Introduction

Agriculture and food systems are confronted with daunting and complex challenges, not the least of which is the ongoing effort to increase food production by 25–70% above current levels while maintaining and enhancing ecosystem resilience (Hunter et al., 2017). Traditional farming practices, on the other hand, are still used to manage an agricultural field uniformly, ignoring the inherent variability in topography, soil, crop growth conditions, and other agronomic factors (Neupane & Guo, 2019). As a result, the excessive and inappropriate use of agrochemicals, fossil fuels, natural resources, and machinery is jeopardizing the ecological integrity of agroecosystems (Singh & Singh, 2017). The prevailing discourse on the future of agriculture calls for food production to increase while becoming more environmentally sustainable (Hunter et al., 2017). Sustainable intensification is emerging as the most frequently referenced new paradigm to produce more from the same area of land by increasing efficiency, reducing waste, conserving resources, reducing negative impacts on the environment, and enhancing the provision of ecosystem services (Wezel et al., 2015). Sustainable intensification is achieved through increased inputs, improved agronomic practices, improved crop varieties, and other innovations (Tilman et al., 2011).

Precision agriculture (PA) is widely acknowledged as a contributor to farming efficiency and environmentally friendly farming practices, and it is essential to long-term intensification (Lindblom et al., 2017). It assists farmers in making precise and optimized use of crop-specific inputs, resulting in lower production costs and a lower environmental impact (Bacenetti et al., 2020; Canaj et al., 2021). Nitrogen (N) is an essential and often the most yield-limiting nutrient for winter wheat production. However, often N fertilization in wheat is commonly based on yield goals, derived by applying uniform rates without considering the spatial and temporal variability (Gobbo et al., 2022). As a result, the N supply and crop demand are misaligned, resulting in low time and space efficiency (Denora et al., 2022) and economic and environmental losses (Fiorentino et al., 2020; Gobbo et al., 2022). The precise management of N fertilizer application is essential for improving crop productivity, use efficiency and environmental sustainability. Variable-rate technology (VRT) is a pivotal technology in PA, aiming to perform site-specific chemical, lime, gypsum, irrigation water, and other farm input management across a field (Vatsanidou et al., 2020). Because it tackles in-field heterogeneity in soil N availability and crop response, variable rate fertilization provides a technique for more effective site-specific management (Stamatiadis et al., 2018). The empirical findings suggest that variable-rate fertilizer application can have both environmental and economic benefits. Many studies, however, fail to investigate the links between the environment and production, as well as the environmental and economic implications of the product's life cycle. Precision agriculture frequently necessitates the use of advanced machinery and technological systems, the construction, maintenance, and use of which may reduce the potential environmental and economic benefits of its implementation (Bacenetti et al., 2020).

