In the EU, greenhouse structure and cultivation show considerable variation, ranging from complex energy intensive structures that heavily regulate the indoor climate to simple structures that resemble open-field practices. This variation is dependent on a variety of factors, including, but not limited to, local climatic and socio-economic conditions [1
]. According to the FAO, an estimated 405,000 hectares of greenhouses are spread throughout the EU [2
], a figure that includes both glass and plastic covered structures. By contrast, data from Eurostat provide an overview of the area of vegetables, flowers and permanent crops under glass in the EU-27 from 2005–2013, highlighting that there are around 135,000 hectares of glass covered (excluding plastic covered) greenhouse cultivation in the EU [3
Greenhouse agricultural production is generally seen as one of the most intensive parts of agricultural production [4
] as compared to energy use in open-field agriculture [6
], however the role of greenhouse agriculture in the European food systems is not well-documented [1
], while a small number of studies have been conducted that document energy use in greenhouse production. In 2013, the European Commission published a report on greenhouse production, however this only includes only some minor energy use statistics [5
]. Campioti et al. (2012) investigate some of the energy parameters in the greenhouse sector in four European countries [7
]. Mohamed et al. (2017) investigate the energy profile of greenhouses in Cyprus [8
], while Wageningen University releases an annual report monitoring energy use within the Dutch greenhouse sector [9
]. Regarding geography, the Netherlands in particular has the largest and most relevant energy use data available [9
]. Nikolaou et al. (2021) investigates energy parameters in greenhouse tomato and cucumber production in semiarid Mediterranean areas [10
]. Other studies that look at agricultural production or energy use in the agricultural sector generally provide no data on energy use in greenhouses specifically [11
In this context, where it is clear that greenhouse cultivation is a significant energy consumer but there is relatively little data available, and taking into account the implementation of EU climate targets and its farm to fork strategy in agriculture, it is apparent that there is a need to move towards sustainable production practices. In order to understand the future of greenhouse production in the EU and how they can contribute to a sustainable transformation of our agricultural systems, it is necessary to understand the energy use in greenhouses. As such, the goal of this paper is to provide an overview of the available data on energy use in the EU and to discuss significant transitions occurring in the sector. The rest of this section provides a brief overview of greenhouses in the EU and their energy use parameters, Section 2
discusses the methodology, Section 3
provides the results presenting available energy use data for greenhouse cultivation in the EU, and Section 4
provides an overall discussion, potential future scenarios, and a proposed framework for measuring energy use in greenhouse agriculture.
Greenhouses in the EU and Their Energy Use Parameters
Greenhouses are complex structures which aim to create ideal conditions for plant growth and production throughout the year, by controlling temperature, humidity, water, light, and carbon dioxide [12
]. There are different types of greenhouses currently in operation in the EU, ranging from intensive structures that heavily regulate the internal environment to those that are solely plastic sheet covered structures in which production inputs are similar to open-field crops [5
]. Over the last two decades, the technology associated with the construction of and agricultural production within advanced greenhouses has progressed considerably with significant changes in design, materials, agricultural techniques, etc. Consequently, the potential yield in ‘technology intensive’ greenhouses has seen dramatic increases, for instance, what is considered a ‘good’ tomato yield has increased from 100 tonnes per hectare to 600 tonnes per hectare in recent years [13
There are a variety of options for heating a greenhouse, including air heaters, central heating through pipes, boilers, cogeneration, natural gas, electricity, heat pumps [14
]. Energy sources vary, with small systems often running on direct fossil fuels, such as gas or oil, while larger systems may run on cogeneration/combined heat and power from power plants, or geothermal heat pumps. It is important to note that a large proportion of greenhouses, especially in Southern Europe, are not heated, while in temperate and northern European countries greenhouses often have large heating requirements. Conversely, there are a number of options for cooling a greenhouse, including shading, whitening, natural, mechanical or evaporative cooling/ventilation. One or a combination of these are utilised in areas that have high average temperatures for parts of the year [5
All production in greenhouses requires irrigation for which there are a variety of systems, generally powered through fossil fuels directly or via the electricity network [5
]. Other areas that consume direct energy include lighting and machinery use. Greenhouses generally are also dependent on a number of indirect sources of energy; energy that is not directly used in agricultural production but can clearly be assigned to the agricultural sector. Indirect energy use includes energy use associated with the building of greenhouses and greenhouse machinery, as well as energy embedded in the production of fertilizers and pesticides [15
There are a number of trends and parameters that affect the energy inputs of greenhouses. Multiple studies highlight that older systems are generally less energy efficient and require higher energy inputs per output [16
]. Geography and climate conditions also clearly play an important role [18
], while the type and philosophy of agricultural practices and techniques-conventional, organic, conservation, soilless cultivation, hydroponic, etc., employed in greenhouses are also likely to impact energy inputs. In practice, greenhouse systems in the EU vary significantly ranging from commercial large-scale systems to small-scale subsistence systems and their design and structure varies considerably, having large implications on energy use and energy sources [5
The extent of greenhouse production per country is not well documented. For most EU countries, other than the Netherlands, there are rough estimates on the extent of greenhouse cultivation, in particular regarding greenhouses with a low technological profile. Eurostat used to provide an indicator on the area of vegetables, flowers, and permanent crops under glass EU-27 (ha), this has now been discontinued. Based on the data presented in Table 1
, from Eurostat, the countries covered in this study cover around 72% of the total area under glass in the EU. In addition, in order to overcome the general lack of detailed available information on greenhouse production this study focused on the ‘hubs’ that are part of the Horizon AgroFossilFree project as this allowed local agricultural experts in each country to provide us with relevant data.
