Improving energy efficiency and carbon balance through conservation and precision agriculture with respect to conventional agriculture: a methodological approach

1 Introduction

Andalusia is an eminently agricultural region located in the south of Spain, accounting 21% of the national agriculture area (MAPA, 2024). Rainfed herbaceous crops occupy 25% of the region’s total agricultural area, more than 888 thousand hectares (MAPA, 2024). This activity accounts for over 6% of the Andalusian economy, compared with a national average of 3% (CAGPDS, 2020). Furthermore, agrarian employment in the region represented almost 9% of total employment in 2019 (CAGPDS, 2020).

Agriculture, like other human activities, has an associated carbon footprint. In Andalusia, agriculture accounts for 7005.07 Gg CO₂eq of the 23,962.77 Gg CO₂eq emitted by the diffuse sector (primarily transport, agriculture, and waste), representing 29.2% of the regional inventory (CAGPDS, 2022). Agricultural emissions arise mainly from the use of nitrogen fertilisers and fossil fuels (Franks and Hadingham, 2012). Fuel consumption in conventional tillage (CT) systems, such as those commonly used in Andalusian agricultural, increases compared with systems that minimise ploughing (Borin et al., 1997; Holland, 2004; Ferraro, 2012). In addition, CT often reduces fertiliser-use efficiency (Stavi and Lal, 2013), due to the oxidation of soil organic carbon (SOC) and a decline in natural fertility (Cárceles-Rodríguez et al., 2022).

Several scientific studies have shown that Conservation Agriculture (CA), based on three interlinked principles: minimal soil disturbance, permanent soil cover and crop rotations (FAO, 2025; González-Sánchez et al., 2015), provides important environmental benefits (Holland, 2004; Triplett and Dick, 2008; González-Sánchez et al., 2012: Kassam et al., 2012; Kassam and Kassam, 2020; Márquez-García et al., 2024). A range of techniques falls under the CA umbrella. Some are adapted to annual crops, such as no-tillage (NT) and cover crops. These techniques offer agri-environmental benefits, including reduced fuel consumption by avoiding tillage, which is a fuel-intensive operation. Lower fuel requirements under CA have been documented in the study area (Hernanz et al., 1995; Sánchez-Girón et al., 2007; Moreno et al., 2011). One of the potential problems associated with the long-term use of CA is soil compaction in deeper layers, which can limit water infiltration and root development (Blanco-Canqui and Lal, 2007; Blanco-Canqui et al., 2022). To minimize these drawbacks, the use of high-flotation tires and controlled traffic farming is recommended (Perez-Ruiz et al., 2012).

Fuel savings are key to improving energy balance and reducing greenhouse gas (GHG) emissions. Energy performance in agricultural systems can be assessed using parameters such as energy efficiency (EE) and energy productivity (EP). EE relates energy inputs to energy outputs, whereas EP relates energy inputs to production.

In addition to CA, the energy balance of the system and energy-related parameters can be improved through Precision Agriculture (PA). PA technologies enable more rational use of agricultural inputs, thereby reducing the energy associated with manufacturing, transport and field application. This optimization is particularly important for fertilisers (Auernhammer, 2001; Borch, 2007), as they are typically the most energy-consuming input in the agricultural production process (Khaledian et al., 2010; Alluvione et al., 2011; Paris et al., 2022).

Global Positioning System (GPS)-based guidance assistance (GA) optimises the performance of various agricultural operations (Perez-Ruiz et al., 2012; Vellidis et al., 2013), including the application of plant protection products and fertilisers, by reducing overlap between tractor passes (Zhang et al., 2002; Kharel et al., 2020).

Therefore, implementing combined NT and GA techniques can increase the EE and EP of rainfed annual crops. However, energy consumption and production are strongly influenced by crop type (Sartori et al., 2005), soil type (Arvidsson, 2010), management system (Rathke et al., 2007) and intensity (Sartori et al., 2005; Nassi o Di Nasso et al., 2011), as well as by the calculation methodology and the specific farms evaluated. Different methodologies for quantifying energy in agricultural systems have been developed worldwide and have been used for decades (Leach, 1976; IFIAS 1978; Fluck, 1992; Hülsbergen et al., 2001; Camargo et al., 2013; Hernanz, 2005).

