Life Cycle Assessment of Land Use Trade-Offs in Indoor Vertical Farming ^†

Abstract

Urban agriculture (UA) is emerging as a promising strategy for sustainable food production in response to growing environmental pressures. Indoor vertical farming (IVF), combining Controlled Environment Agriculture (CEA) with Building-Integrated Agriculture (BIA), enables efficient resource use and year-round crop cultivation in urban settings. This study assesses the environmental performance of a prospective IVF system located on a university campus in Portugal, focusing on the integration of photovoltaic (PV) energy as an alternative to the conventional electricity grid (GM). A Life Cycle Assessment (LCA) was conducted using the Environmental Footprint (EF) method and the LANCA model to account for land use and soil-related impacts. The PV-powered system demonstrated lower overall environmental impacts, with notable reductions across most impact categories, but important trade-offs with decreased soil quality. The LANCA results highlighted cultivation and packaging as key contributors to land occupation and transformation, while also revealing trade-offs associated with upstream material demands. By combining EF and LANCA, the study shows that IVF systems that are not soil-based can still impact soil quality indirectly. These findings contribute to a broader understanding of sustainability in urban farming and underscore the importance of multi-dimensional assessment approaches when evaluating emerging agricultural technologies.

1. Introduction

As the demand for food continues to rise, driven by shifting global consumption patterns and a growing population, arable land has been expanding slowly and unevenly, at a rate of 0.2% per year, accompanied by persistent land degradation worldwide [1]. To ensure food security, innovative sustainable agricultural practices and policy interventions are taking shape [2]. Advanced agricultural technologies and novel farming systems can help sustain current production levels and meet future food demands while claiming to minimize soil use [3,4].

Urban agriculture (UA) has been increasingly explored as a localized food production strategy that can contribute to addressing emerging environmental and social challenges. By repurposing underutilized urban spaces, UA reduces the need for new land use and brings food production closer to consumers, promoting also community engagement [5,6,7]. Controlled Environment Agriculture (CEA) emerges as a critical technology within UA, optimizing crop growth through the use of enclosed spaces that provide controlled microclimates, hydroponic systems, software-controlled environmental parameters, Light-Emitting Diode (LED) lighting, and nutrient dosing [8]. These controlled environments are frequently incorporated into vertical farming systems, which maximize the use of available space while mitigating the challenges posed by climate variability, effectively neutralizing it as an uncontrollable factor [9]. While vertical farms and CEA systems consume significant amounts of energy, their integration with Building-Integrated Agriculture (BIA) and the use of renewable energy can significantly reduce their environmental impact [10]. This approach, indeed, leverages urban infrastructure to power energy-intensive vertical farming operations, thereby enhancing UA through efficient and localized food production systems [4,11,12].

Beyond optimizing space and energy use, UA enables the cultivation of nutrient-rich crops, further contributing to the rising demand for nutritious food.

One such example is the production of microgreens, vegetables grown and harvested at the first leaf stage, which offer high nutritional content as a final consumer product [13]. The compact growing systems used for microgreens reduce direct land occupation, which is an important consideration in densely populated areas [11].

Despite those positive aspects, there is limited data to clearly assess, compare and validate these innovative technologies regarding two critical environmental effects. Firstly, the role of urban farms on soil health and, secondly, the role renewable energy can play in reducing the consumption of electricity from other energy sources and their effects on soils. Life Cycle Assessment (LCA) represents one of the most widely used frameworks for assessing the potential environmental impacts of these systems. It is applied to address the environmental effects of products, processes, and services throughout their entire life cycle [14]. LCA is able to cover a wide variety of pressures and impacts associated with human health, ecosystem and natural resources and enables multi-criteria assessments and, for this purpose, it is extensively applied to analyze the environmental performances of food products, including to prospectively forecast impacts of innovative agri-food systems in their initial planning stages [15]. In recent years, LCA has been applied to hydroponic systems [16,17], urban and peri-urban agriculture [18,19,20], and precision farming [21,22,23]. Moreover, several recently published studies analyze the application of LCA to vertical farms [24,25,26].

Still, most existing LCA studies on CEA do not include some critical environmental indicators related to soil health and land use, limiting their ability to fully assess the system’s potential benefits. The UNEP-SETAC Life Cycle Initiative has performed extensive work to develop and systematize impact assessment indicators for soil quality, biotic productivity and biodiversity, but analysis of land use and soil degradation remains underrepresented in LCA [27,28]. Among the land use assessment tools for LCA, the LANCA model has been identified as highly relevant and applicable, with its indicators aligning well with ecosystem service mid- and endpoints and life cycle inventory data [29]. Although the LANCA model is still in its experimental phase, it offers a comprehensive multi-indicator approach to assess land use impacts, accounting for six distinct soil ecosystem services and clearly distinguishing between the impacts resulting from land occupation and those from land transformation [30]. Regardless of its possibilities, the LANCA model remains underutilized in current studies of vertical farming and UA systems.

