Life Cycle Assessment in Protected Agriculture: Where Are We Now, and Where Should We Go Next?

Abstract

Researchers and practitioners use life cycle assessment (LCA) as a powerful tool to thoroughly assess the environmental impact of protected agriculture. However, the literature in this field has shown heterogeneity, which is characterized by inconsistent methodologies and assumptions. Identifying prevailing trends and resolving existing limitations is necessary to generate robust results and guide future work. Here, we conduct a bibliometric and systematic review to explore how LCA applications have addressed protected agriculture. The bibliometric analysis unveils trends in scientific productivity, spanning temporal evolution and geographic distribution, while also identifying prominent research avenues. The systematic review traces the historical trajectory of agricultural LCA and scrutinizes methodological decisions across the standard LCA phases: (i) objective and scope, (ii) life cycle inventory, (iii) impact assessment, and (iv) interpretation. We summarize and discuss the reported environmentally friendly practices and provide a qualitative interpretation of the LCA findings. Moreover, we pinpoint key methodological challenges and propose research horizons. It is crucial to note that the environmental benefits of protected agriculture are context-dependent, with climate change emerging as a critical factor influencing crop yields and the system’s input and output resources. This impact is particularly pronounced in terms of water and energy consumption and carbon emissions. In regions with extreme climates, protected agriculture provides solutions for producers aiming to attain high yields of top-quality crops. The integration of circular bioeconomy strategies in this context allows mitigation of the environmental trade-offs identified by LCA.

1. Introduction

Food production within greenhouse structures, commonly referred to as protected agriculture, has emerged as a pivotal element in modern agriculture, making substantial contributions to global food security [1]. Protected agriculture ranks among the most intensified approaches to food production, measurable through various metrics such as material inputs, yields per unit area, energy consumption, greenhouse gas emissions, and cost [2,3]. This agricultural technique has spurred substantial technological advancements, marked by the utilization of protective structures and air-conditioning equipment with varying technological sophistication. These components collectively enable the conditioning and management of the microclimate, maintaining optimal conditions for the growth and development of plants [4]. The adoption of protected agriculture has facilitated the maximization of both the quantity and quality of harvested products, efficiently optimizing water and soil resources and ensuring a consistent, year-round food supply [5,6].

It is vital to recognize that protected agriculture, while offering numerous advantages, has also brought about challenges associated with increased consumption of natural resources and energy. This heightened resource usage contributes to a negative environmental footprint, raising sustainability concerns for productive systems [7]. Reports indicate substantial consumption of construction materials, encompassing both the greenhouse structure and the materials used for enclosure and covering. This consumption poses notable environmental implications, with the manufacturing and transportation of these materials contributing to elevated energy consumption. Additionally, there is a risk of excessive plastic waste and increased demand for fossil fuels to power air-conditioning systems, potentially leading to water source contamination [8]. The excessive use of chemical substances for pest and disease control and plant nutrition adds a high toxicity load. Authors have highlighted concerns about landscape deterioration due to the construction of extensive greenhouse clusters globally [9,10]. The coexistence of benefits and environmental impacts emphasizes the necessity for a meticulous evaluation and optimization of protected agriculture practices to ensure a more sustainable and environmentally responsible approach to food production [11].

In the quest to understand and mitigate the environmental implications of protected agriculture, the use of the life cycle assessment (LCA) tool has gained considerable traction. LCA, a widely employed environmental management technique [12], standardized by the International Organization for Standardization, operates in four iterative phases: defining objectives and scope, conducting inventory analysis, evaluating impacts, and interpreting results [13]. LCA facilitates the quantification of environmental loads across the life cycle phases of products and services, offering valuable insights for environmentally conscious decision making [14]. Despite the increasing adoption of LCA, a lack of uniformity persists in its application across diverse research endeavors [15]. This lack of consistency poses reliability challenges, potentially skewing statistical outcomes. The key factors of divergence include variations in databases and non-compliance with required standards throughout different phases [16,17].

In the specific context of protected agriculture, variability in the application of LCA has not been explored. Shedding light on this aspect is pivotal for promoting a more standardized and informed approach to LCA. This exploration will notably contribute to filling knowledge gaps, particularly in areas such as the use of biomaterials in greenhouse structures, the integration of 4.0 technologies, and the adoption of circular bioeconomy practices [1,2]. Therefore, identifying the predominant trends and addressing their limitations is imperative for advancing our understanding and enhancing the implementation of LCA in protected agriculture.

In this paper, we undertake a comprehensive exploration of the current landscape of LCA application in protected agriculture through a combined bibliometric and systematic review. The bibliometric analysis shows scientific productivity trends and identifies prominent lines of research within this domain. In tandem, the systematic review delves into the historical development of agricultural LCA and assesses its operability within the context of protected agriculture. These two analytical approaches are elucidated in the sections following the methodology described.

2. Materials and Methods

2.1. Data Collection and Search Strategy

The search, selection, and consolidation of the literature reported on LCA in protected agriculture were performed using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [18]. This method is based on four steps, and its workflow is presented in Figure 1.

The first step is the identification of publications, which was carried out in two sub-steps: defining the search equation and selecting the database. We followed the guidelines provided by Koutsos et al. [19] to conduct systematic reviews of the literature on agricultural sciences. The search equation includes keywords, boolean operators, and truncation symbols. Three parentheses were used to represent the central topic of this review: the first was associated with the agricultural sector, the second with LCA, and the third with the specific protection structures of protected agriculture. Each parenthesis included the most representative keywords using “OR” as a Boolean operator. This operator enables the use of one or more defined keywords. Between the three parentheses, an “AND” was used as a Boolean operator to represent the intersection of the two topics. As the keyword “greenhouse” can be used in the context of both protected agriculture and climate change, it was necessary to use a fourth parenthesis to avoid retrieving off-topic posts (e.g., “greenhouse effect”) using the Boolean operator “AND NOT”. The truncation symbol (*) was also used to find both singular and plural words, and the proximity operator (“ ”) was used to find an exact sentence [20]. The defined search query is shown in

Table 1. Search query used in the databases to consolidate the literature—own elaboration.

Search Query

References

((agri * OR agron * OR “food system *”) AND (“life cycle assessment *” OR “life-cycle assessment *” OR “life cycle analys *” OR “life-cycle analys *” OR “life cycle sustainability assessment *” OR “life-cycle sustainability assessment *” OR “life cycle sustainability analys *” OR “life-cycle sustainability analys *” OR “lca” OR “life cycle thinking” OR “life-cycle thinking” OR “life cycle costing” OR “life-cycle costing” OR “life cycle impact assessment *” OR “life-cycle impact assessment *” OR “life cycle inventory” OR “life-cycle inventory” OR “life cycle impact analys *” OR “life-cycle impact analys *”) AND (“greenhouse” OR glasshouse OR nethouse OR mesh-house OR screenhouse OR “protected agriculture”) AND NOT (“greenhouse effect” OR “gas emissions” OR “greenhouse gas *” OR “greenhouse gas emissions”))

[21,22,23]

*: truncation symbol used to find both singular and plural words; “ ”: proximity operator used to find an exact sentence.

.

We chose to assess the databases Scopus and the ISI Web of Science (WoS) because they are frequently employed in bibliometric research [19] and are managed by Elsevier and Thomson Reuters. The search equation was applied in a single day (21 June 2022) using “topic” as the search scope, which included the title, abstract, and keywords. The set of retrieved publications included all source types (e.g., articles, reviews, conference papers, books, etc.). The number of publications retrieved in Scopus was 105, whereas it was 101 in WoS. After removing duplicate publications between these two databases, a set of 112 publications was compiled.

The second step denotes the initial filtering or screening of the set of retrieved publications. The screening consisted of two sub-steps. The first was to exclude publications that were not entirely in English. Although the keywords were defined in English, some publications are not in English but follow the guidelines that the abstract should also appear in English. The number of publications excluded was 4, of which 3 were in Chinese and 1 was in Spanish. The second substep was to check downloadable and unrelated publications based on an abstract review.

The third step represents the eligibility check, which is the second and final filtering of the set of publications. Eligibility was based on whether the full text of the publications was closely related to LCA in protected agriculture. Finally, the fourth step is the consolidation of the set of publications included as the literature to be analyzed and scrutinized (Table S1) [24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120]. Data from the final set of publications with all available information were exported as comma-separated value (CSV) files and then imported into Microsoft Excel 2016.