The life cycle thinking has been considered one of the most fitting methodologies to deal with farming sustainability. Life cycle assessment (LCA) is widely regarded as the most effective method for assessing the impact of crop production-related emissions and resource consumption. It generates a better understanding of the energy, water, and material inputs and evaluates the output impacts of any production system from a life cycle perspective. LCA has been carried out on various precision agriculture applications, including irrigation (Canaj et al., 2021; Fotia et al., 2021); fertilization (Bacenetti et al., 2020; Jovarauskas et al., 2021; Li et al., 2016; Meza-Palacios et al., 2020; Sanches et al., 2021; Vatsanidou et al., 2020); mechanized field operations (Ashworth et al., 2022; Lagnelöv et al., 2021; Lovarelli & Bacenetti, 2017); and land leveling (Nguyen-Van-Hung et al., 2022). It is applied to olives in Greece (Fotia et al., 2021; van Evert et al., 2017), zucchini in Italy (Canaj et al., 2021), rice in Italy (Bacenetti et al., 2020) and Asia (Nguyen-Van-Hung et al., 2022), pear orchards in Greece (Vatsanidou et al., 2020), nectarines in Greece (Núñez-Cárdenas et al., 2022), corn in the USA (Li et al., 2016), vineyards in Greece (Balafoutis et al., 2017; Pradel et al., 2022), wheat in Lithuania (Jovarauskas et al., 2021) and sugarcane in Brazil (Sanches et al., 2021) and South Africa (Van Der Laan et al., 2015). Previous LCA studies in wheat production (Fabiani et al., 2020; Jovarauskas et al., 2021; Kazlauskas et al., 2021; Medel-Jiménez et al., 2022; Scuola et al., 2017) found that variable fertilization rates may reduce overall energy consumption and greenhouse gas (GHG) emissions. However, other direct and indirect environmental benefits from the reduction of synthetic resources in crop production could be realized. Understanding how alternative agricultural input efficiency, such as variable rate fertilization, contributes to a variety of environmental effects is essential for reducing crop production's environmental impact. This study applied life cycle energy analysis (LCEA) and a multi-indicator life cycle assessment (LCA) to evaluate the energy performance, environmental impact, and external environmental costs of durum wheat production in southern Italy by using different N fertilization strategies: variable rate technology (VRT) and uniform application (UA). The findings provide the first detailed assessment of the energy and environmental benefits that can be realized when precision farming technologies are used to support N fertilization in rainfed wheat production in a Southern Mediterranean context. Moreover, the study is the first of its kind to estimate the indirect economic benefits of variable rate fertilization in cereal crops by monetizing the LCA results.

Material and methods

Case study and system description

The data for this study were retrieved from field data collected in 2018–2019 at Genzano di Lucania (Potenza province, Basilicata region), latitude: 40.82° N, longitude: 16.08° N. The Basilicata region primarily produces cereals, accounting for 72% of arable land. The experimental field had a total area of 4.07 ha. The area is located on the clayey hills of the Bradanica grave and the basin of Sant’Arcangelo (Fig. 1).

Fig. 1
figure 1

Location of the study site and delineated maps of N fertilization in uniform management and variable-rate application

Full size image

Across the whole field, wheat was sown with a row spacing of 0.13 m, and 250 kg ha−1 of seeds were used. Soil tillage consisted of a 40 cm deep plowing (August 28, 2018) and two harrowings (November 11, 2018, and December 5, 2018). Pre-sowing fertilization was broadcast applied with 92 kg ha−1 of P2O5 and 36 kg ha−1 of N. A dose of 35 kg ha−1 of N (Urea 46%) was spread in pre-sowing over the entire field. In the uniform application (UA) plots, we applied a dose of N equal to 85 kg ha−1, which corresponds to the amount generally applied by the farmer, and slightly over the average, the dose of N applied in the three zones. The amount of nitrogen fertilizer to be applied by VRT was calculated based on estimated crop nitrogen uptake and soil characteristics of the area determined by electrical resistivity (Denora et al., 2022). Crop potential N uptake was estimated using the previous year's crop yield in each homogeneous area, and was corrected to account for the N contribution provided to the crop by organic matter mineralization. Soil property maps derived from low induction electromagnetic measurements were used to calculate N balances for a field application of VRT nitrogen fertilization. A low-induction electromagnetic mini explorer (GF Instruments Brno-CZ) was used to investigate the spatial variability of the soil. For the variable rate nitrogen treatments, the final prescription map was created using the QGIS 2.18.4 software, and N doses were applied in each homogeneous area using a Kuhn Axis-40–2-w fertilizer spreader mounted on a John Deere 6910 tractor.

LCA modeling

This LCA study was based on the LCA framework's four main phases: goal and scope definition, life cycle inventory, life cycle impact assessment, and life cycle interpretation of results.