2. Materials and Methods
2.1. Search Strategy and Selection Process
This study conducts a review of the available literature on energy use data in greenhouse cultivation within the EU. Our search strategy used a key word search of Google Scholar and SCOPUS and the selection criteria for a paper’s inclusion in this study were: the presentation of energy use data based on either LCA methodologies or surveys, whether they were peer-reviewed, and whether they were published by well-respected journals or institutions. In some cases, data from reports were used if they were extensively cited in existing literature. The methodology, relevancy and accuracy of each study was evaluated for relevancy and accuracy and the wider project partners of the AgroFossilFree consortium [20
]. Overall, 24 relevant published articles and reports were located in the first part of the search process. Of these, after the evaluation process, 13 were discarded due to methodological issues. From the remaining 11 studies, data were extracted and used for this review.
For this study, energy use was defined according to operations and included direct energy uses—on-farm operations, heating and cooling, lighting, irrigation—and indirect energy uses—energy embedded in the production of fertilizers and pesticides. Energy assigned to the production and establishment of agricultural infrastructure, such as agricultural machinery and greenhouse constructions, was not included due to difficulties associated with accurately measuring and reporting on this and most studies do not provide data on this. This is mainly because agricultural infrastructure is often in use for multiple decades and undergoes a variety of production processes which makes it difficult to accurately assign a yearly energy use values to these components. These energy input data for greenhouse production were extracted from each study and compiled and compared to existing studies for verification.
The research questions that guided our research included: what are the specific energy sources in greenhouse production processes? What are the energy needs for different types (high intensive to low intensive) greenhouses? How do the energy needs of greenhouses vary depending on geography and climate across the EU?
2.2. Types of Greenhouses
It is widely accepted that significant variations in the energy intensity and overall energy use between greenhouses exist and are dependent on a variety of factors. In practice, this variation makes it difficult to collate results without further assigning greenhouses to conceptual categories. Due to this, this study divides results according to high and low energy intensity greenhouses. The benchmark employed for this is that all studies with energy inputs below a 1000 GJ/ha were assigned as low energy intensity greenhouses and everything above 1000 GJ/ha as high energy intensity greenhouses. In practice, such an approach works well for this study as all studies were either much lower or higher than 1000 GJ/ha allowing for more meaningful comparisons and is justified as it splits energy use into intensive systems that use extensive climatic control, generally energy inputs in these systems are around 8–12 times more in these systems, and non-intensive systems that implement minimal climatic control, generally energy inputs are around 8–12 times less in these systems.
In addition, it is important to note that data on the energy use in greenhouse cultivation in the EU are fragmented and that available data are extremely limited. Therefore, our results provide data on greenhouse energy use both on a country level, for which there is some data available (for The Netherlands, Italy, Spain, Germany, Ireland), and specifically for tomato production, a crop which has by far the most available data. For a more detailed breakdown of our results, please see the Appendix A
2.3. Limitations and Bias Risk
A main risk of this study is the lack of data on energy use in greenhouse cultivation in the EU. As such, this study is dependent on a relatively small number of studies. In effect, our understanding of energy use in EU agriculture may be disproportionately influenced by these studies and in instances may not be representative of the EU as a whole. This is a crucial limitation and highlights the importance of future studies investigating energy use in greenhouses across crops and climatic conditions in the EU.