This study aims to evaluate the environmental performance of an indoor vertical farm (IVF) integrating CEA and BIA technologies, using a prospective analysis of a not yet operational IVF system previously analyzed by Parkes, Cubillos Tovar et al. (2022) [7] as a case study. Two energy scenarios were compared through an LCA analysis: a standard operating scenario relying on Portugal’s current grid mix and a greener alternative leveraging a rooftop solar panel to utilize photovoltaic electricity, aligned with BIA principles. To enhance the multi-criteria assessment, this study adopts the Environmental Footprint (EF) method, which evaluates 16 impact categories to provide a comprehensive analysis of environmental effects associated with transitioning to cleaner energy in UA. The EF method’s normalization and weighting capabilities yield a single, interpretable score, while its structured framework supports hotspot analysis to identify critical processes driving environmental impacts. Additionally, alongside the EF method, the study employs the LANCA model to specifically assess impacts on land use and soil quality.

This research seeks to evaluate the added value of integrating these two impact assessment methods while assessing the environmental performance of photovoltaics as an energy solution for UA. It aims to generate insights into the environmental dynamics of CEA and vertical farming systems, provide robust data for evaluating future evaluations, and identify key environmental trade-offs.

2. Materials and Methods

2.1. Case of Study Description

The IVF system, located in one of NOVA University’s campus buildings in Portugal, was at a prototype stage, namely, it was only used for testing and was not yet in commercial production. Its installation in the technical services area of the basement allowed additional points of integration with the building’s air ventilation and energy systems. The IVF system, covering a total of 46 m², comprised two main areas: a controlled environment growth chamber with a vertical soilless growing system and a separate preparation area for seeding and harvesting operations. The system utilized CEA technology, integrated with Heating, Ventilation and Air Conditioning and plant factory artificial lighting using LEDs, and was equipped with Internet-of-Things devices to optimize resource efficiency [31]. Tests made in the farm aimed to grow fresh broccoli microgreens (Brassica oleracea var. Raab), selected for their high nutritional value and suitability for this cultivation method and to test the supply either locally or within a 10 km radius. The growth cycle during the test phase lasted 14 days, with LED lights operating for 12 h/day and the climate control system running continuously. Proposed production varied from 5.28 kg to 7.28 kg depending on management of farms, and the annual goal was 1900 kg over 360 days. After each harvest, the plants were packed and delivered, followed by a cleaning process to prepare the area for the next cycle. A more comprehensive description of this IVF installation can be found in the Supplementary Materials, which is based on Parkes, Cubillos Tovar et al. (2022) [7] and Parkes et al. (2023) [32].

2.2. Scenarios

The combination of technologies used to operate any IVF system requires a high energy demand for these climate-controlled farms [33,34]. This includes CEA components responsible for regulating environmental conditions, such as heating or cooling, as well as those responsible for LED lighting and hydroponics, which circulate flows of water and nutrients within the system. Due to the prominent role of energy in the sustainability of IVF systems using CEA technique, two scenarios of different electricity sources were evaluated: one utilizing the Portuguese national grid mix, and another incorporating a predominant solar energy mix through the installation of a photovoltaic system. These scenarios are referred to as GM (Grid Mix) and PV (PhotoVoltaic mix). The PV scenario aimed to replicate a prospective situation in which the electricity was a mix coming from both the Portuguese national grid (about 30%) and a non-ground-mounted photovoltaic system (70%).

Both scenarios include the construction of the chamber, its transportation to the operating location, the assembly of the IVF system, and the processes needed for cultivation and harvest. Packaging occurring on-site and transport to deliver the final product to retailers or consumers within a 10 km radius were also included in the evaluation.

2.3. Life Cycle Assessment

The LCA procedure used complies with the European Commission ILCD Handbook [35] and the Product Environmental Footprint (PEF) guidelines [36].

2.3.1. Goal and Scope

This study compared the two scenarios outlined in Section 2.2 to evaluate the environmental impacts of the IVF system described in Section 2.1, using two impact assessment models: the EF v.3.1 and LANCA v.2023.1, and their respective 16 (EF) + 6 (LANCA) impact categories with a particular emphasis on land use effects. The goal was to identify system hotspots in the two scenarios and support improvements to the production method.

The functional unit was 1 kg of fresh weight microgreens produced, packed and delivered daily, modeled as the average production over a 12-month period. The system boundaries were set from grower to grocer, considering the processes involved from the IVF construction to the broccoli final distribution, excluding post-delivery activities (Figure 1).

2.3.2. Life Cycle Inventory

This study was built upon the inventory used by Parkes, Cubillos Tovar et al. (2022) [7], incorporating modifications and updated data. The infrastructure, equipment (including the photovoltaic system), consumables, water, and nutrients remained consistent with the original study. Life Cycle Inventory activity data were collected either on site or through estimates provided by facility operators and modeled with LCA for Experts v.10.9.0.20. The databases used to access background data were a combination of Sphera and ecoinvent v.3.9.1, supplemented by Agribalyse v.3.0.1. Assumptions and limitations regarding these processes, along with detailed lists of their specific inputs, are provided in the Supplementary Material. Details of the modified processes and updates to data inputs compared to the previous work are given hereafter.