Goal and scope

In this study, a cradle-to-farm gate LCA study was performed. Crop cultivation started with tillage for seeding; after that, seeding occurred, plant protection and fertilization were performed for crop growth, and at the last stage, harvesting took place. A flow chart of the system boundary is shown in Fig. 2. The analysis also takes into account the production of seeds, fertilizers, pesticides, fuel, tractors, and human labor within the system boundary. We distinguished foreground (direct) and background (indirect) systems when analyzing datasets. Direct field and farm emissions are substances emitted from an agricultural area or directly from the farm. In our model, we accounted for foreground emissions due to agricultural operations (fuel combustion and tyre wear), fertilizer application, and emissions of pollutants (ammonia volatilization, nitrous oxide emissions, nitrate leaching, and phosphorus compound emissions). Indirect emissions denote emissions that occur in upstream processes, such as purchased inputs used in agriculture or transportation (production of seeds, fertilizers, pesticides, fuel, lubricants, and tractor units). Both hectare (1 ha) and ton of grain (1 ton) production were used as functional units to highlight possible contrasting results on crop yield and the effect of agricultural intensification. No allocation criteria were used for allocating the impacts because it was assumed that straw was left on the field.

Fig. 2
figure 2

A flow chart diagram for the system boundary for wheat production

Full size image

Life cycle inventory (LCI)

The inventory data are summarized in Table 1. The direct agricultural input data (foreground system), such as seed rate for sowing, plant protection product, fertilization amount and types, fuel consumption, and machinery working hours, were collected at the farm during field tests and surveys. Nitrogen emissions (nitrate leaching, ammonia volatilization, and nitrous and nitrogen oxide emissions in the atmosphere), phosphate emissions in water, and fossil CO2 emissions to the atmosphere were calculated using Koeble (2014) and Nemecek et al. (2020) guidelines. N2O emissions from atmospheric deposition of N on soils and water surfaces and emissions from N leaching and runoff were included in the indirect emissions. Direct N2O emissions were equivalent to 1% of the amount of N applied as fertilizer (0.01 kg N2O–N). Ammonia volatilization was considered to be 0.1 kg NH3–N per kg of N. The indirect N2O from atmospheric deposition was 0.01 kg N2O–N per kg of NH3–N whereas the indirect N2O from leaching/runoff was 0.0075 kg N2O–N per kg of NO3–N. The nitrate–nitrogen leaching loss was considered 0.22 kg NO3–N per kg of N for UA and 0 for VRT. In the VRT strategy all the N given with the fertilizer was taken up by the crop, whereas in the UA strategy, only 22% was lost. For urea, the emission is 1.57 kg CO2 per kg Urea–N. The secondary emissions of the inputs during the production stage, including fertilizer, agrochemicals, machinery, and infrastructure production, were retrieved from the Ecoinvent database (Ecoinvent Database 3.1 2014).

Table 1 Inventory data for 1-ton wheat production under uniform (UA) and variable rate fertilization (VRT)
Full size table

Energy analysis and life cycle impact assessment

The performance assessment included energy input-output and a series of life-cycle environmental impacts. To evaluate the energy performance, various energy indices such as energy consumption, energy use efficiency (EUE), net energy gain (NEG), energy productivity (EP), and specific energy (SE) were used (Table 2). The energy input was obtained as a product of each input and its corresponding energy coefficient. It was classified into direct and indirect, and renewable and non-renewable. The total energy input was calculated as the sum of all energy inputs for all resources used in crop production. The output energy was obtained as a product of yield and its equivalent energy representative.

Table 2 The average value of energy equivalent coefficient of inputs and outputs
Full size table

Energy use efficiency (EUE) was calculated from the ratio of energy output and energy input (Eq. 1). An increase in the ratio indicates an improvement in energy efficiency.

$$\text{Energy use efficiency }=\frac{\text{Energy output }(\text{MJ }{\text{ha}}^{-1})}{\text{Energy input }(\text{MJ }{\text{ha}}^{-1})}$$
(1)

Energy productivity (EP) was measured from the ratio of crop output of wheat and energy input (Eq. 2). An increase in the indicator denotes high EP and vice versa.

$$\text{Energy productivity }(\text{kg }{\text{MJ}}^{-1})=\frac{\text{Crop output }(\text{kg }{\text{ha}}^{-1})}{\text{Energy input }(\text{MJ }{\text{ha}}^{-1})}$$
(2)