2.2. Bibliometric Analysis

The bibliometric parameters used to show the scientific productivity of the research between LCA and protected agriculture were as follows: (i) number of publications per year, (ii) total number of publications both by continent and by country, (iii) number of publications per institution, (iv) number of publications per author, and (v) trend of topics over time. The open-source, R-based tool Bibliometrix v3.0.1 for thorough scientific mapping analysis provided these data. Prism 8 software (GraphPad Software Inc., San Diego, CA, USA) was used to plot the results.

2.3. Analysis of the Thematic Structure of the Publications

The thematic structure of the publications was assessed through a similarity analysis based on the formation of clusters and a text-mining analysis applied to the abstract section. Initially, data-cleaning tools were applied following the methodology of Abd-Alrazaq et al. [121]. Punctuation, numbers, and incomplete words were then removed from the abstracts. All alphabetic characters have been converted to lowercase. The text was separated into sentences using the corresponding Python procedure. Titles that denote sections of the document (i.e., background, objective, results, conclusion) and those within the abstracts were removed. Subsequently, the Python Natural Language Toolkit library was used to tokenize the sentences of each summary, and the SnowballStemmer procedure was used to convert words to their stems and each abstract to a feature vector. Features were defined by term-frequency and inverse-document-frequency (TF-IDF) weights. TF-IDF represents the importance of a word in relation to a document (an abstract) in a corpus (the collection of abstracts). The TF.IDF for term i in document score j is defined by Equation (1):

TF.IDF_ij = TF_ij × IDF_i

(1)

It is the product of the term frequency (TF) and the inverse density frequency (IDF). TF is defined by Equation (2):

TF_ij = f_ij/(max_k f_kj)

(2)

The frequency of term i in document j is normalized by f_ij by dividing it by the maximum number of occurrences of any term in the same document. The IDF is defined as the following:

IDF_i = log₂ (N/n_i)

(3)

The term i appears in n_i of the N documents in the collection (corpus). The terms with the highest TF.IDF scores are usually the ones that best characterize the topic of the document. This score increases proportionally to the number of times the word appears in the document but is offset by the frequency of that word in the corpus.

For TF-IDF representation of the abstracts, the TfidfVectorizer module of the Python scikit library was used. The number of terms to be used for each document was controlled by the parameters of the TfidfVectorizer algorithm. The parameter min_df = 10 determined that the term should appear in at least ten documents, so terms too rare to be used in finding document clusters were excluded. The parameter max_df = 0.90 determined that the term should appear at most in 90% of the corpus; therefore, some terms that are too common were excluded.

Hierarchical clustering was used for the unsupervised grouping of selected papers based on the similarity of their feature vectors. The resulting clustering tree was cut to select four branches following Ward’s procedure to guarantee internally coherent but well-separated groups. To account for each cluster resulting from the tree, random forest analysis with the scikit-learn Python library for supervised classification of selected branches of hierarchical cluster analysis was applied. The RFA was trained using the bagging method, which combines bootstrapping and aggregation. Using these features, the mean of the TF.IDF score was calculated to create a star graph and identify the key terms linked to the main subjects in the classes found from the cluster analysis.

The number of features required to construct each tree corresponds to the square root of the total number of features resulting from the analysis using the TfidfVectorizer module. With the grid search result from the random forest analysis, the list of feature importances to select the first relevant features in a number equal to the tree features was found. Finally, the mean of the TF.IDF score was calculated to make a star chart and identify the key terms linked to the main themes in the groups found.

2.4. Systematic Review

In conducting our systematic review, we adopted a two-fold approach. First, we outlined the historical progression of LCA chronologically, from its inception in a broader context to its specific applications in agriculture, with a special focus on advancements in protected agriculture. This involved incorporating crucial events and noteworthy reports that have significantly influenced the field.

Moving on to the second phase, we examine the operational dynamics of LCA within protected agriculture. Our study followed the steps set out by the International Organization for Standardization (ISO) in standards 14040 and 14044 [122]. The four steps are the definition of objectives and scope, inventory analysis, impact evaluation, and interpretation. Each phase was subjected to scrutiny regarding the methodological aspects of decision making. We quantify the frequency of its use and provide detailed information on the most relevant findings. To visually represent the prevalence of these aspects, we use bar and pie charts.

Table 2. Review framework to explore multiple aspects of the 4 stages of the LCA applied to protected agriculture—own elaboration.

LCA Stage	Analyzed Aspects
Objective and scope	(i) Objective of the study, (ii) scope definition, (iii) productive system, (iv) system boundaries, and (v) functional unit.
Inventory analysis	(i) Water consumption, (ii) use of fertilizers and pesticides, (iii) greenhouse structure, (iv) energy consumption in air-conditioning work, and (v) harvesting, post-harvesting, and commercialization.
Life cycle impact assessment	(i) Eutrophication potential, (ii) global warming potential, (iii) acidification potential, (iv) human toxicity potential, (v) ozone depletion potential, and (vi) abiotic depletion potential.
Interpretation	Decision frameworks and uncertainty analysis

shows the framework used in this systematic review, in which we included the most predominant aspects of the literature in phases 2 and 3.

The description and discussion of the findings allowed us to identify environmentally friendly practices in protected agriculture, analyze the coherence of methodological decisions with respect to ISO 14044 guidelines, and deduce whether methodological choices limited the value of the study in promoting environmentally sustainable protected agriculture.

3. A Brief History of the Agricultural LCA

LCA has undergone an evolution in both methodology and applications for more than 60 years to become a well-known and widely used tool in industry, academia, and politics [123]. The LCA has presented different stages of development and adoption, including (i) the early years period (1960–1970), (ii) the conception period (1970–1990), (iii) the standardization period (1990–2000), (iv) the elaborations period (2000–2010), and (v) the concept broadening period (2010–until now) [124]. The conceptualization of LCA emerged in the 1960s as Resource and Environmental Profile Analysis (REPA) in the US or Ecobalances in Europe [125]. Initially rooted in industrial packaging research, LCA primarily focused on energy use and pollutant emissions. Knowledge dissemination of early LCA advances was limited, leading to largely uncoordinated method development in the US and Northern Europe [12]. Notable reports and committees reflecting the knowledge generated in the 1970s include “The Limits to Growth”, “A Blueprint for Survival”, and the “Committee on the Survey of Materials Science and Engineering” [125].

Methodological innovation and international coordination of LCA between private sector stakeholders and the scientific community have substantially increased since the 1980s. The term LCA, particularly the concept of life cycle thinking, became established and normative in the 1990s [124]. Key actions in the late 20th century included the launch of normalized assessment scores for environmental impact categories (CML-92 method), the publication of the LCA sourcebook A European Guide to Life Cycle Assessment, the establishment of the International Journal of Life Cycle Assessment, the introduction of the standards 14040 and 14041 of the International Organization for Standardization (ISO), and the development of the Dutch Eco-Indicator 99 method [126]. In the following decade, consolidation and harmonization of the LCA methodological basis occurred, primarily through the Life Cycle Initiative, the Ecoinvent database, and the ISO 14040 standard [127]. These events expanded the coverage to include more products and systems, with results disseminated through academic papers and government and industry reports [128].

The first set of LCA guidelines for agricultural production systems was developed as part of the “Harmonization of Environmental Life Cycle Assessment for Agriculture” project [129]. This project addressed the greater complexity inherent in the application of LCA to agricultural products due to multiple converging processes and factors. Subsequent efforts on LCA in agriculture led to biannual LCA Food international conferences, recognized as the “reference event for the LCA Food scientific community” [130]. LCA studies of agricultural and food systems continue to intensify, becoming a field of research to respond to growing public concern about climate change, biodiversity loss, and soil and water pollution in productive agricultural areas [128].