Analysing direct and indirect energy is a basic approach for evaluating emissions through emission factors associated with each operation and input, and can reveal differences in the energy intensity and emission profile of each production system. Improving the efficiency in the use of agricultural inputs reduces emissions, as does adopting other C-efficient alternatives (Lal, 2004; Vejendla et al., 2025). These practices can also contribute directly to achieving European Union climate goals: under the European Climate Law, the EU mandates a net GHG emissions reduction of at least 55% by 2030 (relative to 1990 levels) and the attainment of climate neutrality by 2050 (EC, 2021).

Despite the increasing interest in CA and PA techniques for input optimisation, there is limited research assessing overlaps in farming operations and their influence on energy expenditure and associated emissions across different crops. Under the hypothesis that the comparison with CT can provide evidence of the potential of CA and PA practices to reduce system energy use and mitigate the climate impact of agricultural production, the main objectives of this work are to:

• Develop a simple methodological approach to facilitate energy and carbon balance calculations.
• Analyse the influence of crop type on the energy efficiency and energy productivity within a crop rotation.
• Evaluate, for each crop studied, the improvement in energy balance provided by combining NT and GA compared with CT.
• Estimate the carbon balance (CO2eq) of field operations and production inputs for the different crops and soil management based on energy consumption.

2 Materials and methods

2.1 Experimental fields

During four growing seasons, 2009/10 to 2012/13, field operations were monitored on three farms located in Andalusia, southern Spain (Figure 1). This area has a Mediterranean climate with a Xeric soil moisture regime, according to the USDA Soil Taxonomy (Soil Survey Staff, 2014), and thermic temperature regime. The climate is characterised by two contrasting periods: a cold, humid period in autumn and winter, when 80% of the rainfall occurs, and a warm dry period in spring and summer, when the plants experience severe water stress.

The experimental fields were representative of rainfed arable farming systems typical of the Andalusian countryside. Their locations and nomenclature are as follows: Córdoba (Field 1), 37° 55’ 50” N, 4° 43’ 07” W; Carmona (Field 2), 37° 25’ 31” N, 5° 3’ 01” W; and Las Cabezas de San Juan (Field 3) 36° 56’ 37” N, 5° 55’ 13” W, the latter two in the province of Seville. The physicochemical characteristics of the soils are summarised in Table 1. Methods used to determine soil parameters are described in Sparks et al (1996).

Table 1. Physical and chemical characteristics of the experimental fields (±: standard deviation).

N, total nitrogen; OM, organic matter; CO₃^-2, carbonates; pH in calcium chloride, (CaCl₂); CEC, cation exchange capacity; K, potassium; P, phosphorus; C-L: clay-loamy; C, clayey.

During the 4-year study (2009/10 – 2012/13) 90 ha were monitored (30 ha per experimental field). A typical Andalusian rotation was implemented: cereal (10 ha), sunflower (10 ha) and legume (10 ha). These crops were grown under two soil management systems (SMS): conventional tillage (CT) and no-tillage with tractor guidance assistance (NT&GA). Thus, each experimental unit (Plot) comprised 5 ha under a given crop and management combination. In summary, the same crop was grown in each plot once every three years, following the rotation described above of cereal in one season; after that the next year sunflower was grown and finally the next season the plot was cultivated with a legume. Therefore, six independent results were obtained each year, one for each combination of crop (cereal, sunflower, and legume) and soil management system (CT and NT&GA). Fields were managed according to the owners’ guidelines; therefore, the crops grown and the operations performed varied among fields and growing seasons.

2.2 Instrumentation used and fieldwork

Three tractors were instrumented to monitor the mechanised operations carried out in each field. Field 1: John Deere 6420 (82.5 kW); Field 2: Fendt Vario 712 (97.5 kW); Field 3: New Holland TM 120 (90 kW). The equipment installed on the tractors is described below (Figure 2):

Each system installed on the tractor (A, B, C in Figure 2) enabled monitoring different parameters of the mechanised operations. A potentiometer (A) detected the position of the rear hitch, indicating whether the tractor was performing a task or was in transport or manoeuvring. GPS measurements (B) were used to calculate task duration, worked area, average speed, theoretical work capacity, actual work capacity, and overlap. A flow meter (C) measured fuel consumption for each field operation. In addition, all tractors used for the GA treatments were equipped with an AgGPS EZ-Guide 500 (Trimble) guidance system installed for phytosanitary treatments and fertiliser application, and an EZ-Steer (Trimble) assisted steering system for sowing operations.