The energy mix reported in the inventory represents the average production mix used in Portugal from January 2024, which consisted of 5.0% natural gas, 3.8% fossil Combined Heat and Power, 38% hydro, 28% wind, 6.7% bioenergy, 10% solar, and 7.9% pumped storage, according to the Portuguese National Energy Association [37]. As indicated in the ecoinvent database, there was a 4.4% loss taken into consideration for the non-photovoltaic energy vectors as a result of the conversion from high voltage to low voltage [37].

The “GLO: market for rape seed, organic” dataset from ecoinvent was chosen as the most appropriate option for modeling the seeds production due to its similarity to broccoli, as both belong to the Brassica plant family. This specific input reflects the Brassica napus species. This dataset replaced the previous model’s dataset, “Cauliflower seed, conventional, at production site/FR U” from Agribalyse, which had been rebuilt on ecoinvent using equivalent flows. Despite cauliflower also belonging to the same plant family, the earlier study’s dataset replication used equations to address the lack of an input flow, introducing additional variability to the model. By opting for the rape seed dataset, this study aims to reduce uncertainty associated with seeds, which were identified as the most significant input from the environmental point of view in the previous paper.

The “Coconut fiber, at regional storehouse/kg” dataset was selected for modeling substrate production. The dataset was copied from Agribalyse and replicated on ecoinvent introducing a key modification: coconut fiber was reclassified as a by-product. This reclassification shifts the environmental burden associated with its production to the coconut, unlike in the previous model. However, the processes involving husk transformation and transportation were still included in the assessment.

The organic waste produced by the harvesting process was counted as “AT: Open windrow composting (incl. compost application and crediting) BOKU/Sphera”, while in the previous study it had been allocated either to generic municipal solid waste or as a circular, locally used byproduct for compost production.

Input flows to account for the direct land use were considered. The chosen flows were “Urban, continuously built (regionalized, PT) [Occupation]”, applied to both the preparation and work areas.

2.3.3. Life Cycle Impact Assessment

As previously introduced, this study used two Life Cycle Impact Assessment methods: EF v.3.1 and LANCA v.2023.1. This dual approach enabled a more in-depth evaluation of land use impacts specific to the IVF system, while also considering the system’s environmental effects on other impact categories. The EF impact assessment method includes a Land Use category indicator based on the LANCA model, but it provides only a single aggregated impact value, lacking specificity in addressing the various soil functions directly impacted by the processes [38,39]. Given this specific focus on differentiated land use impacts, the full LANCA model was deemed appropriate, allowing for a detailed assessment of environmental impacts across various soil ecosystem services. Indeed, the LANCA model adopts a multi-indicator approach to assess land use impacts by evaluating six key soil ecosystem services: erosion resistance, mechanical filtration, physicochemical filtration, groundwater regeneration, biotic production and soil organic carbon [30]. It also establishes a primary distinction between land use indicators, differentiating between “Occupation” and “Reversible transformation”. Occupation refers to the condition of using a specific land area, which is considered static for a certain duration. Under this condition, the level of ecosystem quality is expressed relative to a defined reference quality. Conversely, transformation identifies the variation in ecosystem quality between the pre-use phase and the post-use phase, when the soil has the potential for regeneration [30]. While occupation categories deal with long-term or continuous impacts during land use, Reversible transformation evaluates short-term, recoverable changes to land characteristics.

Results from the characterization phase of the EF were normalized to ensure a fair comparison across its different impact categories. This was achieved by dividing the category values by specific normalization factors defined for the EF v.3.1 impact assessment method. The dimensionless and standardized results enable comparisons among categories that originally have different units of measurement and facilitate the interpretation of environmental impacts. Normalized results were weighted as proposed by the EF method [36,40]. The weights are used to calculate a single score for each impact category, which can be compared to the overall score of other product systems. Finally, a hotspot analysis was carried out following the PEF guidelines [36], considering only those categories that, when added together, contribute at least 80% to the overall impact of the system [40].

3. Results and Discussion

In this section, results are presented and discussed in detail, highlighting their significance, comparisons with previous studies, and potential limitations. The two scenarios, GM and PV, have been evaluated with the EF v.3.1 impact assessment method first, then with LANCA model for land use. Specific normalization and weighting factors provided by the PEF guidelines were applied to the EF v.3.1 impact scores to perform the hotspot analysis. Subsequently, the scores were aggregated to obtain a result that is more easily communicable [36].

3.1. Environmental Footprint Results

Characterization impact scores of each impact category are reported in

Table 1. Characterization results on the sixteen impact categories of EF v.3.1.