In 2012, the World Food LCA Database Project established stronger and more comprehensive guidelines for evaluating new environmental aspects [131]. Initiatives such as the Envi Food Protocol (Food SCP RT 2013) and the Product Environmental Footprint (PEF) Pilot were launched to address the new stages of the life cycle of a food product [12]. Mitigating environmental damage to a greater extent became a central goal, capturing the interest of researchers who conducted studies on different agricultural production systems, including protected agriculture [132]. Valuable information on trade-offs between alternative agricultural processing, distribution, and consumption behaviors associated with food systems has been discussed and reported [133]. Similarly, data for the verification of environmental technologies, including food process wastewater treatment [134], formulation of forecast policies for public payments for organic production [135], sustainable comparison of specific goods of food choices [136], and improvements in eco-design [137,138], are available. In protected agriculture, LCA has promoted the study of innovative systems for sustainable greenhouse production, including the exchange of energy between the greenhouse and its associated building, as well as light and shade transmission processes [72]. LCA methodological development continues to this day, employing a variety of approaches to study the nutritional, environmental, economic, and social impacts of the agricultural sector under its different production systems.

4. Bibliometric Overview of LCA in Protected Agriculture

Research on LCA applied to protected agriculture began more than 20 years ago, with its inception in 2000, marked by two publications (Figure 2). These initial publications explored the advances and potential of bioenergy from biomass crops in the United States and delved into the role of consumers in food purchases [120,139]. In the first decade, scientific production exhibited variability, averaging less than five publications annually. Notably, no publications were reported between 2001 and 2004. The inaugural study intrinsically linked to protected agriculture surfaced in 2005, introducing an LCA methodology encompassing various greenhouse types and emphasizing the benefits of hydroponic crops and recycled plastics [119]. No papers were published in 2006 or 2007. The publication count in the first decade peaked in 2008 with four publications, followed by three in 2009 and one in 2010.

In the second decade of research on LCA in protected agriculture, the number of publications more than doubled compared with the first decade. However, the produc-tivity trend varied from year to year. We observed continuous growth in the number of publications over a three-year period, from 2010 to 2012, from one to five publications, and from 2016 to 2018, from seven to ten publications. The year with the largest scientific production was 2020, with fifteen publications, followed by 2019 and 2018, both with ten publications, and 2017, with nine publications. A minor decrease in publications was noted in 2021, with ten reported. Up to the current database search in 2022, four publi-cations have been found.

Europe emerges as the leader in scientific productivity of LCA in protected agri-culture by continents, with Asia, America, Oceania, and Africa following suit (Figure 2). The top five countries are Spain, Italy, China, the United States, and Germany. The top five countries driving this productivity are Spain, Italy, China, the United States, and Germany. These observations align with the notion that Europe is a pioneer in protected or intensive greenhouse agriculture, boasting over 43% of the world’s greenhouse area [140,141]. Spain, France, Greece, Italy, and the Netherlands stand out with the largest areas of protected cultivation, with Spain leading at 71,783 hectares, notably concen-trated in agricultural regions such as Almeria province [142,143].

Examining the most prolific affiliations and authors per continent based on the number of publications reveals a notable focus on protected agriculture, urban agriculture, and sustainable agricultural production (Figure 2). European research centers its themes on urban agriculture, industrial ecology, rooftop greenhouses, and the circular economy applied to specific crops. Notably, Spain, Italy, and France are often featured in case studies exploring models for ecological compensation, environmental and economic cost–benefit analyses, and the evaluation of water and nutrient recirculation [59]. In Asia, research aligns with sustainable agriculture and environmental impact analysis for crops such as strawberries, wheat, tomatoes, sunflowers, and cucumbers. This research investigates greenhouse gas emissions, carbon footprints, nitrogen footprints, pesticides, and solar energy sources [64,83].

The American continent, influenced by European research, has shown interest in rooftop greenhouse studies aimed at establishing urban gardening in commercial parks. Additionally, there is research on hydroponic and biogas systems to close cycles in greenhouse processes [38,70,97]. Oceania’s focus includes the adoption of algal biofertilizers for nutrient circularity, the evaluation of greenhouse typologies in various cropping environments, and the reuse of municipal wastewater [28,37]. Finally, African researchers have explored the limits and prospects for smallholders in protected vegetable cultivation, the use of stem residues for banana production, and innovations in cucumber irrigation using groundwater [43,75].

5. Main Research Lines concerning LCA in Protected Agriculture

The implementation of the random forest algorithm enabled the classification of group names into categories such as “water management”, “technology management”, “energy management”, and “impact measurement” (Figure 3). These categorizations are derived from the prevailing themes within the clusters, potentially reflecting the research directions outlined below, deduced from the content of the publications, and the relevance of keywords (Figure S1).

Efficient water management: This cluster focuses on optimizing water usage within greenhouse environments. Key considerations include addressing water scarcity [108], enhancing irrigation efficiency [3], ensuring water quality [43], managing leachate drainage [44], and mitigating the impacts of climate change [73]. Emphasis is placed on proper leachate management to facilitate reuse and prevent environmental concerns such as groundwater contamination. The integration of decision support systems and innovative irrigation technologies, such as smart irrigation, is a crucial avenue to significantly enhance efficiency and minimize costs.

Technology management: This cluster delves into technologies aimed at reducing environmental impact and enhancing efficiency, particularly in regions with challenging climates. It encompasses strategies for achieving high-quality vegetable production in low-energy systems [13,16] and the use of alternative energy sources such as biomethane [47], biodiesel [57], wind power [38], and solar panels [64].

Efficient energy use: This cluster focuses on optimizing energy consumption for heating and transportation. Key aspects include initiatives to reduce fuel consumption [18] and design optimizations to minimize heat losses in structures [116]. Research within this cluster has investigated the environmental impact of tomato production and transportation on consumers [25] and explored diverse marketing methods for greenhouse vegetables [31]. Strategies to reduce transportation distances are considered crucial, with investigations into approaches such as urban agriculture, rooftop agriculture, and zero-mile agriculture [25,31,41].

Impact measurement: This group delves into evaluating the environmental footprint of various agricultural systems and practices. Research within this group includes analyzing the impact of cultivating various crops such as tomato [49], pepper [65], wheat [66], cucumber [67], apricot [80], lettuce, and barley [90]. Some studies have explored the combined impact of multiple species grown together, considering aspects such as recycling drainage effluents [29] or using urban and rooftop greenhouses [48].

6. Technical Discussion of the Operability of the LCA

6.1. Goal and Scope Definition

LCA emerges as a powerful tool not only for comprehensive comparisons between competing and potential systems but also for the optimization of existing systems. The successful achievement of results in an LCA is intrinsically linked to the precise formulation of the objectives and scope; therefore, this phase is an integral part of the LCA execution process and is interrelated with each of the subsequent phases [17].

6.1.1. Goal

The objectives of the LCA analysis should be clearly defined, encompassing the target audience, the context of the work performed, and the causal factors driving the LCA, along with the expected utility of the results [144,145]. These aspects facilitate the understanding of the findings, enabling different stakeholders to focus on specific and relevant aspects of the analyzed product or process [146]. The current overview of the objectives in the reviewed literature is illustrated in Figure 4, which is consolidated into four main groups.

The first group focuses on comparative environmental studies that analyzed various greenhouse cultivation alternatives. For instance, it assesses climate optimization alternatives or diverse types of greenhouses, studies the efficiency of water and fertilizer use in soilless cropping systems, explores the integration of various renewable energy sources, and examines pest and disease management strategies. These studies propose viable alternatives to mitigate the environmental impacts associated with conventional intensive agricultural production [147].

The second group comprises publications that conducted descriptive environmental studies, aiming to propose production strategies with a circularity approach under greenhouses. This approach enhances the utilization of waste generated in agricultural production. Likewise, researchers have identified studies that aim to increase crop yields and improve consumption habits to reduce the negative environmental impact generated during the commercialization phase of the food supply chain [148].

The third group encompasses studies that aimed to compare production in open fields and greenhouses, analyzing various irrigation and fertilization management strategies, along with the impact of diverse climatic conditions. As for the fourth group, we identified reports that analyzed the food supply chain from both an environmental perspective and the socio-environmental standpoint of the water–energy–food nexus. This context has recently gained significant attention in international science and policy to promote integrated resource management and sustainability [84,149].

6.1.2. Scope

The scope of LCA can be categorized into descriptive analyses, comparative analyses, or combinations of both (Figure 5). Descriptive analysis aims to identify environmental burdens in the production process of a product or service, whereas comparative analyses seek differences in environmental loads within modified or optimized processes.