A data logger (Data Taker DT 85) with 48 analogue and digital input signals and a storage capacity of up to 5 million data was used as the data acquisition system. The system was programmed to log data every two seconds, recording the following information:

• UTM position
• Heading
• Speed
• Instantaneous fuel consumption
• Hydraulic lift position

Each day, a PC remotely connected to the data acquisition system. Data were transmitted to the PC via GPRS using an ETM9560–1 Mäternick modem.

To calculate the average crop production, the plots were subdivided into 10 homogeneous grids of 0.5 ha each. This was carried out with a Pocket PC equipped with EZ Maps (Trimble) software. Samples were taken at the centre of each grid using a 1×1 m² frame, with two replicates per sampling area. The yield for each experimental unit was calculated as the arithmetic mean of the 10 sampled points.

As additional information, the type of implement, working width, product applied and application rate were recorded daily for each mechanised operation carried out in the experimental plots.

2.3 Data analysis

2.3.1 Meteorological data

The main meteorological parameters (temperature and precipitation) were monitored throughout the 4-season study period. Data were obtained from the Agricultural Meteorological Stations Network (RIA, Red de Información Agroclimática) of the Andalusian Regional Ministry of Agriculture (CAPADR, 2025).

2.3.2 Energy

Daily data from remote monitoring of mechanised operations in each experimental field were analysed using ad-hoc software developed in the Basic language ‘Reporter Life’. By manually entering the implement width and the dimensions of the worked experimental field, this software automatically calculated:

• Average speed
• Average fuel consumption
• Theoretical working capacity
• Actual working capacity
• Overlap

The methodology used for the energy analysis is based on that proposed by the International Federation of Institutes for Advanced Studies (IFIAS, 1978). It associates the amount of non-renewable energy with each factor in a process (Hernanz (2005)) and defines two types of energy: (i) direct-use energy, related to fuel consumption, and (ii) indirect-use energy, related to production factors. Direct-use energy was calculated by multiplying fuel consumption by 36.8 MJ l^-1. This value is based on the lower heating value of fuel (43.3 MJ kg^-1) and its density (δ = 0.85 kg l^-1). Indirect-use energy includes energy associated with production factors such as the manufacture and maintenance of machinery, as well as the energy used to obtain seeds, fertilisers, and phytosanitary products. The approach “from cradle to gate” was used for indirect-use energy inputs.

This energy analysis considers two main indicators: Energy Efficiency (EE), defined as the ratio between the energy contained in the final product (grain) and the energy required to cultivate it. For example, how many MJ ha^-¹ of energy are obtained in the harvested seed for each MJ ha^-¹ of energy applied to the crop. Energy Productivity (EP), defined as the amount of product produced (kg ha^-1) per unit of energy supplied (MJ ha^-1). That is, how many kilograms of food are produced for each MJ ha^-¹ of energy used for crop production. Energy produced per hectare (i.e., energy contained in grain production) was calculated by multiplying the energy content of harvested grain (MJ kg^-1), obtained from Villalobos and Fereres (2016) by the crop yield (kg ha^-1).

The indirect energy associated with each production factor was calculated using previous work:

• Manufacture and maintenance of machinery: Dalgaard et al. (2001)
• Seeds:Heichel (1980); Hernanz (2005), defined as Equation 1

(1) Es(MJ ha−1)=D(kg ha−1)×I (MJ kg−1)

• E_s: Indirect energy used in seeds
• D: Sowing rate
• I: Energy content of seed per unit mass

•Fertilisers: Leach (1976); Pimentel (1980) ; Mudahar and Hignett (1987); Equation 2.

(2) Ef(MJ ha−1)=EAI(MJ kg−1 or MJ l−1)×AI (%)×D(l ha−1or kg ha−1)

• E_f: Indirect energy used in fertilisers
• E_AI: Energy of active ingredient manufacture or nutrient concentration
• AI: Percentage of active ingredient or nutrient concentration contained in the product
• D: Applied dose rate

•Plant protection products:Pimentel (1980); Fluck and Baird (1982); Green (1987); Zentner et al. (2004); Hernanz (2005); Audsley et al. (2009), Equation 3.