Specific energy (SE) was estimated from the ratio of energy input and crop output (Eq. 3). An increase in the indicator denotes lower energy efficiency and vice versa.

$$\text{Specific energy }(\text{MJ }{\text{kg}}^{-1})=\frac{\text{Energy input }(\text{MJ }{\text{ha}}^{-1})}{\text{Crop output }(\text{kg }{\text{ha}}^{-1})}$$
(3)

Net energy gain (NEG) was approximated by the deduction of input energy from output energy (Eq. 4).

$${\text{Net energy}}\left( {{\text{MJ}}\,{\text{ha}}^{{ - 1}} } \right) = {\text{Energy output}}\left( {{\text{MJ}}\,{\text{ha}}^{{ - 1}} } \right) - {\text{Energy input}}\left( {{\text{MJ}}\,{\text{ha}}^{{ - 1}} } \right)$$
(4)

The life cycle impact (LCIA)-model ReCiPe 2016 (Huijbregts et al., 2017) was used to analyze environmental performance. We calculated twenty-one (21) environmental indicators (Fig. 3): eighteen (18) at the midpoint level (e.g., global warming, acidification, eutrophication, and toxicities) and three (3) at the endpoint level (human health, ecosystem quality, and resources). Midpoints were used for a more specific and detailed analysis, whereas endpoints were used to communicate the results obtained to a broader, non-expert audience. To easily compare the environmental impact of fertilization strategies, a single score index was calculated by aggregating environmental impacts into a single score expressed in a physical value (ReCiPe single score) (Fig. 3). Afterward, the computed environmental impacts were converted into externalities (environmental costs) by applying monetization weighting factors (Canaj et al., 2021). Monetizing LCA results is one way of expressing environmental impacts in terms of costs. The openLCA 1.10.3 software (https://www.openlca.org/) was used to model the study system and to calculate the selected performance indicators. The standard deviation of the impact categories was simulated as a function of seed rate (± 10%), crop yields (± 10%), diesel fuel (± 10%), and fertilization rates (± 10% and ± 20%).

Fig. 3
figure 3

ReCiPe 2016 impact pathway from inventory to aggregation to a single score

Full size image

Result and discussion

Energy performance indicators

Figure 4 and Table 3 show the results of the energy analysis for wheat production. The energy input was calculated to be 7113.3 ± 729.3 MJ t−1 and 6282.8 ± 438 MJ t−1 for UA and VRT, respectively. Fertilization used the most energy (Fig. 4), accounting for 59% and 53% of total energy consumption for UA and VRT, respectively. In rain-fed wheat production, chemical fertilizers are one of the top contributors to total energy consumption and environmental footprint (Canaj & Mehmeti, 2022; Ilahi et al., 2019; Taki et al., 2018).

Fig. 4
figure 4

Process input energy required for rainfed durum wheat production using uniform (UA) and variable rate fertilization

Full size image
Table 3 Indicators of energy performance for wheat production with uniform (UA) and variable rate fertilization
Full size table

Table 3 presents the energy use efficiency (EUE), net energy gain (NEG), energy productivity (EP), and specific energy (SE) scores. In wheat production with UA, the EUE, SE, EP, and NEG were calculated as 1.83 ± 0.18, 7.11 ± 0.73 MJ kg−1, 0.14 ± 0.014 kg MJ −1, and 15 659 ± 3325 MJ ha−1, respectively. The values for wheat with VRT were 2.07 ± 0.14, 6.28 ± 0.44 MJ kg−1, 0.16 ± 0.011 kg MJ−1, and 18 084 ± 730.7 MJ ha−1. Accordingly, VRT increased EUE and EP by 13.3%, reduced SE and total energy inputs by 11.7%, and increased NEG by 15.3%. Both systems relied on non-renewable energy sources (> 80%). The fossil energy dependence was found to decrease in VRT, as the use of non-renewable energy decreased by 14.4% from 5884.1 MJ ha−1 to 5036.6 MJ ha−1.