In the scope of comparative analyses, constituting 67.8%, studies such as He et al. [83] compare conventional and organic tomato production under greenhouse conditions. While reporting that organic production yields up to 54% less environmental impact than conventional methods, this study stresses the need to optimize the use and conditions of organic fertilizers to limit soil ecotoxicity. Similarly, Martínez-Blanco et al. [110] compared tomato production in open fields and greenhouses by examining fertilization practices using mineral fertilizers and compost. They found that greenhouse production can increase by up to 50% compared with open fields, albeit with higher resource and energy requirements during construction. Additionally, using compost in the greenhouse can replace part of the mineral nutrition and reduce environmental impacts on open-field cultivation.

The descriptive scope, representing 27.8%, includes works describing environmental loads in the production of certain foods in greenhouses [72]. Research also focuses on evaluating environmentally friendly technologies to identify innovation priorities for agricultural production [107].

Regarding studies of combined scope, accounting for 4.4%, their objective is to showcase reduced environmental impact through process optimization or innovation. Hosseini-Fashami et al. [55] assessed the energy use and life cycle of strawberries under greenhouse conditions and found that integrating solar technologies reduces negative environmental impacts by at least 16% compared with diesel fuel.

6.1.3. Types of Food Products Analyzed

The products under analysis in each study were categorized into eight groups, as illustrated in Figure 6. Among these, tomatoes emerged as the most extensively studied product, featured in 42 of the analyzed papers. Other vegetables, including chard, spinach, and eggplant, were the subject of 10 studies, while cucumbers, lettuce, and peppers were each examined in 8, 7, and 4 studies, respectively. Consequently, the total number of studies dedicated to the analysis of horticultural products reached 71. This trend likely stems from the realization that both low- and high-tech greenhouse horticulture engenders various environmental impacts, posing a significant threat to sustainability [150]. Consequently, there is a growing emphasis within the scientific community, among producers and consumers, on the evaluation and transparent communication of these impacts. This extends beyond understanding them to actively seeking strategies for optimizing production systems to mitigate these environmental burdens [90].

We also identified 21 studies addressing other types of products, such as bananas, citrus fruits, tilapia grown in aquaponic systems, beans in covered greenhouses, and beans in open-field greenhouses. Furthermore, four studies focused on strawberries, and an additional four explored cut and ornamental flowers as well as potted plants. Notably, considering the global significance of greenhouse flower production, the number of publications in this domain remains comparatively low. Despite ongoing investigations into strategies for enhancing flower quality and yield, there appears to be limited interest in analyzing their environmental impacts [116,151,152].

6.1.4. System Boundaries

System boundaries present a formidable challenge for LCA because they require meticulous establishment of the limits for the analyzed system, service, or product. This requires a rigorous definition of each activity, input, and unitary process integral to the analysis [35]. Figure 7 illustrates the breakdown of approaches found in the collected papers, with 37% of the papers delineating the system from cradle to gate. In this comprehensive approach, LCA encompasses everything from procuring and extracting basic raw materials to production, harvesting, and, at times, packaging or packing of the final product.

The gate-to-gate approach was followed closely and adopted by 17% of the studies. This approach focuses solely on activities within the production nucleus, disregarding the entry of raw materials to the farm, the end-use of the final product, and the waste generated during production. These two approaches, both cradle-to-gate and gate-to-gate, have been widely favored for analyzing environmental impacts on various industries, including agriculture [153]. They prove particularly effective for evaluating the environmental performance of technological changes implemented in production processes, a common occurrence in covered agriculture where multiple technological options exist for food production [49].

A subset of papers, constituting 15%, narrowed the LCA to “cradle to consumption”. This approach typically excludes consideration of the product’s final use, limiting the analysis to the point of delivery to a wholesale or retail market or when the consumer makes the purchase, irrespective of the product’s eventual use or disposal. Lastly, 8% of the papers embraced a cradle-to-grave approach, comprehensively examining impacts at each stage of a product’s life cycle—from raw material extraction and processing to production, harvesting, transformation, transportation, product use, and, ultimately, disposal [154].

6.1.5. Functional Unit

Establishing a functional unit is pivotal for constructing and modeling a product system in LCA. This unit quantifies the product’s function and serves as a crucial reference point for all impact assessment calculations in both descriptive and comparative studies [83,144]. The functional unit can be derived from various product characteristics, primarily those related to the production system [155]. In agriculture, the unit of mass of the final product, such as 1 mg, 1 g, 1 kg, or 1 ton of the product, is commonly chosen, as indicated in Figure 8. Over half of the studies opt for functional units of this typology, referred to as product quantity, with kg and tons of the product being predominant.

Nevertheless, in some studies, the definition of a functional unit varies based on the objective and system limits. For instance, four studies designate the unit of the product as the functional unit. Bonaguro et al. [79] considered a single plant in its 14 cm pot as a functional unit, revealing that substituting 10% expanded perlite for 10% fresh rice husks can reduce the environmental impact of cyclamen pot production without compromising the quality of the final product. Another study by Canaj et al. [49] used one hectare of tomatoes grown under an unheated greenhouse in Albania as the functional unit. The authors concluded that these results could serve as a “springboard” for detailing and quantifying the environmental impacts of other crops and as a foundation for extending sustainability thinking in agricultural production.

Approximately 10% of the studies used the amount of product obtained per unit of harvested area or simply the amount of product obtained, indicating yield and productivity. These studies focused on comparing the environmental impact of different cultivation systems in rooftop greenhouses and plastic waste generated in greenhouses across various parts of the European Union. About 3.3% of studies used water and energy resources as functional units, basing their analysis on resource consumption. These studies compared the environmental impacts of greenhouse air-conditioning and the water footprint resulting from different irrigation methodologies and growing systems. Additionally, 2.2% of studies adopted the cost of the product as a functional unit. This scenario arises, for example, when comparing a conventional greenhouse tomato production system with a greenhouse cultivation system equipped with photovoltaic panels. In such cases, the authors mention that the functional unit is generated to have a monetary basis for integrating crop production and electricity, enabling the analysis of a system that produces goods with different commercial values [64]. Finally, we noted that almost a third of the studies did not specify the functional units of the LCA.

6.2. Life Cycle Inventory

Inventory analysis involves collecting information from sources, creating flowcharts, and applying calculation methods [83,144]. These steps enable the accurate estimation of relevant inputs and outputs associated with a product and its system [104]. Sources can be primary (measured in situ or through field surveys) or secondary (database) [83,144]. In LCA applied to protected agriculture, several databases were used, such as GaBi from Germany [156], CLCD from China [157], NREL-USLCI from the USA [158], Ecoinvent from Switzerland [159], LCA Food from Denmark [160], and ELCD from Europe [161]. Most studies clarify whether the calculation of potential environmental impacts can refer to the global, regional, or site scale. Furthermore, some studies fully describe the assumptions made regarding agricultural production.

We identified five categories within inventory analysis, as shown in Figure 9. Among these categories, the most frequently reported was the use of fertilizer and pesticides, followed by water use and greenhouse infrastructure. The categories of energy consumption for air-conditioning and harvesting, post-harvest, and commercialization are also reported.

6.2.1. Water Consumption

Water depletion, a crucial environmental impact category in LCA, aims to curtail crop water consumption [66]. A transformative strategy to achieve this is the digitization of agriculture, which provides real-time insights into water inputs and outputs throughout the crop growth cycle. This involves harnessing data from soil moisture sensors, infrared plant images, and water meters [162].

Ensuring a genuine and verifiable water inventory over the crop cycle is imperative for precisely assessing the water footprint and devising strategies for enhanced water management [36]. Detecting inefficiencies in irrigation water usage is critical for implementing intelligent irrigation practices, which, despite initial investment costs, yield substantial benefits in terms of water conservation and reduced energy for water pumping [27,38].

Diverse techniques can be employed to reduce crop water demand. Controlled-environment agriculture, such as aquaponics and hydroponics, has emerged as a promising solution because of its high productivity and minimal use of soil and water [38]. Using rainwater for irrigation reduces reliance on the water supply network, while reclaimed water offers an alternative to address water scarcity [43]. In addition, desalinated seawater can help alleviate aquifer overexploitation, albeit with a caveat—the environmental impact hinges on the sustainability of the desalination or wastewater treatment process, particularly in terms of clean energy use [29].

While agronomists traditionally measure water use efficiency in absolute volumes, a recent method rooted in LCA provides a comprehensive perspective. This method gauges water footprints in terms of freshwater scarcity, considering the nature of the water source, whether it is treated wastewater, groundwater, or desalinated water for irrigation [99].