(3) Epp(MJ ha−1)=EAI(MJ kg−1 or MJ l−1)×AI (%)×D(l ha−1or kg ha−1)

• E_pp: Indirect energy used in plant protection products (Pesticides, fungicides and herbicides)
• E_AI: Energy of manufacture of the active ingredient
• AI: Percentage of active ingredient contained in the product
• D: Rate of product applied

2.3.3 CO₂ Emissions

Emissions were measured only for crop operations, without considering soil carbon sequestration or soil respiration. Since the purpose of this study was exclusively to evaluate how NT and GA management practices affected these gas emissions specifically in relation to crop operations. Emissions of CO₂eq were calculated for direct-use energy using fuel consumption and the diesel conversion factor (0.074 kgCO₂eq MJ^-1), derived from the emission factor provided by the Ministry for Ecological Transition MITECO (2025), 2.719 kgCO₂eq l^-1, and the lower heating value of diesel (36.8 MJ l^-1).

Emissions associated with production factors were calculated from the estimated energy (Manufacture and maintenance of machinery, seeds, fertilisers and plant protection products), using energy factors obtained from literature. Emission factors were obtained from the average of Industrial Energetic Mix for the years 2009-2012 (MITECO, 2025).

Carbon balance was expressed on an area basis (kgCO₂eq ha^-1). Additionally, carbon use efficiency (CAUE) was determined as the C content of grain production per C emissions associated with the energy consumed (kgC kgC^-¹). That is, the kg of C obtained in the harvested seed for each kg of C emitted during crop production. For this efficiency only C content in grain was considered; carbon in the biomass was not taken into account. The methodological approach of ASABE (2011) was used, which does not consider the carbon content in residue biomass. Carbon concentration in grain was estimated from Ma et al. (2018). Similarly to EP, an indicator of carbon use productivity (CUP) was calculated as the amount of production (kg ha^-1) per CO₂eq emitted (kgCO₂eq ha^-1). This indicator can be summarized as the number of kilograms of food produced per kilogram of CO₂ equivalent emitted during crop production.

2.3.4 Statistical analysis

Statistical analyses were carried out using the statistical programme Statistix 9.0. A split-split-plot design was considered with the following dependent variables: energy consumed and produced, EE and EP, CAUE and CUP. The experimental field was taken as the main factor, crop as the secondary factor (subplot) and the Soil Management System (SMS) as sub-subplot. The year (season) was taken as a replicate. Analysis of variance was conducted for the variables to determine statistically significant differences; Tukey’s post-hoc comparison of means test was used at p ≤ 0.05.

A Pearson correlation matrix was used to determine the main statistical relationships among variables.

3 Results and discussion

The results and their discussion are presented in 4 sections. The first covers the harvest obtained under each soil management system: CT on the one hand and NT&GA on the other. The second section examines fuel consumption for each crop and system. The third section addresses energy in terms of expenditure, production and the most representative parameters. Finally, the last section estimates the carbon balance (CO₂eq) of field operations and production factors based on energy consumption and includes several indicators.

The rationale for integrating these sections in the analysis is to compare results holistically. Erroneous conclusions may be reached if only energy, carbon balance, or fuel consumption are considered. Prioritizing reduced energy-based management could mean overlooking the fundamental objective of agriculture: to produce sufficient food and fibre and to achieve the productive potential of the area.

3.1 Harvest

The first factor to consider when carrying out an energy balance is crop yield, as it is the source of energy production (Bechine and Castoldi, 2009). However, yield is strongly influenced by the meteorological conditions of each crop year (Lopez-Bellido et al., 2007), which, in our case, showed considerable variability over the four years of study (Figure 3). The first two years were exceptionally wet, with precipitation close to 1000 mm, almost twice the annual average for the area (CAPADR, 2025). The third was very dry, with total rainfall below 300 mm, and the fourth year was again exceptionally wet, which negatively affected the production of the crops planted in the experimental fields. These extreme differences have prevented the production potential from being achieved in all cases.

As mentioned above, the plots followed the typical rotation of the Andalusian countryside: cereal - sunflower (Helianthus annuus) - legumes. The different cereals and legumes sown in the experimental fields during the four seasons are described in Table 2.

Table 2. Cereals and leguminous plants cultivated in the experimental fields. Seasons 2009/10 to 2012/13.