Our results agree with the findings of other studies (Fabiani et al., 2020; Jovarauskas et al., 2021; Kazlauskas et al., 2021; Scuola et al., 2017), in which VRT technology improves energy performance indicators of wheat production. Kazlauskas et al. (2021) demonstrated that using VRT technology could save 5.2% of energy input (12 059 vs. 12 726 MJ ha−1) in wheat production in Lithuania. Jovarauskas et al. (2021) estimated that VRT reduced total energy input by 10.46% in Lithuanian winter wheat production, which resulted in approximately 9% higher energy efficiency (4.58 vs. 4.18) and productivity (0.327 ± 0.015 kg MJ−1 vs. 0.299 ± 0.012 kg MJ−1). In Central Italy, Scuola et al. (2017) estimated a 30.15% (12 732 vs. 18 228 MJ) reduction in non-renewable energy consumption. Fabiani et al. (2020) discovered that using VRT applications in Greek wheat production could increase EUE by 14% (2.51 vs. 2.21) and decrease SE by 12% (5.7 vs. 6.56 MJ kg−1) compared to the Czech Republic, where the authors estimated marginal effects with less than 2% benefits.

Environmental performance at the midpoint and endpoint level

Table 4 shows the results of impact category indicators at the midpoint level for 1 hectare and 1 ton of product. The findings show that VRT had a negligible impact on many environmental impacts (such as mineral resource scarcity, ozone formation, human toxicity, water consumption, and so on), with benefits of less than 5%. The VRT demonstrated a general reduction in potential impacts for 1 ha of wheat cultivated. For one ton of wheat, the VRT had a minor negative impact on freshwater eutrophication, freshwater, marine, and terrestrial ecotoxicity, and land use. In our study, the yield of wheat with VRT was slightly lower than in UA. Nevertheless, our model results show that the application of the VRT for a precise N-fertilization system allows reducing several environmental impacts, such as global warming (− 17.9%), fine particulate matter formation (− 19.7%), stratospheric ozone depletion (− 28.7%), terrestrial acidification (− 22.3%), and marine eutrophication (− 87.8%). These environmental impacts were mitigated by reducing on-farm (foreground) emissions. The higher land application of N compounds as chemical fertilizers had a negative influence on the environment through the release of N-containing gases such as NH3 and N2O, and nitrate (NO3−) losses via leaching and runoff. Further, the use of every kg of urea essentially induces CO2 emissions after its usage. The reduction of soil N2O emissions and CO2 releases after urea applications reduced global warming. Reduction of ammonia (NH3) volatilization and nitrogen oxide (NOx) emissions had the greatest impact on fine particulate matter formation and terrestrial acidification. Marine eutrophication occurred due to the nitrate originating from agricultural runoff and leaching (waterborne N-emissions).

Table 4 Average cradle-to-farm gate midpoint environmental impacts of wheat production with uniform (UA) and variable rate fertilization (VRT)
Full size table

The relative contribution of the agricultural inputs to the environmental impacts of wheat is presented in Fig. 5. For both UA and VRT, fertilizers had the greatest environmental impact (12 out of 18). Photochemical ozone formation was greatly affected by mechanized field operations (i.e., diesel fuel emissions), whereas pesticide use caused freshwater, marine, and terrestrial ecotoxicity. The greatest impact on water consumption was caused by seed production.

Fig. 5
figure 5

Contribution of agricultural inputs and processes to the environmental impacts of wheat production: A UA, B VRT

Full size image

Figure 6 depicts the numerical endpoint scores for 1 ton of product. The benefits of VRT to areas of protection (human health, ecosystems, and resources) ranged from 3.3% to 13.5% for 1 ton of product and from 4.4% to 14.2% for 1 ha of land. For UA, the damage to human health, ecosystem quality, and resource availability was 9.43E−03 ± 9.77E−04 DALY t−1, 4.15E−05 ± 4.6E−06 species.yr t−1 and 58.28 ± 6.53 USD2013 t−1, respectively. For VRT, the damage to human health, ecosystem quality, and resource availability was 8.16E−03 ± 5.52E−04 DALY t−1, 4.01E−05 ± 1.9E−06 species.yr t−1 and 52.9 ± 2.9 USD2013 t−1, respectively. The aggregation of the weighted results into a single score showed that damage to human health is controlled by fine particulate matter formation, which is due to the volatilization of ammonia (NH3). In terms of ecosystem quality, agricultural land occupation accounted for more than 47% of the footprint. The scarcity of fossil fuels is the primary determinant of resource availability.