6.2.2. Use of Fertilizers and Pesticides

Analyzing the entry of fertilizers and pesticides into LCA involves systematically collecting data on material flows, energy, and emissions throughout the production, transportation, use, and disposal of these substances in protected agricultural systems [163]. Efficiency in using fertilizers and pesticides within protected agriculture is pivotal for mitigating adverse impacts, including increased nutrient leaching, greenhouse gas emissions, and excessive water and energy consumption [37,50,164]. Achieving this involves determining optimal dosages and carefully selecting products and application methods.

Recent research has presented promising alternatives to reduce the dosage of fertilizers and pesticides. Firstly, the application of organic amendments, such as plant waste and composted animal manure, emerges as a standout option. These amendments contribute to soil carbon sequestration and enhance crop productivity by providing nutrients, thereby maintaining fertility and structural stability [68,80]. However, proper treatment is crucial to prevent the transmission of pathogens, the diffusion of hormonal substances, or the spread of antibiotic-resistant genes [83]. In this context, the ‘One Health’ approach gains relevance, promoting balanced well-being among humans, animals, and the environment in managing organic amendments [122].

Secondly, the use of microbial inoculants with metabolic capacities that promote plant growth is noteworthy not only to replace mineral fertilizer doses but also to mitigate the effects of abiotic and biotic stresses on crops [165,166,167]. Based on the meta-analyses reported so far, Pseudomonas is the most effective bacterial genus for improving crop yields by alleviating stress, while Enterobacter is the most effective for improving plant nutrient availability [165,166,167,168]. In terms of crops, these two genera have the greatest beneficial effect on horticultural crops. Also highlighted are endophytic microorganisms that live in the plant tissue of various fruits, vegetables, and medicinal plants. These can generate phenotypic benefits from seed germination, fruit development, and even the post-harvest period. The most prominent endophytic bacterial genera are Burkholderia and Azospirillum [165,166,167,169,170]. Additionally, microbial inoculants serve as useful biological seed treatments, counteracting chemical seed treatments and improving seed germination, seedling emergence, plant biomass, and crop yield [122,171,172]. All these virtues of the use of microbial inoculants translate into LCA in the reduction of the environmental impact of agricultural production in terms of mineral extraction, eutrophication and aquatic ecotoxicity, acidification, and global warming. [4,173].

Thirdly, the use of bioactive plant substances, such as tansy flower extract, has been highlighted as an effective alternative for pest control in horticultural systems [174]. Furthermore, microbial inoculants of the genus Bacillus have proven to be bioalternatives with numerous benefits in pest control [175,176]. This approach helps avoid negative impacts on biodiversity and ecosystem health, contributing to more sustainable and ecologically responsible greenhouse agriculture without compromising productivity.

6.2.3. Greenhouse Structure

Within the LCA framework, this category typically encompasses all aspects related to greenhouse infrastructure, including foundations, structures, enclosures, climate control equipment, and even the energy and man-hours consumed during construction [90]. The diversity of greenhouse typologies used globally, coupled with the varying scales and technologies employed in each country, makes this category highly heterogeneous—ranging from small to large scale and low to high technology [4]. Incorporating the greenhouse structure into the LCA analysis is crucial as it provides an opportunity to pinpoint strategic areas for improvements to reduce environmental impact. Potential enhancements include the use of more sustainable materials and the integration of environmentally friendly technologies [177]. It is worth noting that some authors opt to exclude infrastructure from the analysis because of challenges in obtaining accurate inventory data from builders or manufacturers. They argue that, given the long useful lives of many structural components, their contribution to environmental impact may not be as significant compared to other inputs [178].

The greenhouse structure often poses the most significant environmental impact concerning global warming and abiotic depletion. Primarily, this is attributed to the extraction of minerals and fossil fuels for raw materials like steel, concrete, and plastic, leading to substantial CO₂ emissions [179]. Studies have quantified that these structural materials contribute up to 35% and 47% to these impact categories in low-tech greenhouses [100,101]. In high-tech greenhouses, the contribution to global warming is around 10%, driven by greenhouse gas emissions and the high energy consumption during the manufacturing of materials such as aluminum and glass—key components of venlo-type greenhouses [179]. The distinction between greenhouse types is significantly influenced by the equipment and energy used for heating, lighting, and cooling, which are common practices in high-tech greenhouses [180].

6.2.4. Energy Consumption in Air-Conditioning

In LCA, the energy consumption associated with air-conditioning emerges as a critical factor, particularly in production systems featuring high-tech greenhouses and regions with extreme climatic conditions preventing traditional agricultural practices [32]. In such scenarios, these advanced greenhouses exhibit a substantial environmental impact compared with their low-tech counterparts and open-field crops. Despite the benefits of climate control and enhanced efficiency in agricultural production, high-tech greenhouses significantly contribute to the environment, primarily because of the consumption of fossil fuels and electrical energy for heating and cooling operations [181,182]. The continual need for energy to sustain optimal climatic conditions results in a noteworthy carbon footprint, elevating greenhouse gas emissions and contributing to the broader challenge of climate change. Consequently, ensuring the sustainability of high-tech greenhouses becomes imperative, necessitating the adoption of eco-friendly practices and the advancement of renewable energy sources to alleviate their environmental impact [183].

Venlo greenhouses, a prevalent type of high-tech greenhouse, exhibit climate control activities as major contributors to impact indicators related to global warming, air acidification, abiotic depletion, eutrophication, and photochemical oxidation, constituting 81–97% of these indicators [144]. However, the extent of these impacts is contingent on the type of energy used for climate control and improvements in the energy efficiency of associated systems. Torrellas et al. [183] demonstrated that in a Venlo-type greenhouse with an electrical cogeneration system, electricity production exceeded the greenhouse’s needs, resulting in negative values for the eutrophication impact indicator.

Current efforts to address the negative environmental impacts associated with high-tech greenhouses concentrate on innovations aimed at optimizing the thermal insulation of materials used in their roofs and enclosures [72]. Furthermore, strides have been made in adopting sustainable energy sources, such as solar and geothermal energy, to fulfill greenhouse air-conditioning needs [64]. A noteworthy recent development involves the construction of greenhouses on urban building roofs, capitalizing on residual energy from domestic activities and building thermal comfort, leading to enhanced energy efficiency [70].

6.2.5. Harvesting, Post-Harvesting, and Commercialization

Harvesting, post-harvest, and commercialization processes encompass several activities, with transportation, packaging, and drying playing pivotal roles in LCA considerations. The impact of transporting goods from farms to retailers or consumers is substantial, influencing fuel use and metrics such as carbon footprint, fuel consumption, global warming potential, carbon dioxide equivalent (CO₂ eq), and fossil fuel depletion. However, quantifying transportation poses a challenge in LCA due to its complexity, marked by diverse modes and means influenced by spatial and temporal variations [184]. While some studies model transportation, a significant portion of LCA analyses in protected agriculture opt to overlook this stage, citing data scarcity and associated complexities [171]. In the realm of protected agriculture, only 12.4% of LCA studies address the impact of transporting and distributing food products.

The energy efficiency of transportation, contingent on vehicle types, is a crucial aspect. Urbano et al. [25] conducted a life cycle analysis on eight scenarios of fresh tomato supply to urban consumers, revealing that long transportation distances from production sites to supermarkets exert the highest environmental impact on global warming. Lorries, especially those meeting EURO 6 standards with a 32-ton capacity, are typically used for long-distance tomato transport. Conversely, lorries with a capacity of 3.5–7 tons are employed for shorter distances. Pineda et al. [41] found that rooftop greenhouses in urban areas outperform conventional rural greenhouses because they produce tomatoes with a lower environmental impact due to reduced transportation distances and smaller heat losses during colder months.

The environmental impact of air transport warrants careful evaluation, especially considering its potential harm to certain products. In Switzerland, where 51%–60% of fruits and vegetables are imported, Stoessel et al. [106] emphasize that reducing food transportation distances is the most effective means of reducing environmental impact. Consumption of seasonal produce from local sources is advocated, with air transport particularly discouraged due to its disproportionately high environmental impact.