Bread wheat, (Triticum aestivum); Durum wheat, (Triticum durum); Broad bean, (Vicia faba); Chickpea, (Cicer arietinum); Pea, (Pisum sativum).

The high meteorological variability meant that average crop yields over the four seasons were below the regional average (MAPA, 2024). Yields across the three crops were reduced around 10% compared with the average yield in Andalusia over the last 10 years: 2.65 Mg ha^-1 for wheat, 1.24 Mg ha^-1 for sunflower and 1.15 Mg ha^-1 for legumes. These climatic variations, mainly due to water scarcity, resulted in some crops not reaching harvest in certain growing seasons, as shown in Table 3. It should be noted that the resulting loss of some data did not affect the statistical analyses performed, as demonstrated in the Supplementary Information.

Table 3. Average annual yields for each crop, management system, and farm. Average yields for the four seasons for each farm and overall mean production. Study seasons 2009/10 to 2012/13.

SMS, Soil management system; NT, No tillage; GA, Guidance assistance; CT, Conventional tillage.

As shown in Table 3, yields varied substantially by crop, farm, and year. Wheat was the highest-yielding crop, with a four-year average across the three experimental fields of 2.39 Mg ha^-1 under NT&GA and 2.24 Mg ha^-1 under CT. Sunflower and legumes had similar yields: 1.08 Mg ha^-1 for sunflower under NT&GA and 1.10 Mg ha^-1 under CT; and 1.06 Mg ha^-1 for legumes under NT&GA and 1.08 Mg ha^-1 under CT. Thus, mean differences of 0.153 Mg ha^-1 in favour of NT&GA for wheat, 0.016 Mg ha^-1 for sunflower and 0.020 Mg ha^-1 for legumes (both in favour of CT), were recorded. In all cases no statistically significant differences were found. Therefore, under rainfed conditions in a Mediterranean climate, yields under NT did not differ on average from those under CT, which is consistent with other studies under similar conditions (Hernanz et al., 1995; Lopez-Bellido et al., 2007; Ordóñez-Fernández et al., 2007).

The mechanized operations performed varied according to local farming practices, crop type, and weather conditions, as well as the criteria of the farmers. Therefore, results obtained accurately reflect real farming conditions and are representative of those expected on commercial farms. NT systems should be adapted to local conditions to ensure successful soil management. In general, the ability of CA to increase SOC contributes to improving soil properties and crop yields (Page et al., 2020).

3.2 Fuel consumption

The highest fuel consumption was recorded for primary tillage (mouldboard: 22.5 l ha^-1, and chisel: 14.1 l ha^-1). The increase in fuel consumption of NT seeders compared with conventional seeders was notable, due to the greater weight of the NT seeder.

Operations using spraying and fertiliser distribution equipment showed greater overlap under CT, as GA was not used. Under CT, the highest overlap was recorded for the disc harrow (35.1%), which was the most frequently used tillage operations alongside the cultivator. The most common field operation under NT was the treatment sprayer because, in this soil management system, tillage is replaced by herbicides to control weeds (Kassam et al., 2009). However, the difference in the number of sprayer operations between NT&GA and CT was 29. This is much lower than the total number of field operations carried out under CT, (mouldboard plough, chisel, disk harrow, semi-chisel, scarifier, cultivator, baler, roller), which totalled 136 operations that are not conducted under NT systems (Table 4).

Table 4. Most common operations carried out in the experimental fields, their fuel consumption, overlap and the number of times they were performed over the four-season study period (2009/10 - 2012/13).

* Agricultural operations that do not have overlapping of two or more working width passes. NT, No tillage; GA, Guidance assistance; CT, Conventional tillage.

It should be noted that the baler and the roller have no overlap, as with the harvester, since the pass track can be easily identified visually.

The number of tillage operations, together with the relatively small difference in the number of sprayers passes, led to a large increase in fuel use under CT. The crop with the greatest reduction in diesel consumption due to the application of CA was legumes, 55.7%, followed by sunflower, 52.9% and wheat 51.6% (Figure 4). The analysis of variance showed statistically significant differences in fuel consumption, with lower values under NT&GA in all cases. Overall, across the four years of the study and the three crops, average fuel consumption was reduced by 53% with the application of CA and PA systems. These values are consistent with those reported in other studies in Spain on similar crops (Hernanz et al., 1995; Moreno et al., 2011; Arrúe et al., 2013).