Categories	Unit	GM	PV
Acidification (AC)	Mole of H+ eq.	4.85 × 10−2	4.07 × 10−2
Climate Change—total (CC)	kg CO2 eq.	7.11 × 100	5.20 × 100
Ecotoxicity, freshwater—total (EF)	CTUe	7.50 × 101	7.12 × 101
Eutrophication, freshwater (EuF)	kg P eq.	2.05 × 10−3	1.79 × 10−3
Eutrophication, marine (EuM)	kg N eq.	1.08 × 10−2	9.06 × 10−3
Eutrophication, terrestrial (EuT)	Mole of N eq.	1.21 × 10−1	9.95 × 10−2
Human toxicity, cancer—total (HTc)	CTUh	1.03 × 10−8	9.14 × 10−9
Human toxicity, non-cancer—total (HTnc)	CTUh	2.42 × 10−7	2.24 × 10−7
Ionizing radiation, human health (IR)	kBq U235 eq.	6.44 × 10−1	5.44 × 10−1
Land Use (LU)	Pt	1.15 × 102	5.70 × 101
Ozone depletion (OD)	kg CFC-11 eq.	3.83 × 10−7	3.29 × 10−7
Particulate matter (PM)	Disease incidences	3.68 × 10−7	3.08 × 10−7
Photochemical ozone formation, human health (POF)	kg NMVOC eq.	3.36 × 10−2	2.71 × 10−2
Resource use, fossils (RUf)	MJ	9.63 × 101	7.28 × 101
Resource use, mineral and metals (RUm)	kg Sb eq.	6.65 × 10−5	5.98 × 10−5
Water use (WU)	m3 world equiv.	1.58 × 101	6.43 × 100

. It shows that the GM scenario consistently yields higher environmental impact scores than the PV scenario across all 16 EF categories for the IVF producing microgreens. Relative differences in impacts between the PV and GM scenarios range from 5% lower in the PV scenario for Freshwater Ecotoxicity (FE) to 59% for Water Use (WU), with an average difference of approximately 20%. Notable lower impacts are observed in the PV scenario for Land Use (LU; 50%), Climate Change (CC; 27%), and Resource Use, fossils (RUf; 24%), primarily due to the GM scenario’s reliance on fossil-based electricity. Human Toxicity, cancer (HTc; 11% lower) and Human Toxicity, non-cancer (HTnc; 7% lower) also show smaller differences. In four of these categories (FE, HTc, HTnc, RUf), however, the savings from adopting the PV scenario over the GM scenario amount to approximately 10%.

Among the standout categories, LU shows a significant difference, with the PV scenario exhibiting nearly half the impact of the GM scenario (PV = 5.7 × 10¹ Pt vs. GM = 1.15 × 10² Pt), reflecting the spatial efficiency of BIA systems employing IVF producing microgreens. This result aligns with the PV scenario’s use of a non-ground-mounted solar system, as assumed in the study design, which limits direct land occupation compared to traditional ground-mounted systems. These results are consistent with those obtained in two very recent studies that highlight the advantages of applying photovoltaic technology to urban agriculture. The study by Jing et al. (2022) [41] proposes the co-location of photovoltaic panels and vegetable gardens on urban rooftops as a solution to achieve clean energy production together with “zero-kilometer” agricultural output. By integrating a Geographic Information System (GIS) with modeling simulations, the study estimates the annual electricity production and the amount of horticultural products that could simultaneously be obtained in the city of Shenzhen, China, yielding interesting results. In an even more recent article by Sobuj et al. (2024) [42], it is highlighted how the high electricity demand of vertical farming systems can be addressed through the use of renewable energy sources (solar and wind) and more efficient electrical equipment. However, neither of the two cited studies applies an LCA to the proposed systems and therefore do not quantify the benefits in terms of avoided environmental impacts resulting from the use of photovoltaic technology in IVF systems.

These findings reinforce the well-documented environmental advantages of photovoltaic systems in both land use and climate change impact categories. For land use, rooftop-mounted PV avoids the need for ground-mounted infrastructure, enabling electricity generation without additional land transformation. Moreover, the compact spatial footprint of PV infrastructure contrasts sharply with the indirect land occupation and ecosystem disruption linked to fossil-based energy sources. Regarding climate change, PV electricity production is virtually emission-free during operation, significantly reducing lifecycle greenhouse gas emissions compared to grid mix scenarios reliant on fossil fuels. As seen in this study, the PV scenario reduced CC impacts by 27%, which can be directly attributed to the substitution of fossil energy inputs with solar-based electricity for high-demand processes such as LED lighting. These operational benefits are amplified when considering the avoided upstream emissions from fossil extraction, transport, and combustion, further confirming the systemic climate advantages of photovoltaic integration into IVF systems.

Figure 2 shows the relative contributions of the five processes (Construction, Installation, Cultivation, Packaging, and Delivery) to each impact category in both scenarios. As explained in Section 2.2, the two scenarios differed primarily in terms of the electricity source.

By comparing the relative contribution of each process, most impact categories show a similar trend across both scenarios. The main difference is seen in LU, where Packaging accounts for a much larger share in the PV scenario (about 80%) compared to the GM scenario (around 39%). Although Cultivation maintains the main contributor in both cases, its percentage is lower in the PV scenario. This is because Cultivation is energy-intensive, and the PV scenario uses more renewable energy, which reduces its impact. As a result, with Cultivation contributing less, other processes like Packaging represent a higher share of the total impact in the PV scenario.