Fig. 6
figure 6

Scores for human health, ecosystem quality, and resource availability in wheat production using uniform (UA) and variable rate fertilization (VRT), with input/process and indicator contributions

Full size image

LCA single score analysis (physical weighting)

Figure 7 depicts the aggregated single-score indicator, expressed as a physical value (ReCiPe single score). Wheat production with UA and VRT was estimated to have an environmental footprint of 182.3 ± 18.8 and 160.1 ± 11.2 points ton−1 respectively. The footprint for 1 ha was 484.9 ± 49.9 points and 421.1 ± 29.4 points for UA and VRT, respectively.

Fig. 7
figure 7

Single score environmental impact of wheat production with uniform (UA) and variable rate fertilization (VRT). A Process contribution; B Subsystem contribution; C Midpoint impact contribution; D Endpoint impact contribution

Full size image

With VRT, the fertilization environmental footprint of wheat production was reduced by 23%, from 100.7 points per ton to 77.6 points per ton. Considering the cradle-to-farm gate perspective, VRT could reduce the total environmental footprint by 12.2% per ton of product or 13.1% per hectare cultivated. The background subsystem (production and transport of N-fertilizers) was responsible for about 6% of the reduction, while the foreground subsystem was responsible for 14% (application of N-fertilizers). The highest benefits were due to the reduction of fine particulate matter formation as a result of NH3 reduction.

LCA single score analysis (external environmental cost)

Figure 8 depicts the aggregated single-score indicator, which is expressed in monetary value (EURO) and represents the external environmental cost. Wheat production with UA and VRT has external environmental costs of 1151.3 ± 80.4 and 1075.2 ± 73.2 Euros ton−1, respectively. Considering the cradle-to-farm gate perspective, wheat with VRT can reduce the external environmental cost by 6.6% for 1 ton of product and 7.7% for 1 ha of land. Differently from physical weighting, money gives more value to land occupation, an indicator that is related mainly to crop yield and no farm inputs. Production of wheat crops needs adequate land requirements (Romano et al., 2021). Land use is the main driver of global biodiversity loss, and its environmental relevance is widely recognized in research on LCA (De Baan et al., 2013), as there are external costs associated with biodiversity loss associated with land use (De Bruyn et al., 2018). The economic analysis literature indicates that the production costs of wheat production in southern Italian regions were 992 EUR ha−1 (Pazienza & Zanni, 2009), 512.52 to 693.96 EUR ha−1 (Tiberti, 2013), 379 and 784.1 EUR ha−1 (Todorović et al., 2018) and 926.5 to 1023.8 EUR ha−1 (Bux et al., 2022). These figures show that indirect costs can be as high as or higher than production costs. This confirms that the true cost performance of variable rate technology will be greatly underestimated if the environmental cost is not considered. Environmental impact monetization could be considered in cost–benefit analyses as a further evaluation attempt.

Fig. 8
figure 8

The external environmental cost of wheat production under uniform (UA) and variable rate fertilization (VRT)

Full size image

Comparison of our findings with other studies

Several LCA studies on wheat production have been conducted, but with a limited focus on the benefits of variable fertilization (Jovarauskas et al., 2021; Kazlauskas et al., 2021; Medel-Jiménez et al., 2022; Scuola et al., 2017). As a result, we provided an overview and compare findings with other several other LCA studies on variable rate fertilization that have been published internationally (Table 5).