3.3 Consumptions and energy productions

For the crops studied, higher fuel consumption and greater overlaps under CT meant that, in all cases, direct-use energy expenditure was significantly higher (Table 5). With respect to indirect-use energy associated with production factors, NT&GA required around 50% less energy for machinery maintenance. This pattern was consistent with that observed for direct-use energy, with statistically significant differences for all crops (Table 5). For seed and fertiliser, there was a trend towards lower expenditure under NT&GA than under CT due to reduced overlap, although these production factors did not show statistically significant differences. However, energy expenditure on plant protection products increased under NT&GA, since weed control in NT systems is achieved through herbicide application. For this production factor, energy consumption was significantly higher under NT&GA in wheat and sunflower. The contribution of plant protection products was very small in wheat, accounting for 3.9% of total energy consumption under NT&GA and 1.3% under CT. For sunflower and legumes, this contribution increased, particularly under NT&GA, where expenditure represented 31.8% and 20.4% of total energy consumption, respectively. These values are close to those reported by Moreno et al. (2011) under similar conditions and crops, and lower than those measured by Zentner et al. (2004) in the Canadian prairies.

Table 5. Average energy consumed, produced, efficiency and energy productivity for the different crops and soil management systems. Seasons 2009/10 to 2012/13.

SMS, Soil management system; NT, No tillage; GA, Guidance assistance; CT, Conventional tillage; EE, Energy fficiency (GJ GJ^-1); EP, Energy productivity (kg GJ^-1).

Across all crops, NT&GA reduced total energy consumption, with statistically significant differences in every case. Wheat was the most energy-consuming crop (17.54 GJ ha^-1 under NT&GA and 19.92 GJ ha^-1 under CT), due to high nitrogen fertiliser use and the lower fuel consumption. For this crop, CA systems accounted for 45.7% of the total reductions achieved, mainly due to lower fuel consumption. Likewise, PA accounted for 54.3% of the reduction in energy consumption by reducing overlaps, particularly during fertilizer application. These values were similar to those obtained by Khaledian et al (2010) in south-east France and by Paris et al. (2022) at the European level for CT, while they were lower than those reported by Moreno et al. (2011) and Hernanz et al. (1995) in Spain, due to lower fertiliser use in these studies. In sunflower and legumes, fertiliser consumption was lower than in wheat, and energy expenditure was therefore reduced. Sunflower required the least energy: 2.75 GJ ha^-1 under NT&GA and 3.73 GJ ha^-1 under CT. These values are slightly higher than those reported by Moreno et al. (2011) under similar conditions and approximately five times lower than those obtained by Nassi o Di Nasso et al. (2011) in Italy, due to the much higher fertiliser use in the latter study. They were also lower than the European-level estimate reported by Paris et al. (2022) for sunflower under CT (17.54 GJ ha^-¹). For legumes, mean consumption was 5.15 GJ ha^-1 under NT&GA and 6.31 GJ ha^-1 under CT. These were similar values to those reported by Hernanz et al. (1995) for the same crops under comparable conditions. For these two latter crops, as fertilizer use was very limited, CA accounted for more than 90% of the total reduction in energy consumption.

Overall, the mean reduction under NT&GA was 12.0% for wheat, 26.3% for sunflower and 18.4% for legumes. These reductions were slightly higher than those reported by Moreno et al. (2011) for wheat and legumes (8.6% and 15.0% respectively) and very similar for sunflower (26.9%). Although wheat had the smallest percentage reduction, it had the greatest absolute energy saving (2.38 GJ). Of this saving, 35.8% was due to the elimination of tillage, while the remaining 64.2% was mainly attributable to fertiliser savings resulting from reduced overlap under GA. The next highest saving was for legumes (1.16 GJ relative to CT). Since fertiliser use was much lower than in wheat, only 11.2% of this saving was attributable to GA, while the remaining 88.8% resulted from the elimination of tillage. Finally, sunflower saved 0.98 GJ, with GA accounting for only 5.1% of the saving and NT for the remaining 94.9%.

In addition to reducing tillage operations, GA systems across the full crop rotation reduced input use and improved equipment efficiency, with potential economic and environmental benefits (Kharel et al., 2020).