The other impact categories where Cultivation was not the major contributor were Ozone Depletion (OD) and Resource Use, mineral and metals (RUm). Across both scenarios, OD was mainly affected by Packaging while RUm was predominantly influenced by Construction. In this study, the Packaging contribution to OD came from the production of polyethylene terephthalate. This result aligns with the observations of Morales-Méndez & Silva-Rodríguez (2018) [43], which indicate that ozone depletion, during plastic bag manufacturing, is strongly correlated with the production of low-density polyethylene and other polyethylene-based materials. On the other hand, the Construction phase contributed substantially to RUm due to the intensive material demands involved in setting up an IVF system. These systems, in fact, typically require large quantities of metals and high-tech materials for lighting, climate control, structural framing, and hydroponic equipment. This aligns with findings from Dorr et al. (2021) [44], who highlighted the construction of controlled environments as a major environmental hotspot due to the embodied impacts of materials and components.

Due to its significance in these results, Cultivation was broken down into its main percentage contributions in Figure 3. Sub-processes contributing less than 1% in all categories, such as the production and use of trays and sensors, were excluded from the figure. The highlighted sub-processes are as follows: Cleaning procedures, Equipment electricity consumption, LED lighting, Clothing for workers, Substrate production and transportation, Seeds production and transportation, Water consumption, NPK fertilizers production, Treatment equipment production and Composting.

Figure 3 shows that LED lighting dominates contributions to energy-related impact categories such as CC, (70–80%); RUf, (60–70%); and WU (60%), driven by the energy intensive nature of microgreen production in the IVF system. This outcome reflects the work of Martin & Molin (2019) [45], where they discuss the high energy demands and associated environmental costs in urban farming systems. The PV scenario shows reduced contributions in these categories compared to GM, reflecting the benefits of renewable energy (e.g., 27% reduction in CC, Table 1), while material-related sub-processes like Equipment electricity and Treatment Equipment contribute significantly to HTc, HTnc and RUm in both scenarios. Notably, Composting showed significant negative contributions (i.e., avoided impacts) in the LU category, highlighting waste reduction benefits. The benefits of composting can be supported by findings from the literature outlining its advantages in sustainable waste management systems [46,47].

In accordance with the PEF methodology, a hotspot analysis was carried out to identify the most relevant impact categories contributing to the overall environmental performance. This analysis is based on the normalized, weighted, and aggregated results, with the aim of determining which impact categories account for at least 80% of the total score. The total scores were 8.36 × 10⁻⁴ for the GM scenario and 6.26 × 10⁻⁴ for the PV one. The hotspot analysis for both scenarios is reported in Figure 4, where the most relevant categories were CC, accounting for 24% (GM) and 23% (PV), followed by RUf, 15% for both scenarios, WU, 14% (GM) and 8% (PV), and RUm (9% and 11%, respectively). In order to arrive to 80%, Particular Matter (PM; 7% for both scenarios), Acidification (AC, 6% and 7%, respectively), Photochemical Ozone Formation (POF; 5% for both scenarios) and Eutrophication Freshwater (EuF; 5% for the PV scenario) must also be added.

The results show differences in the percentage contributions to the total impact score between the two scenarios. While in the GM scenario the three main impacted categories are, in order, CC, RUf and WU, in the PV scenario, they are CC, RUf and RUm. Interestingly, the contribution of WU to the total impact decreases by 4% when moving from the GM to the PV scenario, highlighting the potential benefits of photovoltaics’ energy use to improve one of the weaknesses of IVF systems, as in their water consumption [5,41]. In the GM scenario, in fact, hydroelectric reservoirs contribute significantly to water consumption, mainly due to pronounced evaporation losses and the operational dynamics of hydroelectric generation [48,49]. By contrast, in the PV scenario, the contribution of the RUm category is the third most significant, accounting for 11%, which is influenced by the upstream production of photovoltaic components and associated infrastructure [50,51].

When examining the contributions of the top four impacted categories, the highest relative impacts were from electricity related sub-processes including Electric equipment, LED lighting, and Packaging, followed by Substrate production. In the CC category, electricity-related sub-processes accounted for 57% (GM) and 42% (PV) of the impact, while the substrate production contributed 12% (GM) and 17% (PV). The results of this study align with existing research, highlighting the high energy demands of IVF and CEA systems. The study from Barbosa et al., (2015) [52] reports substantial electricity use in indoor hydroponic systems for lighting and water circulation, while Graamans et al., (2018) [53] emphasize the heavy reliance on artificial lighting in plant factories compared to greenhouses. Moreover, Benke & Tomkins, (2017) [5] confirm that lighting and climate control are the most energy-intensive processes in vertical farming.

Substrate production, as the second highest sub-process, reflects the research of Grasselly et al., (2009) [54] who conducted an LCA on coconut fiber substrates produced in Sri Lanka, identifying emissions of 0.4 kg CO₂ eq./kg of substrate, excluding transcontinental shipment impacts. Similar studies on substrates for UA report values of 1.75 kg CO₂ eq./kg [17] and 0.4 kg CO₂ eq./kg [55]. The latter aligns with this study, where 3.5 kg of coconut fiber per FU contributed 0.9 kg CO₂ eq., corresponding to 0.3 kg CO₂ eq. per kilogram of substrate, reflecting the 12% (GM) and 17% (PV) impact found in this hotspot analysis.