Table 5 Literature LCA studies on variable rate fertilization
Full size table

Jovarauskas et al. (2021) and Kazlauskas et al. (2021) found that variable-rate fertilization on wheat production could reduce the GHG emissions by 5.2% to 9.5%. Scuola et al. (2017) estimated a 32% lower carbon footprint in the cultivation of bread wheat through precision agriculture in Central Italy. Further reductions were estimated for blue water, acidification, and eutrophication potential. Medel-Jiménez et al. (2022) estimated an 8.6% reduction in the climate change impact by using the ground-based optical crop sensor for variable rate nitrogen application in Austrian conditions. Other remarkable benefits were observed for freshwater eutrophication (− 21.23%), human toxicity (− 20.20%), and marine eutrophication (− 9.05%). According to Van Der Laan et al. (2015), total energy input and GHG emissions in sugarcane production in Brazil could be cut by 20% and 25%, respectively. According to Li et al. (2016), sensor-based nitrogen application in corn production in the USA could reduce life cycle non-renewable energy consumption, global warming, acidification potential, and eutrophication potential by 7, 10, 22, and 16%, respectively. Variable rate nutrient application, according to Balafoutis et al. (2017), could reduce the carbon footprint of the vineyard in Northern Greece by 28.3% when compared to conventional production. Vatsanidou et al. (2020) demonstrated the environmental benefit of variable rate fertilization by reducing air emissions from fertilizer application in pear orchards in Greece by nearly 50%. Variable-rate fertilization could reduce the environmental impact of rice production in Italy by up to 13.6% when compared to uniform N application (Bacenetti et al., 2020). Meza-Palacios et al. (2020) showed that a decision support system for NPK fertilization in sugarcane farms could reduce on average damage to human health by 11%, damage to ecosystem quality by 9%, climate change impact by 14.5%, and resource availability by 11.5%. Sanches et al. (2021) estimated that applying fertilizer at variable rates in sugarcane production could reduce climate change by 3.4% and fossil fuel depletion by 4.2% per ton of product. According to Núñez-Cárdenas et al. (2022), using precision agriculture practices in Spanish conditions could reduce the carbon footprint of nectarine production per kg of fresh fruit at the farm's gate by 20.5%. Casson et al. (2022) found that variable-rate drip irrigation and fertigation in Italian grape farms can significantly reduce the CO2-eq emissions generated during grape production by over 50%. In general, the majority of LCA studies show that variable-rate fertilizer application has environmental benefits. These benefits of VRT technology vary from study to study depending on data availability and accuracy, system boundaries, modeling approach, functional unit, and life cycle impact assessment method. Future case studies are thus required to test new indicators, new LCIA methods, and their outcomes.