Only in term of energy of manufacturing plat protection products, the NT&GA system showed higher energy consumption (Table 5). On average, NT&GA spend 45% more energy in pesticides, fungicides and herbicides than CT. The increase mainly corresponds to herbicide use as CT controls partially the weeds mechanically and the use of pesticides and fungicides are similar in both systems. This might imply environmental issues, however, the significant reduction in runoff and erosion avoids non-point source pollution into water bodies (Clement et al., 2024; Fleming et al., 2025).

Furthermore, continuous NT soil management increases soil organic matter (Carbonell-Bojollo et al., 2015; Mazzoncini et al, 2016) thereby enhancing the rate of herbicide degradation in soil (Sharipov et al., 2021).

In terms of energy production, no statistically significant differences were found between the soil management systems, since, as noted above, there were also no differences in yields. Energy production varied among crops. Wheat not only consumed the most energy during production, but also had the highest yields and therefore produced the most energy (30.10 GJ ha^-1 under NT&GA and 28.18 GJ ha^-1 under CT). These values were higher than those reported by Moreno et al. (2011) and slightly lower than those measured by Hernanz et al. (1995), both conducted in Spain under rainfed conditions. Cereal energy production was approximately twice that of sunflower (14.98 GJ ha^-1 under NT&GA and 15.23 GJ ha^-1 under CT), substantially higher than the values reported by Moreno et al. (2011) in Spain and similar to those obtained by Nassi o Di Nasso et al. (2011) in Italy. Legumes produced about 2.5 times less energy than wheat (12.01 GJ ha^-1 under NT&GA and 12.74 GJ ha^-1 under CT). These values match those reported by Moreno et al. (2011); however, they were approximately 7 times lower than those obtained by Hernanz et al. (1995).

Although sunflower did not produce the most energy, it used energy most efficiently, with the highest EE and EP among the three crops. It was the only crop for which statistically significant differences on EE and EP were observed in favour of NT&GA. These two indicators were approximately half as large for legumes and around one-third as large for wheat. In all cases EE and EP was higher under NT&GA, consistent with comparisons of NT and CT reported by Hernanz et al. (1995); Moreno et al. (2011), and Nassi o Di Nasso et al. (2011) under rainfed Mediterranean conditions. For sunflower, EE increased by 33.5% under NT&GA and EP by 34.5%. For wheat, EE increased by 21.3% and EP by 27.3%; for legumes, the increases were 15.6% and 23.5%, respectively, relative to CT.

3.4 Carbon balance and indicators

Carbon balance (i.e., emissions) associated with the energy consumed in field operations and production factors followed a similar pattern to the energy balance. Higher emissions were observed in wheat and under CT. In all crops, NT&GA significantly reduced GHG emissions (Figure 5). On average, across all fields and seasons, reductions of 26.4% were observed for sunflower under NT&GA, 18.4% for legumes, and 12.0% for wheat. These percentage of reductions are consistent with those reported by Saldukaitė et al. (2022) for oilseed rape. In line with our findings, Guardia et al. (2016) reported substantially higher GHG emissions in cereals than in legumes, mainly due to greater fertiliser inputs. Their study also showed that NT reduced emissions associated with field operations and input use compared with CT, consistent with our results.

The carbon balance developed in this study does not consider the carbon sequestration conducted in soil when it is minimally disturbed and the crop residue is retained. The relation between CA and carbon sequestration in climate change studies has been researched deeply (Román-Vázquez et al., 2025). CA systems are widely demonstrated the soil carbon increase when the three core principles are jointly carried out (González-Sánchez et al., 2012; Aguilera et al., 2013; de Oliveira Ferreira et al., 2024). Overall, the carbon balance is more favourable to CA systems against CT when carbon sequestration is considered (Guardia et al., 2016). For that reason, this study is centred exclusively to the carbon footprint associated to the energy consumption direct and indirectly.

Likewise, the GHG emissions from soils (soil respiration) were not measured in this research. When the CO₂ emissions from soils are evaluated comparing CA and CT systems, several studies report lower emissions under CA (Bregaglio et al., 2022), Tillage tends to engender accelerated oxidative breakdown of organic matter with accelerated release of increased volumes of CO₂ to the atmosphere, beyond those from normal soil respiration processes (Kassam et al., 2009). The particular case of N₂O emissions is more dependent on fertilisation strategy, although NT system usually reduces the amount of N₂O emitted (Corrochano-Monsalve et al., 2020; Carbonell-Bojollo et al., 2022).