Packaging and drying activities contribute to the environmental burden, necessitating the exploration of (bio)alternatives, especially for materials with high carbon emissions and dependencies on fossil energy. While packaging consumption is addressed in a mere 2.1% of agricultural production system LCA analyses, exploring recycled boxes and promoting direct consumption can circumvent the need for cardboard, wood, and polypropylene packaging. Post-harvest drying, mentioned in only 1% of studies, amplifies energy demand but concurrently reduces the impact of climate change, as highlighted by Bosona et al. [68], emphasizing the benefits of drying organic tomatoes in terms of reducing product losses and enhancing shelf life.

6.3. Life Cycle Impact Assessment

Life cycle impact assessment is a pivotal stage in LCA, translating collected data into a nuanced array of potential impacts strategically weighted through category indicators. This approach offers practitioners and decision makers a deep insight into the negative consequences of resource use and associated emissions [83,144]. In practice, this phase involves classification, characterization, normalization, and weighting. While the first two are mandatory, normalization and weighting are considered optional components of this third phase [83,144].

The classification step entails identifying the impact assessment method, often relying on established methods such as CML (including versions like CML 2 baseline 2000 V2/world and CML 2000), ISO 14044 (2006), ISO 14040, and IPCC 2001 GWP 100 [144,185]. In the characterization phase, impact categories are selected, and the environmental impact on each category is quantified. Figure 10 illustrates the most frequently reported impact categories in the studies. Eutrophication potential and global warming potential were the most utilized, with 55 and 54 papers, respectively. Acidification potential and human toxicity potential had 49 and 34 papers, respectively, while ozone and abiotic depletion potential were featured in 31 and 28 papers, respectively. The subsequent subsections provide detailed descriptions of these six impact categories.

The increasing need to identify the most pertinent impact categories and achieve unambiguous results has spurred additional research into normalization and weighting [83,144,186]. Normalization offers insights into the relative significance of impacts, enhancing the interpretation of the results. Various approaches to normalization exist and are categorized as follows: global normalization, production-based normalization, and consumption-based normalization [83,187]. Despite notable strides in normalization approach development, several questions remain. Crucial inquiries involve methodological disparities between the studied and reference systems, alignment of the spatial extension of the reference system with the system under study, and alignment of the reference year with the study year [188].

The weighting process facilitates the ranking of alternatives and aids stakeholders in making informed decisions. Over the last few decades, various weighting sets have been developed. Different techniques are available for normalization weighting: (i) binary weighting (for zero weights or equal weights); (ii) panel weighting, involving the consideration of opinions of stakeholders and experts; (iii) monetary weighting; and (iv) distance to target, where weighting factors depend on the distance between the current impact level and a desired state based on regulations [144,189]. The weighting process is challenging because the composition of the panel or the design of the questionnaire may influence the weighting factors. Therefore, it is advantageous, if possible, to apply different weighting methods to verify the robustness of the study’s conclusions [83,186].

6.3.1. Eutrophication Potential

Eutrophication potential involves the augmentation of nutrients in soil or water, with nitrogen and phosphorus being the primary contributors to this enrichment. The surplus nutrients lead to the overgrowth of algae and plants. In aquatic environments, this excessive proliferation of microorganisms causes oxygen depletion and reduced penetration of solar energy. Consequently, it contaminates plants and groundwater, contributing to terrestrial eutrophication [190].

This review uncovered notable results, particularly in recent years, where studies have compared production in open fields with that in conventional and indoor greenhouses. For instance, Martínez-Blanco et al. [110] found that tomato production in open fields has an eutrophication-related environmental burden up to 57% higher than that in greenhouses. Similarly, Page et al. [99] reported a greater environmental impact of growing tomatoes in open fields and low-tech greenhouse production systems than in medium- and high-tech greenhouses. Payen et al. [96] noted that the tomato crop produced in Morocco and marketed in France significantly contributes to freshwater eutrophication, accounting for 66% of the impact, mainly due to phosphate emissions during fertilizer production, while the manufacturing of containers and transportation contributed 20% and 12%, respectively. Regarding rooftop greenhouses, Muñoz-Lieza et al. [32] concluded that this alternative for horticultural production in highly populated cities requires structural improvements and enhancements in roof materials to halve the current environmental impact associated with potential water eutrophication, both sweet and salty.

Additional studies have concentrated on optimizing water usage, cropping systems, and fertilizer application. Martín-Gorriz et al. [29] suggested that a significant reduction of up to 72% in the eutrophication potential is attainable through the implementation of hydroponic growing systems with leachate treatments. Similarly, Canaj et al. [27] indicated that employing smart irrigation systems in protected Mediterranean agriculture could reduce eutrophication by up to 15% compared with the current situation.

6.3.2. Global Warming Potential

Global warming potential serves as an indicator that measures the overall impact of a process by assessing the absorption of thermal radiation in the atmosphere caused by the emission of greenhouse gases associated with that process [191]. The analysis of global warming potential quantifies the contribution of greenhouse gases to climate change, encompassing various activities such as energy consumption for maintaining optimal microclimatic conditions, energy use linked to resource consumption, and waste and emissions management from specific agricultural practices [192]. This metric plays a key role in conducting a comprehensive environmental assessment of indoor agricultural production [193].

Greenhouse agricultural activities have substantially contributed to the global warming impact category [105]. Whether it is the production of structures, the implementation of automated technologies, the selection of materials, or specific growing practices, each stage in the greenhouse supply chain and operation significantly influences greenhouse gas emissions [101,194].

The assessment of global warming potential in protected agriculture reveals noteworthy findings when comparing various systems with different technological levels. For instance, fully automated, climate-controlled greenhouses used in urban agriculture exhibit lower environmental performance than conventional plastic-covered greenhouses. These structures yield values of 2.59 kg (CO₂ eq) for tomatoes and 26.51 kg (CO₂ eq) for lettuce, in contrast to 0.59 kg (CO₂ eq) and 0.92 kg (CO₂ eq) in conventional greenhouses [82]. Similarly, the roof greenhouse, despite its operational efficiency with automated technologies, shows an environmental load between 21% and 66% higher in the heating potential category than conventional glass and ethylene-vinyl acetate plastic film greenhouses. This increase is primarily attributed to the energy considerations associated with automation and air-conditioning [32].

In the context of supply chains, the transportation of tomato crops from the Mediterranean to export European markets has emerged as a noteworthy contributor to climate change, representing 44% of the impact for tomatoes transported from Morocco to France. Concurrently, the cultivation of tomatoes in climate-controlled greenhouses in France significantly contributes to the climate change impact, with a notable share of 45%. This substantial contribution is primarily associated with CO₂ emissions incurred during the manufacturing processes of greenhouse components and the electricity consumption for fertigation and irrigation [96]. These findings align closely with those reported by Bojacá et al. [101], except for the non-heated greenhouses used in Colombia for tomato production.

In soilless cropping systems, particularly ornamental crops, the acquisition and transportation of peat have been identified as significant contributors to global warming, accounting for 66% of the impact. In contrast, expanded perlite and rice husk make minor contributions at 8.5% and 8%, respectively. Therefore, when selecting substrates, it is important to consider environmental aspects alongside technical suitability [79]. In the context of greenhouse banana production, approximately 55% of the global warming impact is attributed to the incineration of crop residues. Adopting practices such as implementing biogas to heat irrigation water in this greenhouse crop can reduce the environmental footprint by up to 5% compared with scenarios where irrigation water is heated with natural gas [46].

6.3.3. Acidification Potential

Acidification potential is an indicator of the impact of acidifying emissions throughout the agricultural life cycle [195]. This metric gauges the capacity of compounds, such as sulfur and nitrogen oxides, to induce acidity in their surroundings [43]. In the realm of greenhouse LCA, potential acidification is influenced by various components of the production system. Notably, the application of nitrogen-based fertilizers, especially those containing ammonium and nitrate ions, plays a significant role in potential acidification [196]. Additionally, the type of energy used in greenhouse climate control, particularly if it involves the consumption of fossil fuels, can contribute to emissions of sulfur dioxide and nitrogen oxides, thereby contributing to phenomena such as acid rain and the acidification of natural resources [197].

A thorough investigation by Canaj et al. [49], focusing on a greenhouse tomato plantation in Albania, identified ammonia volatilization as the primary contributor to fine particle formation and soil acidification potential. The authors underscored the importance of factoring in geographical and site-specific elements in LCA studies for more precise assessments of the environmental impacts linked to tomato and other greenhouse crop production. One potential solution to alleviate these issues could involve the adoption of intelligent irrigation and fertilization systems, contributing to a 5% reduction in the potential for acidification compared with traditional irrigation systems [27].