The studies that provide energy use data on greenhouses generally present data without any uncertainty attached, while there are no established standardised methodologies for accounting for energy use in greenhouses [15
]. LCAs in greenhouse systems often employ simple yet heterogenous methodologies which render comparisons difficult [22
]. In our view, in order to obtain a clearer understanding of the energy use in greenhouses in the EU, the development of standardized methodologies and terminologies for approaching the specific investigation is a prerequisite for future studies. Based on this review we discuss some of the prerequisites and framework needed for the development of such a methodology.
Furthermore, greenhouse agriculture as a term is used to refer to a wide array of production systems. In practice, this means that production systems that have little in common, except that they occur in covered structures, are labelled together. Further conceptual distinctions and categorisations, as used in this paper, would, allow for improved analysis of greenhouse production.
The data presented in this study illustrate that energy use in greenhouses in the EU varies considerably depending on the type of greenhouse, geographical area, and crop grown. In advanced greenhouse systems, heating is the dominant energy requirement and in some studies accounts for up to 99% of all energy inputs. In our view this needs more detailed research with more data points from accepted methodologies as 99% appears high. In any case, our review indicates that energy requirements for heating and cooling in energy intensive systems are so considerable that other energy inputs, such as fertilizers, are extremely minor. Indeed, it has been estimated that the heating and cooling of greenhouses represents 1.5% of Europe’s total energy consumption [41
]. Methods of heating greenhouses vary throughout the EU; gas boilers are generally popular, as are air-unit heaters for small scale installations [42
], as well as cogeneration (Combined Heat and Power) in certain countries (mainly The Netherlands) [9
]. In most cases, heating and cooling systems are powered by energy from fossil fuels but in recent years sustainable sources of heat, mainly geothermal, have been growing rapidly.
In less energy intensive systems, overall energy requirements per hectare are significantly less (50–70 times less energy per hectare) as compared to high energy systems, but generally still multiple times higher than the energy requirements of open-field agriculture. The mixture of energy inputs in these systems are split more evenly, depending on the crop and production system, between direct (lighting, heating/cooling, irrigation, machinery use) and indirect (fertilizers and pesticides). It is important to note that indirect energy sources, mainly fertilizers, constitute a considerable amount of energy inputs in low energy intensity greenhouses (6–27%).
Studies also indicate that there is significant geographical variation in energy intensity between greenhouses in the EU as, in general, more advanced greenhouses are located in northern Europe and basic greenhouse structures in southern Europe. An FAO study which focuses on greenhouse production in South-Eastern Europe (including non-EU countries) highlights that around 18% of greenhouses are glasshouses and 82% plastic greenhouses (a higher proportion than the EU average), and that of all area under greenhouses in the region 97% are not heated. By contrast, in Northern Europe, especially in the Netherlands, where most greenhouses are heavily managed, it is normal for greenhouse cultivation to be heated, consuming around 11,000 GJ/ha [9
]. This difference can to a large extent be attributed to climatic conditions as crops grown in Northern European greenhouses for large parts of the year are dependent on external heating sources. However, it is likely that other factors, including market conditions, accessibility, knowledge, also play a role.
Identifying these areas where energy use is concentrated can help support the development and implementation of energy efficiency measures (EEM) and renewable energy sources (RES) interventions that support a transition of the entire greenhouse sector. In recent years, various studies have investigated and proposed EEM in different forms: Nikolaou et al. (2021) discuss a range of practical EEM for greenhouses located in Mediterranean countries [10
], while Fabrizio (2012) finds that energy demand in greenhouses in northern Italy can be reduced by 30% by the usage of better insulated transparent materials. [17
]. Zhang et al. (2020) conduct a review of different control strategies for improving the energy efficiency of greenhouses finding that traditional control strategies combined with intelligent algorithms have become a popular way to reduce the energy demand of greenhouses and can lead to substantial energy savings [43
]. Chen et al. (2012) show that such a system can improve energy efficiencies by up to 15.1% [44
]. Various studies highlight that traditional light sources in greenhouses are relatively energy inefficient and generally generate radiant heat and highlight that switching to LED sources can lead to considerable energy savings [45
]. Studies generally show that improving the design, operation, and technology used in greenhouses can also significantly improve overall energy efficiency [43
]. However, implementing these systems can be expensive and, due to variation in greenhouses, EEM interventions are often context specific, highlighting that further research in these areas would be useful [43
]. Various studies investigate the transition to RES for greenhouse production, including the potential for: integrating solar technologies within greenhouse production [17
]; biogas and bioenergy [49
]; and geothermal [14
]. Regarding geothermal, Arpa et al. (2016) find that geothermal are economically comparable with conventional systems [14