Emissions-related indicators showed significant differences for sunflower; for wheat and legumes, mean CAUE and CUP were higher under NT&GA, but without statistically significant differences (Figure 6).

NT&GA increased CAUE and CUP by similar amounts within each crop: 34% for sunflower, 21% for wheat, and 20% for legumes. Therefore, CA systems combined with PA can support low-carbon agriculture while maintaining similar production.

The negative effects of climate change are expected to be particularly severe in Mediterranean area such as Southern Spain (Kovats et al., 2014). CA and PA are especially relevant to adapt to future climate scenarios as, acting jointly, they are more energy and carbon efficient (González-Sánchez et al., 2017; Awais et al., 2026).

Table 6 shows that the crop grown had a strong influence on all variables studied: energy consumed and produced, EE and EP, as well as CAUE and CUP. The experimental field did not influence energy produced and consumed and the other variables studied. Finally, the soil management system was a key factor affecting energy consumed in crop production, with no differences in the energy produced, since yields did not vary significantly between management systems. This factor also influenced EE, EP, CAUE and CUP, as values were higher under NT&GA. In addition, the FieldCrop interaction showed statistically significant differences only for energy consumption, as did the FieldSMS interaction, although the latter showed lower p-values and, therefore, a higher level of statistical significance. Other interactions (FieldSMS, Year and FieldCrop*SMS) were not significant in any case.

Table 6. Statistical analysis of the relationships between the different study variables.

E, Energy; EE, Energy efficiency; EP, Energy productivity; CAUE, carbon use efficiency; CUP, carbon use productivity; SMS, Soil management system.

p value: *<0.05; <0.01; *<0.001; † not significant. Seasons 2009/10 to 2012/13.

It should be noted that energy-related and carbon-emissions-related parameters are closely aligned, as carbon emissions were calculated from energy consumption.

4 Conclusions

The remote monitoring system developed proved to be an effective tool for accurately characterizing field operations and assessing management practices. Under rainfed conditions in southern Spain, the adoption of NT combined with GA maintained crop yields and energy output, confirming its agronomic and productive viability.

Furthermore, the simple methodological approach has proven to be an efficient and straightforward method for performing energy analyses, providing reliable results based on real data that can be replicated. When choosing energy-saving and efficiency strategies, good practices must be differentiated and tailored to the target crop. In our study, energy savings in wheat cultivation were mainly due to lower fertiliser consumption, which benefited from the use of guidance assistance. In sunflower and legumes, the greatest savings came from the suppression of tillage under the NT system, leading to a reduction in fuel use of more than 50%.

In energy-balance studies, energy efficiency and, above all, energy productivity must be considered. If the analysis focuses only on energy consumption, this can lead to incomplete conclusions. In contexts where energy is accessible, farmers can be expected to use what they need; therefore, when developing technical training programmes, it is important to understand where energy is best invested in relation to production. In other words, the EP of each crop under each soil management system should be evaluated. Likewise, the carbon footprint of the system should be expressed not only per unit area, but also in terms of CUP, taking into account production per kg CO₂ emitted.

Wheat consumed and produced the most energy and had the highest carbon footprint. However, sunflower was the most efficient crop in terms of energy and carbon use, with EE and EP values twice those of legumes and three times those of wheat. In terms of carbon balance, NT&GA reduced emissions by 12-26%. Similarly, CAUE and CUP were 15-35% higher under NT&GA than under tillage-based systems without guidance assistance.

Energy consumption and GHG emission were not only influenced by crop and management system, but also by the experimental field, since operations varied by field, and by study year, due to the fact that weather variability required farmers to adapt cultivation practices to prevailing conditions.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fagro.2026.1916869/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Ivan Vasenev, Moscow Timiryazev Agricultural Academy, Russia

Reviewed by: Ernesto Viglizzo, Independent Researcher, Santa Rosa, Argentina

Diego Pierotti Procópio, Federal University of Mato Grosso, Brazil

Gergana Kuncheva, Agrotechnologies and Plant Protection “Nikola Pushkarov”, Bulgaria