Likewise, an analysis of the supply chain has identified that transportation activities to the end market contribute up to 50% of the total impact in the land acidification category. Following closely are fertilization and pest and disease management activities with 39%, and finally, post-harvest or processing activities such as packaging with 10% [96]. The impact during transportation is mainly attributed to nitrogen oxide emissions generated by truck transportation. Moreover, sulfur dioxide emissions, which are related to fertigation, fertilizer production, and energy consumption, also play a prominent role. In comparison, slightly newer technologies involving the use of rooftop greenhouses proved to be a more sustainable option, generating 10% less environmental burden compared to the heated glass greenhouse in terms of potential acidification. However, it is worth mentioning that this technology exhibited a 33.3% increase in potential acidification compared with the ethylene-vinyl acetate plastic film greenhouse [32]. Finally, in crop fertilization activities, it has been observed that potential acidification can be up to three times greater in heated greenhouses than in simple, unheated tunnels. Sulfur dioxide and nitrogen oxide emissions cause this increase [109]. These variations are mainly attributed to the implementation of automation technologies and the choice of energy-efficient roofing materials. This automation approach has significantly contributed to operational efficiency, and the careful selection of roof materials has positively impacted the reduction of energy consumption.

6.3.4. Human Toxicity Potential

Human toxicity potential is a commonly employed metric in LCA, serving to quantify the inherent toxicity burden associated with emissions from a process or product [191]. Within greenhouse agriculture, the potential for human toxicity is linked to the use of agrochemicals, including pesticides and fertilizers. The application of these chemicals can result in residues in crops or contamination of soil and water, posing a risk of toxicity for humans consuming the cultivated products [192]. Additionally, the manufacturing process of raw materials for greenhouse construction and the use of fossil fuels for air-conditioning contribute to the negative environmental impact quantified in this category [109].

It is noteworthy that LCA analyses integrating toxicology with greenhouse production are still limited due to differences of opinion among various authors and a lack of consensus on the methodology for pesticide evaluation [183]. Studies addressing this impact category reveal that in heated greenhouses, tomato cultivation with air-conditioning practices contributes five times more to human toxicity than tunnels for soil cultivation and multi-tunnel greenhouses for soilless cultivation. However, when examining the impact derived from pesticide use, it was observed that this impact is 5 and 22 times higher in tunnel cultivation with soil compared with multi-tunnel greenhouse and heated greenhouse scenarios [109]. Bojacá et al. [101] reported that in Colombian greenhouses used for tomato production, pest and disease control practices contribute up to 45% of the impact in this category. Furthermore, implementing climate-smart agriculture practices has been shown to reduce the impacts of this category by up to 12% through precise control of irrigation and fertilization [27]. Therefore, adopting responsible environmental practices in plant nutrition and the management of chemical agents for pest control is essential to mitigate the negative impacts of greenhouse production [82].

6.3.5. Ozone Depletion

The ozone depletion potential is an environmental indicator that reflects the reduction of the Earth’s ozone layer in the atmosphere. This phenomenon is associated with the presence of halocarbons, which are compounds that break down ozone molecules and contribute to ozone layer depletion. The consequences include increased radiation, including ultraviolet radiation, reaching the Earth’s surface, potentially leading to various climatic alterations with implications for ecosystems and human health [193].

Greenhouse production contributes to ozone depletion primarily through the use of ozone-depleting substances. Certain refrigerants, solvents, and blowing agents employed in greenhouse air-conditioning and thermal insulation systems may contain chlorofluorocarbons and halons [194]. When released into the environment, these substances can reach the stratosphere, decomposing and releasing chlorine and bromine atoms that participate in ozone-depleting chemical reactions [195]. Additionally, the excessive or inappropriate use of specific fertilizers containing nitrogen compounds can result in the emission of nitrogen oxides [192].

In greenhouses, the ozone depletion potential can be reduced by up to 40% compared with that in open fields when proper fertilization practices are employed [90,110]. Similarly, the combustion of fossil fuels such as natural gas or oil, which are used in air-conditioning, tillage, and transportation, can release chemical substances such as sulfur dioxide and nitrogen oxides. Activities in greenhouses contribute between 10% and 60% to the ozone depletion potential [90,109].

6.3.6. Abiotic Depletion

The abiotic depletion potential is centered on measuring the use of non-renewable resources such as minerals and fossil fuels throughout the entire lifecycle of greenhouse production [196]. A typical use of abiotic depletion potential is to tackle direct reliance on non-renewable resources [197]. A lower abiotic depletion potential value not only signifies decreased strain on these resources, implying more environmentally friendly practices but also acts as a crucial tool for making strategic decisions to enhance the efficiency of agricultural production [198].

Abiotic depletion, particularly during the fertilization stage, has emerged as a crucial factor in the environmental impact of tomato and other vegetable production, contributing up to 30% to this category [101,110]. Moreover, elements such as galvanized structures and steel cables in metal greenhouses can impact abiotic depletion by 16%. The introduction of recirculation systems adds significant variability, as both the infrastructure supporting crops and water recirculation can contribute up to 40% to this impact. Key materials such as polyethylene covers, polypropylene in gutters and doors, and mosquito nets also exert a notable influence, emphasizing the importance of considering various aspects of greenhouse design and operation [119]. Abiotic depletion resulting from energy consumption ranges between 30% and 35%, emphasizing the pressing need for energy-efficient strategies in greenhouse agricultural practices [55,89].

In the particular scenario of tomato production in suburban Beijing, a comparison between conventional and organic farming systems reveals intriguing trends. While the organic system requires 31.5% less non-renewable energy resources, there is a noticeable 1.73% increase in water depletion. Furthermore, the land area expansion for organic production is 29.98% higher compared to conventional production [83]. These findings emphasize the importance of identifying sustainable trade-offs between resource demand and territorial expansion in organic agriculture.

In the specific context of greenhouse fertigation techniques for tomato cultivation, a noteworthy contribution to abiotic depletion is observed, particularly in the category of metal depletion. The greenhouse structure, which is characterized by a significant use of iron, and the fertigation system are the primary drivers of this impact [29]. Tomato cultivation accounts for 34% of this impact, which is primarily attributed to the plastic used to cover the greenhouse [96]. In different crops, such as apricots cultivated in greenhouses, the structure, irrigation system, and canopy material can contribute up to 100% to abiotic depletion [80]. Similarly, studies focused on green bean production have revealed significant results, with abiotic depletion values up to 10 times higher in open fields than in greenhouse systems, primarily due to low efficiency in the use of water and fertilizers [105].

When considering greenhouse construction materials, wood and plastic greenhouses emerge as environmentally favorable options, displaying up to 30% less impact in the abiotic depletion category compared with metal and glass greenhouses [119]. Consequently, the selection of materials and cultivation methods significantly influences the environmental sustainability and efficiency of greenhouse agricultural systems.

6.4. Interpretation

Ensuring robust and reliable information for investors and stakeholders in greenhouse agriculture requires thoughtful methodological decisions during the LCA interpretation phase [199]. These decisions should prioritize transparency, thorough documentation, and consideration of the specific context and objectives of the analysis, aiming for a comprehensive assessment of the environmental sustainability of the studied agricultural system [200]. Moreover, addressing uncertainties through sensitivity analysis and scenario assessments becomes crucial, given the inherent variability in agricultural processes. Agriculture stands out as a sector with particularly uncertain impact results due to the natural variability involved in its processes, making the evaluation of uncertainty a mandatory necessity [201].

The aim of interpreting LCA results is to extract valuable insights and offer a succinct summary of both inventory analysis and impact assessment outcomes [185]. Beyond mere information provision, this phase plays a crucial role in delivering practical recommendations to enhance sustainability and mitigate the environmental impact of protected agricultural systems. A pivotal consideration guiding the interpretation phase is frequently the identification of inputs or impact categories that underscore the unique challenges within the region under study. Presently, this phase relies heavily on the water–energy– food nexus approach, a comprehensive framework that tackles the vital interconnections between these key elements in greenhouse agriculture [84].