3.2. LANCA Results

Table 2. LANCA impact assessment results.

LANCA v2023.1	Indicators	Unit	GM	PV
Occupation	Biodiversity Loss Potential	PBR	5.90 × 101	3.54 × 101
	Erosion Potential	kg	2.21 × 10−1	5.01 × 10−1
	Groundwater Regeneration Reduction Potential	m3	−7.03 × 10−3	−3.24 × 10−3
	Infiltration Reduction Potential	m3	1.10 × 101	6.70 × 100
	Physicochemical Filtration Reduction Potential	mol·a	2.48 × 101	1.51 × 101
	Soil Organic Carbon Reduction Potential	kg	1.55 × 10−1	−1.02 × 10−1
Reversible transformation	Biodiversity Loss Potential	PBR	6.51 × 100	5.88 × 100
	Erosion Potential	kg	1.22 × 100	1.52 × 100
	Groundwater Regeneration Reduction Potential	m3	−4.11 × 10−3	−1.86 × 10−3
	Infiltration Reduction Potential	m3	5.20 × 100	2.47 × 100
	Physicochemical Filtration Reduction Potential	mol·a	1.42 × 101	6.79 × 100
	Soil Organic Carbon Reduction Potential	kg	1.97 × 10−1	3.27 × 10−1

shows the results of the LANCA model application to GM and PV scenarios. Potential impacts are shown across the six indicators within two impact categories grouped by land occupation (upper section ‘Occupation’) and land transformation (lower section ‘Reversible transformation’); for each pair, the highest impact between the two scenarios is highlighted in bold.

As can be seen, the GM scenario exhibited higher Occupation impacts in four out of the six LANCA indicators and three out of the six for Reversible transformation. Consequently, the PV scenario showed greater impacts in two Occupation impact indicators and three for Reversible transformation. These insights align with the EF results, where the PV scenario showed a reduction in LU impacts compared to GM (Table 1).

The GM scenario exhibits the highest environmental impacts in Biodiversity Loss Potential (BLP), Infiltration Reduction Potential (IRP) and Physicochemical Filtration Reduction Potential (PFRP). Significant contributions across these indicators come from the processes of power, heat and co-generation and open-ground photovoltaic systems for energy consumption, seed production and transportation, and carton box production for packaging. Power, heat and co-generation stands out as a dominant source of impact, contributing significantly to land occupation, accounting for 49% of BLP, 25% of IRP, and 20% of PFRP impacts. Open-ground photovoltaic installations also emerge as critical contributors, particularly to land transformation, where they represent around 60% of impacts in both IRP and PFRP, and around 20–25% for land occupation in the same indicators. Carton box production contributes consistently across indicators, especially for land occupation (18% for IRP, 14% for PFRP, and 35% for BLP), with further contributions to land transformation (23%). Seed production and transportation, while specific to BLP, is the primary driver of land transformation in that category, accounting for 59% of the impact.

While the PV scenario substantially reduces land use impacts through the rooftop installation of photovoltaic systems and emission-free electricity during operation, it is important to acknowledge the upstream impacts as revealed by the LANCA results. In particular, the PV scenario shows notably higher values for Erosion Potential (EP) (0.50 kg vs. 0.22 kg in GM) and Groundwater Regeneration Reduction Potential (GRRP) for both Occupation and Reversible transformation (−3.24 × 10⁻³ m³ and −1.86 × 10⁻³ m³, respectively, compared to −7.03 × 10⁻³ m³ and −4.11 × 10⁻³ m³ in GM). These values indicate that while rooftop PV reduces operational land pressure, its component manufacturing and material sourcing can contribute to adverse effects on soil stability and groundwater recharge. Other major contributors to these impacts are seed production and transportation, followed by carton box production. Seed production is the primary driver in both EP and GRRP, accounting for 152% of land occupation and 112% of land transformation in EP, and 289% of land occupation in GRRP. In GRRP for land transformation, the carton box process leads, contributing 60% of the total impact.

For the Soil Organic Carbon Reduction Potential (SOCRP) indicator, the GM scenario shows greater contributions for land occupation, driven by seed production (195%) and photovoltaic electricity production from open-ground systems (150%). Conversely, PV exhibits clear advantages in SOCRP, showing a beneficial negative value under occupation, suggesting that processes, including composting, may contribute to soil organic storage. For land transformation in this same indicator, the PV scenario is majorly influenced by the seed process, which contributes 110% of the impact. These findings emphasize that although PV systems offer significant benefits in minimizing operational land occupation and climate change impacts, a full life cycle perspective is essential.

It is important to note that contributions exceeding 100% occur when other processes in the system yield negative impacts, often referred to as avoided burdens. These compensatory effects are discussed in more detail below.

Figure 5 below presents the relative contributions to each category from the main system processes: Construction, Installation, Cultivation, Packaging and Delivery. Here, the relative shares are normalized to the total value of the respective indicator.

