSS from Earth should also be utilized in the analysis. The BLSS development team should include biologists, environmental scientists, and supply chain experts to help with this analysis.

To meet the mission criterion for a sustainable human presence in space, engineered life support systems tend to be defined by six functional objectives: 1) maintain closed-loop atmospheric pressure and composition within optimal parameters, 2) manage closed-loop water cycling, storage, and wastewater; 3) produce and store food to meet crew caloric and nutrition needs, 4) manage and recycle waste, 5) generate efficient energy for the system, and 6) ensure crew safety (e.g., Eckart 1996; NASA 2017). Table 1 shows how these six functional objectives are supported by the ecosystem services network (i.e., Earth environment basis of natural cycles and the ecological services provided by such cycles, and the supply chains that extend the resulting provisioning services to the location of humans in space) and the ECLSS or BLSS that convert the provisioned resources provided by the supply chain into the resources that meet human consumer need. Table 1 also identifies the human loads and space-based infrastructure and industry loads that impact the sustainability of the whole system.

These load points are the final consumer-resource interactions in the complex network that make up the BLSS and its connection back to Earth and the soil basis. As such, the measures of resources directly consumed by humans are the dependent variable critical factors (DVCFs) of the independent variable critical factors (IVCFs) measured at the control points. The IVCFs can be impacted by feedbacks of the human loads (i.e., IVCFs pulled out of nominal by human load rates) and by disturbances (i.e. expected and unplanned events). Thus, the stability of the BLSS is dependent upon the stability of the human resource DVCF when the IVCFs are impacted. We define the sustainability of a BLSS to be the stability of the human resource DVCF produced by a BLSS in response to human load and disruption impacts on IVCFs at the environmental control points of the BLSS and its ecosystem services network. We apply this definition in our development of a Terraform Sustainability Assessment Framework.

Studies to assess the stability of ecosystem critical factors have used variability, persistence, resistance, and resilience as properties of ecological stability (Pimm 1984). Where the challenge lies in quantifying these stability properties is a lack of clear definitions of the terms and a lack in understanding of how to capture the complexity and variety of ecosystem responses to disturbances (Donohue et al., 2016; Pimm et al., 2019).

Nimmo et al. (2015) proposes quantifying resistance and resilience as R s =   X 2 / X 1 (1) R l =   X 3 / X 1 − R s (2) Resistance ( R s ) is the ratio of the minimum level to which a critical factor ( X ) drops due to a disturbance ( X 2 ) to its value prior to disturbance ( X 1 ), and defined as “the ability to maintain functional output immediately following a disturbance.” Resilience ( R l ) is the difference of the resistance ratio ( R s ) and the ratio of the value to which the critical factor recovers following the drop ( X 3 ) to its value prior to disturbance ( X 1 ), defined as “the ability to return to functional output after a disturbance has passed.” Resistance and resilience, as described here, are the stability properties associated with disturbances.

As resistance and resilience relate to disturbances, variability and persistence relate to load feedback processes. However, whereas variability is defined as how much a critical factor drops as a result of feedback from a load, we propose to use what we call consistence that we define as “the ability to maintain functional output immediately upon establishment of a load.” Persistence is “the ability to return to functional output while loading is ongoing.” Equations for consistence and persistence would then follow the same model as resistance and resilience C =   Y 2 / Y 1 (3) P =   Y 3 / Y 1 − C (4) where consistence ( C ) is the ratio of the minimum level to which a critical factor ( Y ) drops under load feedback ( Y 2 ) to its value prior to loading ( Y 1 ), and persistence ( P ) is the difference of the consistence ratio ( C ) and the ratio of the value to which the critical factor recovers following the drop while still under load ( Y 3 ) to its value prior to loading ( Y 1 ). For clarification, symbols “X” and “Y” are used to differentiate disturbance critical factors ( X ) and load feedback critical factors ( Y ). It is worth noting that a given critical factor could be impacted by both disturbances and load feedbacks.