] and Russo et al. (2014) find that geothermal heat pumps are environmentally beneficial over conventional systems. Other recent research indicates that the installation of these can also be economically advantageous as compared to alternatives [51
]. This suggests that the transition to RES, which is already occurring at a rapid pace in certain European countries, could open the door to major reductions in the dependency on fossil sources in the greenhouse sector. Similar to EEM measures, RES interventions are likely dependent on context specificity and economics, while various factors need to be taken into account to support a sustainable transition. For instance, one of the reasons fossil fuels are used to provide energy in greenhouses is that they are source of CO2
that can be used to enhance productivity [4
Taking local and international contexts into account there are various policy interventions that will likely support the transition to the use of RES and EEM in greenhouse production. In our view, these include increased taxes for fossil fuels, subsidies for RES adoption EEM interventions and higher carbon prices. For instance, direct subsidies as implemented in the Netherlands can accelerate this transition for greenhouse heating [52
]. The new CAP is generally supportive of this by including a range of policies that support the targets of the Green deal including considerable funding for eco-schemes and climate supportive policies [53
Despite our findings, there are relatively little data available and few studies were identified that investigated energy use in greenhouses in the EU, the studies that do exist employ varying methodologies and boundaries for measuring energy use ranging from LCAs to surveys to advanced models. Although this is similar to data and methodological issues experienced in other agricultural sectors [6
], in the greenhouse sector the lack of data is more considerable. In our view, the development of a widely accepted methodology for measuring energy use in agriculture is of utmost importance and its establishment will catalyse future studies on energy use. In addition, one of the limitations highlighted by this review is that greenhouse agriculture shows extreme variation but is often labelled under one term or category. In our view, an approach that further categorises greenhouse production into, for instance, low and high energy intensity greenhouses is useful for our understanding and for comparing greenhouse production across agro-climatic conditions. Our review also highlights a few other general areas for further research, including the potential of EEM and RES technologies, especially in the context of the European Green Deal, as well as on energy demands per crop and within different geographical and agro-climatic areas.
Proposed Framework for Measuring Energy Use in Greenhouse Agriculture
Based on these limitations we propose a framework that would, in our view, allow for considerably more accurate measurement and understanding of energy use in greenhouse production systems and indeed in all agricultural systems. We propose that this framework is centred around conducting several LCAs at regular intervals for all crops in different production systems in all countries. In practice this would first require the categorization of production systems and the identification of ‘standard’ systems in each country. At a later stage non-standard and non-conventional production systems can also be considered. We also recommend that a wide, accurate and measurable definition of direct and indirect energy is included in these studies. Though indirect energy is consumed prior to the farm it can clearly be attributed to the agricultural sector and is important to be included.
In such a system, an effective approach to conducting individual LCA for crops in each production system needs to be developed. For energy use we propose that the LCA methodology uses a wide, standardized, and detailed categorisation of energy inputs. The categories that we propose include energy use for: seeds; soil; growing media; pesticides; fertilizers; diesel use for machinery, heating and ventilation; electricity use for machinery, heating, ventilation, irrigation, storage, machinery. Such a process that also differentiates between the likely main energy sources, diesel and electricity, would support a detailed understanding of energy inputs in greenhouses. We consider this overall categorisation particularly important as, currently, a small number of non-standardized energy use categories are generally used between LCAs. For instance, one LCA might have the categories fertilizers, pesticides, diesel use, and lighting, while another one might have fertilizers pesticides, machinery use, and electricity. In practice, this makes effective comparisons extremely difficult and prone to errors as these categories often overlap, are not exclusive and often there is little transparency on what was included in each of these inputs.
In addition, it would be useful if results are presented as energy use per hectare, as well as energy use per kilogram/tonne and also include uncertainty and confidence intervals. A major drawback of existing LCAs in agriculture is that they generally do not include any degrees of uncertainty. This entire methodology would need to rely on detailed data that rely on accurate real-world parameters and/or accurate database reference values. For this to occur considerable effort is required for the finding of accurate and updatable parameters and reference values. Once a large number of LCAs have been conducted, using appropriate statistic tools, an effective meta-analysis can be conducted that provides accurate greenhouse energy use profiles in different countries and production systems.
In our view such a framework can also be applied to the open-field agriculture and livestock sectors. This is a large piece of work that requires significant funding to be carried out but would support a more detailed understanding of energy use in agriculture and support effective policy and interventions for the transition away from fossil-based agriculture.