As can be seen, Cultivation emerged as the primary contributor across nearly all categories in both GM and PV scenarios. It accounted for the vast majority of impacts on the Reversible transformation categories, particularly in EP and SOCRP, demonstrating almost total influence. Packaging also contributed significantly to the BLP of land occupation, where in the PV scenario it exceeded the contribution of Cultivation, representing 58% of the total relative impact in this category. Delving deeper into the analysis, the negative impact on biodiversity is primarily linked to the carton box production used for packaging. Although outcomes may vary depending on management practices, producing cartons from forest fiber extraction can result in species displacement, habitat fragmentation, and ecosystem loss [56,57]. Interestingly, the same process also contributes positively (negative values in Figure 5 representing environmental benefits or avoided impacts) to the EP and GRRP indicators for land occupation. Those benefits refer to avoided impacts associated with packaging. These benefits arise from the origin of carton boxes in forestry, which can enhance soil stability, support groundwater recharge, and reduce erosion. This observation highlights a trade-off: while carton box production contributes to environmental burdens through material use and emissions, it simultaneously delivers environmental benefits due to its potential positive interactions with ecosystem services.

Regarding the avoided impacts on GRRP sourced from Cultivation, they are largely attributed to the composting process (relative contributions of Cultivation processes are reported in Figure 6). Compost production enhances soil organic carbon levels, which also explains the negative contribution of this process to the SOCRP indicator. This finding is consistent with the EF results, where composting also shows negative values in the LU category (Figure 3). Additional beneficial effects on GRRP, specifically in terms of reversible land transformation, originate from various processes, as illustrated in Figure 6. The negative values for this indicator, in fact, show contributions from the seeds production and transportation process and the electricity-related processes, such as electricity use for equipment and LED lighting. The predominance of these processes is evident across most indicators.

The environmental benefits from the electricity-demanding sub-processes are linked to electricity production, particularly to the hydropower component. These benefits may be associated with the role of reservoirs in such systems, which store water and promote infiltration into surrounding soils and aquifers, enhancing groundwater recharge [58].

Conversely, the role of seed production and transportation emerges as dominant across several LANCA categories, both in terms of soil occupation and Reversible transformation impacts. The significance of the seed process within the LANCA model may be attributed to the extensive agricultural land requirements associated with rapeseed cultivation, in line with recent findings [59]. This cultivation, in fact, involves soil occupation over growing cycles and potential indirect land use changes, such as deforestation or the conversion of natural ecosystems to agricultural land. Additionally, intensive farming practices, often required for rapeseed cultivation, could contribute to increased ecosystem pressure, as highlighted by Ma et al. (2019) [60] and Markowicz (2023) [61]. Nonetheless, environmental benefits can also be associated with seed cultivation due to the deep-rooted system characteristic of rapeseed plants [62]. This root system, in fact, improves soil structure and enhances water infiltration into the ground, aiding groundwater recharge. Moreover, their drought tolerance and ability to access water stored in deeper soil layers reduce the risks of water evaporation and runoff, leaving more water available for aquifer recharge [63,64].

In addition to the contributions from electricity, seed production and transportation, and the waste composting process, other relevant processes within Cultivation, also presented in Figure 6, include the cleaning procedure, clothing production for workers, the substrate production and transportation.

The high impacts on BLP, IRP and PFRP categories are primarily associated with energy consumption from non-renewable sources due to their contribution to emissions, habitat disruption, and resource extraction, all of which negatively affect biodiversity and soil functionality. Specifically, open-ground photovoltaic systems require significant land occupation throughout their lifecycle, directly affecting ecosystems despite their low operational emissions [65]. Similarly, wood chip co-generation for electricity production has been shown to drive substantial land use impacts. This is largely attributed to forestry practices that involve land occupation and carbon stock alterations from logging activities, which may compromise both biodiversity and soil properties [56,66,67]. Equally, energy biomass crops can negatively affect soil and water quality [68,69,70]. Moreover, such energy-intensive processes often necessitate infrastructural development and land use modifications, which may interfere with natural soil infiltration and filtration capacities, thereby further altering hydrological and ecological balance.

3.3. Recommendations for Stakeholders

Based on the first key result highlighted in this study, namely the overall reduction in environmental impacts when replacing grid mix electricity with electricity from rooftop PV systems during the use phase, it is possible to recommend that stakeholders adopt this technology wherever economically and technically feasible.

Regarding economic feasibility, a cost–benefit analysis for GM and PV systems was outside the scope of this study. Moreover, feasibility is highly dependent on factors such as the location of the case study area, the scale of the operation, the specific technology employed and the type of cultivated crop.

In terms of technical feasibility, technological innovations in this field are numerous and rapidly evolving. Photovoltaic panels are no longer limited to traditional rooftop installations; they can now be integrated into building façades, balconies, shading systems, and even windows through semi-transparent modules. This expanded range of applications allows for greater architectural flexibility and enables solar energy generation in urban environments with limited rooftop space. In general, both scenarios showed the minimal contribution from the construction phase to environmental impacts. This can largely be attributed to the implementation of IVF within existing buildings, which significantly reduces the need for additional material and energy, avoiding, then, related emissions. As such, the use of existing buildings appears to be particularly effective for UA. In many urban areas, abandoned or underutilized buildings, after losing their original function, could be repurposed for urban farming. This approach would not only reduce environmental impacts but also offer social and urban planning benefits. Another recommendation stems from the minimal environmental impact observed in the delivery phase. A key advantage of UA is the significant reduction in transport distances between producers and consumers, which leads to improved air quality and associated health benefits. It is therefore advisable to promote urban agriculture that is both located in and serves the city, with the goal of minimizing transport-related emissions.