Discussion

Fertilization is an essential crop input for wheat production; however, improper N application rates can result in serious environmental concerns from fertilizer production and application. Precision farming has been widely expected to show environmental benefits; however, the magnitudes of these effects are largely uncertain and case-dependent (Finger et al., 2019). Here, using a multi-indicator life cycle impact assessment model, we compared the energy and environmental impacts of wheat production under uniform and variable rate fertilization strategies. VRT resulted in a 25% reduction in nitrogen fertilizer with the same level of yield as UA. This level of nitrogen efficiency provided environmental benefits on air-related environmental indicators of particulate matter formation, global warming, and terrestrial acidification, which depended on emissions of ammonia (NH3), nitrogen oxide (NOx), and nitrous oxide (N2O). Our model results showed that the reduction of NH3 had a greater influence on the final environmental benefits of wheat production. Similar previous findings (Medel-Jiménez et al., 2022) have revealed that the amount of applied N fertilizer has a greater influence on NH3 and NO3 indirect soil emissions than on direct N2O emissions. Fine particulate matter formation is an indicator of air pollution that causes primary and secondary aerosols in the atmosphere and can have a substantial negative impact on human health (Huijbregts et al., 2017). For some environmental impacts, a minor negative effect was observed due to the effect of crop yield. According to the single-score analysis, wheat production with VRT has lower pollution-related environmental impacts per unit of product and land area. The findings, which are consistent with previous energy-related (Fabiani et al., 2020; Jovarauskas et al., 2021; Scuola et al., 2017) and LCA  research (Bacenetti et al., 2020; Medel-Jiménez et al., 2022; Vatsanidou et al., 2020), highlight the value of VRT in input management to reduce nitrogen application rates while maintaining crop productivity and providing energy as well as numerous environmental benefits. Yet, our study highlighted that the overall expected benefits of smart agricultural technologies in annual crops are not always straightforward due to trade-offs between environmental indicators. In this study, land-use impacts that are not controlled by crop yield rather than fertilization had a significant effect on the overall co-benefits or co-damages of wheat production. This suggests that the consideration of multiple metrics needs to simultaneously explore trade-offs that may exist between productivity and environmental sustainability. Higher grain yields are expected to have a lower impact on land occupation; thus, the environmental benefits of VRT could be maximized by simultaneously increasing grain yield and optimizing the fertilizer rates. Understanding the spatial and temporal interactions between soil–plant-atmosphere is required for the successful implementation of site-specific N management (Basso et al., 2016). It is demonstrated that soil type, meteorological conditions, and N fertilizer rate and type have significant implications for N availability and crop uptake (Pampana & Mariotti, 2021) and crop yield, energy performance, and economic efficiency (Jovarauskas et al., 2021). Therefore, to realize the full potential of VRT, weather, soil, and landscape data should be combined when implementing variable rate treatments.

The decision to use variable rate fertilization would be based on economic performance. Until now, literature has produced contradictory results on the profitability of such concept. Farm sizes and the level of efficiency of the “business-as-usual scenario” influence the economic impact of the VRT (Fabiani et al., 2020). To be profitable, variable rate N management must accurately match N requirements to crop N demands (Long et al., 2015). Even with an increase in yield and cost savings on crop production inputs, using VRT technology may result in high costs, especially in small-scale farming systems (Späti et al., 2021). For the first time, this paper introduces the concept of monetization life-cycle assessment results to estimate the indirect cost of wheat production under the precision management of fertilizers. Our research found that VRT can have indirect economic benefits because the indirect costs (environmental externalities as external costs) are lower than with uniform management. Thus, we emphasize that a more comprehensive LCA that includes these environmental impact monetizations is required to investigate the “true cost” performance of VRT by quantifying the cost of environmental impacts and directly integrating them with economic costs.

Conclusion

This study used a multi-indicator model and lifecycle-based indicators to compare the performance of rainfed wheat production using uniform (UA) and variable N fertilization (VRT). According to our model results, the VRT can reduce indirect energy inputs while increasing energy efficiency and productivity by at least 10%. The LCA findings show that there is a range of potential environmental benefits associated with VRT on wheat cultivation, including reductions in global warming, fine particulate matter formation, stratospheric ozone depletion, terrestrial acidification, and marine eutrophication. Our model indicated that fertilizer use efficiency drives on-farm environmental benefits (reduction of N losses due to leaching, denitrification, ammonia volatilization, and fossil CO2 emissions) more than indirect benefits (emissions that come from the manufacture of synthetic N fertilizer). Aggregating the results into a single score demonstrated that physical environmental benefits can be up to 12.2% and indirect economic benefits (hidden environmental costs) can be up to 7.7%. These results outline that VRT is a promising option for sustainable extensification and improved eco-efficiency of wheat production in a Mediterranean context.

As a result of our research, we conclude that for annual crops, multiple metrics need to be considered to explore the full range of trade-offs and synergies between different environmental indicators. The analysis shall include mass-based and land-use-based functional units to capture trade-offs between environmental performance, land use, and productivity. It is necessary to improve the methodology by combining life cycle assessment, monetization, and life cycle costing to explore the connection between direct and indirect financial implications and environmental benefits in a life cycle context. This would be a great step for the to support decision-making regarding the “true” sustainability of VRT.