Concerning LCA studies with uncertainty analysis, we observed that there are not many of them and that stochastic modeling or a probabilistic approach using numerical methods is the prevailing way to deal with uncertainty. We highlight the study by Romero-Gámez et al. [77]. They conducted a quantitative uncertainty analysis using Monte Carlo simulation to compare the environmental impacts of three cultivation scenarios: greenhouse, screenhouse, and open field. The results emphasized that the manufacturing of steel structures, production of perlite, use of high-density polyethylene plastics, and consumption of electricity for irrigation and fertilization generated the greatest environmental impacts. The greenhouse scenario showed the greatest impact in most categories, being up to 96% greater than the scenarios with a screenhouse and open field in terms of eutrophication. The identification of these burdens allowed valuable conclusions to be drawn for improving environmental performance in protected agricultural systems, underscoring the importance of considering both methodological decisions and uncertainty in interpreting LCA results.

In the same vein, Gil et al. [34] recently conducted research that employed stochastic multi-attribute analysis (SMAA) to assess the environmental performance of agricultural systems, addressing common limitations in conventional LCA, particularly related to uncertainty and the correlation between impact categories. Unlike traditional LCAs that often focus on averages or mean values, neglecting the inherent variability and correlation in agricultural production, the SMAA method provides advanced insight by offering a single unbiased indicator that stochastically incorporates individual results from selected categories. This improved approach not only overcomes the limitations of traditional methods but also enhances the accuracy of detecting significant differences in LCA comparisons. In addition, the SMAA method tackles potential bias in weighting procedures, providing an objective assessment for selecting the most appropriate alternative from an environmental performance perspective. In summary, this study underscores the importance of considering advanced methods such as SMAA to address the shortcomings of conventional approaches for environmental impact analysis and uncertainty management in LCA.

7. Opportunities for Future LCA in Protected Agriculture

Recommendations and priorities to improve the applicability of LCA have been continually reported [128,146,185,202]. The reported findings thus far highlight challenges and knowledge gaps in each of the four phases of LCA development. Next, we compile and propose interventions to guide future research.

7.1. Comply with the Standards Required by Phase

We found that several LCA studies in protected agriculture did not comply with the ISO standards. All phases showed methodological gaps. We observed superficial descriptions of functional units and system boundaries. Furthermore, there is limited use of normalization and weighting processes in the third phase and uncertainty analysis in the fourth. New approaches are needed to encourage the acceptance of normalization and weighting, as well as the communication of single scores within the broader community [203]. In addition, we strongly recommend more frequent and complete reporting of uncertainty analyses. Identifying, quantifying, and communicating uncertainty in LCA studies is vital for determining confidence levels in the findings. These limitations do not make a study illegitimate [203]; however, they make methodological harmonization and comparison between studies difficult.

7.2. Strengthen Under-Addressed Impact Issues

The life cycle impact assessment of protected agriculture does not provide data on land degradation or nutritional quality. These aspects are associated with ecosystem services and are rarely considered in other agricultural systems [202,204]. This situation limits the adequate perspective needed to robustly dimension environmental trade-offs. Therefore, it is necessary to expand the system of agricultural LCA indices and standardize the respective evaluation methodologies.

Ecosystem services are “blind spots” in LCA for protected agriculture, according to van der Werf et al. [202], because less than 10% of the literature mentions them. LCA research has focused on supporting and regulating ecosystem services, such as nutrient cycling, soil formation, biodiversity, pollination, water and air quality, waste decomposition, climate, and pest control [24,205]. However, provisioning ecosystem services has received little attention. Aspects of raw material management, including biomass, water, genetic resources, and biofuels, provide ecosystem services [206,207].

7.3. Synergistically Integrate LCA with Other Frameworks

Existing research on protected agriculture reflects that few frameworks are used alongside LCA. Cucurachi et al. [146] suggested that LCA should not be used in isolation but should be complemented using other methods. In this way, a more complete perspective of social, economic, and even environmental aspects converging in protected agriculture can be obtained, leading to improved effectiveness in decision making. Van der Werf et al. [202] proposed integrating frameworks focused on ecosystem services, McLaren et al. [128] proposed integrating frameworks focused on nutritional and dietary factors, and Fan et al. [185] proposed integrating frameworks focused on agent-based modeling, data envelopment analysis, and multi-criteria decision analysis. The joint use of the previous three methodologies and LCA has not been explored in protected agriculture. A frequently asked question in the literature is how can LCA become acquainted with assessing sustainability on various temporal and spatial scales [208]. The combined application of the second and third interventions described here, i.e., the previous and current interventions, can be regarded as part of the solution.

7.4. Build Regional Databases

Measuring and modeling the performance of agricultural systems is a complex task [202]. The characteristics of natural resources, such as soil, climate, and ecosystems, strongly influence and vary by region [185]. Therefore, the use of a general global database is a disadvantage that has been widely recognized [128,146]. Biases and inaccurate interpretations can easily occur, affecting decision making. Urgent action is needed to regionalize databases on the necessary local spatial scale [203]. It is crucial to recognize that the backbone of LCA is an up-to-date, reproducible, consistent, and quality-controlled life cycle inventory dataset. Computational frameworks to enhance the effectiveness and efficiency of data collection in LCA have been established [209,210]. Countries such as Switzerland and Peru have shown interest [211,212]. Government efforts guided by global agendas can promote, establish, and finance this goal.

7.5. New Research Horizons

We recommend a comprehensive exploration of the LCA environmental performance of modern integrated greenhouses under different climatic conditions. This includes analyzing effective energy-saving methods (e.g., photovoltaic modules, solar collectors, heat pumps, and other integrated modules) for the design of greenhouses, considering their structures (geometry, orientation, cladding material), ventilation, and lighting systems.

Another avenue for research lies in the implementation of circular agricultural practices, necessitating a thorough understanding of the overall performance of the production system concerning circularity and sustainability. This can be effectively accomplished by integrating circular economy indicators—such as the material circularity indicator, material efficiency metric, product circularity indicator, longevity indicator, and others—into the LCA [185,213]. Simultaneously, applying the circular life cycle sustainability assessment framework, which encompasses environmental and circular dimensions, also extends to economic and social aspects, providing a more holistic evaluation [185,214]. The synergy between the circular economy and protected agriculture is gaining strength to the point that its benefits for the sustainable transformation of territories are beginning to be examined [185,215].

It is expected and suggested to continue studying the integrated use of LCA and the “Nexus” approach (i.e., energy-water-food nexus) since it has enormous potential to quantify the environmental burdens of the systems, considering the interdependent relationships between resources. We underline the pressing need to provide more evidence through LCA in the debate on organic agriculture versus conventional agriculture or the pragmatic integration of these under greenhouse structures, considering the specific needs and capacities of each region. Finally, we highlight the importance of research that delves into the ecosystem services provided by protected agriculture, creating innovative versions of methodologies and establishing new methodological frameworks that include indicators and protocols intrinsic to the missing categories of ecosystem services.

8. Conclusions

The world of LCA is evolving rapidly, exerting influence across various sectors and shaping sustainable practices within protected agriculture. While the implementation of LCA in protected agriculture is not recent, its current popularity is on the rise. Spain, Italy, and China dominate scientific productivity in this field. Research topics vary across continents, with particular interest in urban agriculture, biofuels, hydroponic systems, the circular economy, wastewater management, biomass, and spatial limits. The research carried out encompasses four general lines of research: water management, energy management, technology management, and impact measurement.

The application of LCA unveils discernible patterns aligning with the four standard phases. Concerning the objective and scope, tomatoes, along with cucumber and lettuce, are the most researched foods. Moreover, the cradle-to-gate boundary system is widely implemented. Although the studies largely considered the mass unit of the final product, there is no uniformity in the functional unit chosen, making LCA results difficult to compare. For the inventory, the predominant inputs commonly reported include the application of fertilizers and pesticides, water management practices, and greenhouse infrastructure, with transportation receiving notably scant attention. Eutrophication potential is the most measured impact category in the third phase of the LCA, followed by global warming potential and acidification potential. However, integrating more impact categories is necessary to elucidate the trade-offs more robustly. The interpretation phase is strongly guided by the water–energy–food nexus approach, considering the challenges of the region under study, but largely lacks uncertainty analysis.

Overcoming major bottlenecks requires addressing the negligent approach to the standards required by the four phases, making limited efforts to apply evaluation frameworks that complement the LCA, and emphasizing indirect effects. As possible next steps, consolidating regional databases and developing environmental impact indicators for sustainable agriculture are critical.