Focusing on specific sub-processes, composting emerges as particularly promising due to its ability to prevent emissions to air and associated environmental impacts. Current European regulations do not allow the disposal of organic waste in landfills, making material recovery the only viable management option. A forward-looking recommendation is to promote integrated strategies that combine anaerobic digestion with composting. This dual approach enables the recovery of both energy, through biogas, and valuable organic matter, enhancing overall resource efficiency and contributing to a more circular waste management system.

From a technical, rather than a policy or management perspective, some directions deserve attention. The role of the substrate, specifically coconut fiber, appears particularly significant in terms of environmental impact. Coconut fiber is typically imported from distant tropical regions, which means that its production and especially its transport contribute considerably to the overall sustainability of the system. This underscores the importance of identifying alternative substrates with lower overall impacts. If such materials are locally sourced, or at least originate from geographically closer areas, they could support local supply chains and promote region-specific solutions for urban farming systems.

3.4. Limitations and Future Studies

This study was subjected to several limitations that must be acknowledged. Firstly, the development of a reliable and detailed Life Cycle Inventory required the integration of data from three different databases. This integration may introduce uncertainties and potential errors, as each database is structured and developed differently. The Inventory was also compiled for a prospective technology as the unit is not yet fully operational, and therefore some activity data may differ once the farm begins production. Additionally, land use effects are notably local and can vary substantially even within relatively small regions. Life Cycle Inventories are not yet fully regionalized and, even when land use flows are regionalized, they are represented as country level averages, which may be insufficient detail. Further, the LANCA model still requires further refinement, and some limitations may affect the reliability of the aggregation indices used in the model. The most significant limitation is its reliance on generalized characterization factors (CFs) based on either global or country-level averages, which may lead to inaccuracies in varied ecological contexts [29,39,71,72,73].

Additional studies are needed to implement the LCA to more IVFs and enable a proper comparison among studies to assess the effectiveness of the sustainability of said systems. These may include different substrates and crops to assess the suitability to be cultivated in such systems. Future research might also develop a Life Cycle Costing analysis to evaluate the suitability and performance of a photovoltaic system, included in this study as a potential prospective technology.

While this study focuses on the role of energy use on the performance of a single IVF system, it does not include comparisons with conventional farming. Such comparisons would require different methodological assumptions, system boundaries, and life cycle data for conventional products replaced by microgreens grown in this specific IVF. Product replacement, and therefore their selection for comparison, has to be determined on a case-by-case basis as it depends on the study area location, the season and time of the year, and market availability. As the system is not yet being commercially exploited, we were unable to determine the replaced products, preventing any comparisons. Nonetheless, these comparisons remain important and should be pursued in future research to fully contextualize the environmental performance of vertical farming systems.

4. Conclusions

This study used the LANCA model to complement the EF method and evaluate the environmental impacts of microgreen production within an IVF system, with a particular focus on land use effects. The analysis compared systems implementing either building-integrated photovoltaics (PV) or a conventional grid electricity mix (GM). While the PV scenario results showed lower environmental impacts in most EF categories, mainly due to the use of renewable energy, the LANCA model revealed trade-offs related to land use, especially for ecosystem services such as erosion and infiltration reduction. The cultivation phase, dominated by energy-intensive processes, played a major role in the reduced impacts observed in the PV scenario. However, material-related inputs, such as substrate and packaging, also contributed noticeably to categories like climate change and ozone depletion. Together, cultivation and packaging emerged as central drivers of both energy- and land-related environmental effects. The hotspot analysis further indicated that in addition to climate change, categories associated with the use of renewable and non-renewable materials were also among the most affected. These outcomes underscore the unique contribution of LANCA in capturing function-oriented soil impacts that are often missed in LCA studies of urban farming. This research shows that although IVF systems do not rely directly on soil, their upstream and indirect effects on soil quality remain relevant and should not be ignored. By incorporating LANCA, this study offers a more complete picture of how production systems influence land-based environmental functions, contributing to a more balanced sustainability assessment. This work suggests that policy-makers should implement mechanisms for future UA investments that are powered by renewable energy but implement policies for PV systems during construction and disposal that favor materials that mitigate the negative effects identified here. Other decision-makers such as IVF managers should not assume that because IVF systems reduce the space required to grow food they are immune to indirect effects on soils. The selection of substrate, for example, is an often overlooked factor that deserves more attention. Future work should apply this combined approach across diverse IVF setups and locations, assess alternative substrates and crops, and integrate economic metrics such as life cycle costing to further analyze the implementation of renewable energy infrastructure. Broader evaluations that include food security and socioeconomic factors will also be necessary to fully understand the sustainability of urban agriculture systems using controlled environments.