When X and Y are selected to be the same human resource DVCF of a BLSS, the stability properties represent measures of sustainability for the BLSS, as we defined in subsection 2.7. Ecosystem control points and the associated IVCFs for the given BLSS and its classes of disturbances and load feedbacks must be selected to meet five caveats identified by Nimmo et al. (2015) for use of these equations. The theories and methods discussed in section 2 enable the BLSS developer to meet the caveats. With these theories and methods, BLSS developers can perform analyses to predict how feedbacks from human loads and impacts from disturbances uniquely and commonly affect IVCFs. BLSS developers can then perform computer simulations and physical tests to see how the classes of selected load feedbacks and disturbances affect any given IVCF (while controlling for the other IVCFs), and how the affected IVCF impacts the particular human resource DVCF produced by the BLSS. Values of the human resource DVCF prior to disturbance/loading (X1/Y1), immediately following disturbance/loading (X2/Y2), and at a later time following disturbance/loading (X3/Y3) are then used to calculate resistance, resilience, consistence, and persistence. Using this Sustainability Assessment Framework, values of these four properties of sustainability can be calculated for the worst-case scenario of each class of disturbance and load feedback for each BLSS.

The Sustainability Assessment Framework we have provided thus far quantifies sustainability of a BLSS as a measure of four properties of stability of human resource DVCFs; however, we still have not accounted for the effect of variance of consumer-resource interactions on the measurements of critical factors. We next propose a Terraform Sustainability Assessment Framework to account for the effect of variance on an assessment of BLSS stability by comparing the four stability properties we can calculate for a BLSS to those of a functionally equivalent Earth environment model under proportional load and disturbance effects. We are controlling the effect of variance on the sustainability calculations of a BLSS by normalizing the stability properties to those of an equivalent (in bioregenerative aspects) Earth environment that would also be affected by variance. Any influence of variance on the X3 and Y3 critical factors of Equations 2, 4, respectively, would be assumed to be in play in both the BLSS and Earth model environment(s), and thus, their influence on the calculations is controlled for in the normalization. The results are four new terms that we call terraform specific stabilities: terraform specific resistance ( R s T ), terraform specific resilience ( R l T ), terraform specific consistence ( C T ), and terraform specific persistence ( P T ). R s T = R s B L S S / R s E (5) R l T = R l B L S S / R l E (6) C T = C B L S S / C E (7) P T = P B L S S / P E (8) where each normalization is the ratio of the value of a BLSS stability property (e.g., R s B L S S ) to the stability property of the proportional Earth model (e.g., R s E ). We use the word terraform for the normalized properties considering the calculations are developed on the basis of Earth sustainability; thus, any BLSS that achieves terraform specific stabilities of 1 has effectively terraformed the human habitation supported by the BLSS, making it as sustainable as an equivalent system on Earth.

Finally, we recommend a formalized and adaptable approach (Figure 2) to using the Terraform Sustainability Assessment Framework for BLSS system developers to assess and quantify sustainability. The assessment can be performed on a BLSS design at any phase of development as a tool for iterative improvement of the design and associated risk and adaptive management plans. Development teams should be staffed with engineers, supply chain experts, environmental scientists, ecologists, and science specialists (e.g., soil scientists) to ensure valid application of the theories we utilized to develop this framework.

By defining terraform specific stability properties of BLSSs on the basis of proportional Earth models, we provide a tool to assist in the purposeful advancement of BLSS designs towards the ideal of no longer requiring Earth supply chains. Our hope is that previously unknown questions and development challenges will be discovered in the process of analyzing and managing the sustainability of any given BLSS using this framework. We encourage the development of case studies of BLSS designs using our proposed Terraform Sustainability Assessment Framework. Such case studies will not only help with the development of more sustainable life support for space but also the development of technologies that can be used for sustainable practices on Earth.