Introduction

In 2020, global agrifood systems emissions were 16 billion tons of carbon dioxide equivalents (CO2e), approximately one-third of total anthropogenic greenhouse gas (GHG) emissions, with the farm gate stage representing nearly half of total agrifood systems emissions (FAO 2022). Although agriculture is a significant contributor to climate change, it is also one of the economic sectors most at risk from it, as climate change is already affecting food security through increasing temperatures, changing precipitation patterns, and greater frequency of extreme events (Mbow et al. 2019; Tubiello et al. 2021).

Water scarcity is one of the most important environmental issues facing the wine industry, and climate change may have a substantial impact on temperatures and precipitation during the growing season, resulting in increasingly severe water deficits that affect fruit yield and composition (Fraga et al. 2012; Gao and Giorgi 2008; IPCC 2018).

In several Mediterranean wine-growing regions, annual precipitation levels are below those required for economically viable grapevine cultivation (Medrano et al. 2015). While most vineyards in Europe are currently rainfed, irrigation is an option for growing sustainable yields in more arid conditions. However, the associated effects on water resources use and GHG emissions in the environment must be considered (Daccache et al. 2014; Silva and Silva 2022; van Leeuwen et al. 2019).

Portugal is one of the main grapevine producers in Europe, with 953 thousand tons produced over an area of about 171 thousand ha in 2021; the southern Portuguese region of Alentejo contributes to 28% of production over an area of around 20% of the national territory. The introduction of irrigation in most vineyards in Alentejo has contributed to an increase in regional and national production (INE 2022). Although irrigation contributes to higher and steady crop yields, its effects on the environment, specifically those associated with water resources depletion and GHG emissions, are still unclear (Balafoutis et al. 2017; Zhang et al. 2018). Evaluating how resources are used in irrigated viticulture may help to outline sustainable management strategies to adjust to the new conditions and reduce the environmental impacts (Koushki et al. 2023; Raza et al. 2019).

A “family” of footprint indicators for the evaluation of the environmental performance of different production systems has been developed over recent decades for measuring and monitoring sustainability (Galli et al. 2012). The water footprint (WF) and the carbon footprint (CF) are among the most known environmental footprint indicators (Čuček et al. 2015). The WF is a tool to calculate the water virtually embedded in commodities, representative of the volume of water needed to produce goods and services (Hoekstra 2003; Hoekstra et al. 2011). The WF of a crop can be a quantifiable indicator for measuring the water applied by irrigation, the water stored in the soil, fractions consumed by the crop, and the potentially contaminated water as a result of the adopted agronomical practices (Mekonnen and Hoekstra 2011). The carbon footprint (CF) is a measure of the amount of GHG resulting from a particular activity, product, or service (Wiedmann and Minx 2008). In the context of crop production, CF corresponds to the total amount of GHG per unit yield that results from various activities involved in producing a crop, such as land preparation, planting, harvesting, transportation, and processing (PAS 2050a, b, 2012). Other than mobile farm operations that require fuel use, the main components of energy use in irrigated agriculture are primarily related to processes required to apply water to crops in the field by lifting, conveying, and/or pressurizing it (Rothausen and Conway 2011). Despite that the water and energy consumption associated with irrigation should increase the water incorporated in the crop products and the GHG emissions per unit area resulting from farming, irrigation normally enhances yields, which is a crucial consideration in the calculation of WF and CF (Zhang et al. 2018). Additionally, the adoption of low-input agronomic options, like reduced or water-saving irrigation strategies, conservation tillage, and/or soil and residues management that increase soil carbon sequestration, have the potential to retain water and mitigate GHG emissions (Gan et al. 2014; Litskas et al. 2017; Sapkota et al. 2020). Regardless of these relationships between irrigation and energy use, WF and CF are indicators with different scope and extent: while local GHG emissions contribute to the global stock of CO2 irrespective of their origin, the evaporation of water has a more localized effect, impacting the basin where evaporation takes place (Fereres et al. 2017; Perry 2014).

Expanding WF and CF estimates from on-field measurements in different agri-environmental conditions can portray better the variability present at the farm level estimates and contribute to the decision-making process in selecting sustainable management practices (Herath et al. 2013; Lal 2004), as well as for delineating strategic policies regarding agricultural production, and more specifically, the wine-making sector in water-scarce regions (Ene et al. 2013; Lamastra et al. 2014).

Based on the research data obtained from an on-farm study in vineyards located in Southern Portugal, the aims of the current study were: (i) to estimate WF and CF; (ii) to explore the relationships between the WF and CF components; (iii) to analyze the trade-off between WF and CF footprints and yield, relating them to the agricultural inputs and management practices of irrigated grapevines under Mediterranean conditions.

Materials and methods

Study area

The study was carried out over three years (2018 to 2020) in three irrigated vineyards (V1 (37°52′22ʺN; 07°30′37ʺW), V2 (37°58′13ʺN; 07°33′18ʺW), and V3 (37°60′28ʺN; 07°32′21ʺW)) (Table 1).

Table 1 General characterization of the vineyards assessed in the study
Full size table

The vineyards were in South Portugal, specifically in the hydro-agricultural area of Brinches-Enxoé, part of the Alqueva irrigation plan, presently covering approximately 120,000 ha. The climate in the region is temperate with hot and dry summers (Mediterranean), with annual precipitation and average mean monthly temperature of, respectively, 558 mm and 16.9 °C (long-term means for the 1981–2010 period, (IPMA 2023)). During the three years of study (2018, 2019 and 2020), data from an automatic meteorological station located near the study area (Latitude: 37° 58′06ʺN; Longitude: 07° 33′03ʺW; Altitude: 190 m), showed that the annual precipitation was 603 mm, 343 mm, and 615 mm, respectively. The mean temperature was 16.7 °C, 17.3 °C, and 17.8 °C, respectively, in 2018, 2019, and 2020 (COTR 2022a).

The duration of the grapevines cycle varied between 154 days in V2-2019 and 189 days in V3-2020, with an overall average duration of 170 days (Table 2). Maximum daily temperatures were higher in 2020 (average of 29.9 °C), the warmest year, while the lowest precipitation values occurred in 2019 (average of 76 mm), the driest year. In 2018 and 2019, average seasonal precipitation was 147 mm and 207 mm, respectively.

Table 2 Duration of the vineyards cycle per year
Full size table

Management practices data were provided by the farmers and are described in Table 3 and 4. The fertilizers used were primarily formulations of nitrogen (NO3−, NH4+, urea), phosphorus (P2O5), and potassium (K2O). Water-soluble and liquid fertilizers, applied over the crops’ cycle through irrigation water, were mostly nitrogen fertilizers but also included other fertilizers containing formulations of iron (Fe chelates), calcium (CaO), or sulfur (SO3). Foliar applications, containing boron (B) and/or manganese (Mn), were also employed. The pesticides used were mainly herbicides and fungicides, and some insecticides. In vineyards V1 and V2, 100 Hp tractors were used, while in V3, the tractor used had 110 Hp power.

Table 3 Summary of crop management data per vineyard per year
Full size table
Table 4 Average workdays, type and number of mechanical operations per vineyard per year
Full size table

The vineyards were drip irrigated, with 50 cm spacing between drippers with flow rates of 2.2 L h−1 in V1 and V2, and 2.3 L h−1 in V3. The irrigation dose and schedule were established by the farmers based on recommendations of the Irrigation Management Model for the Alentejo region (MOGRA—Modelo de Gestão da Rega para o Alentejo (COTR 2022b). The MOGRA model performs a daily soil water balance, based on the FAO-56 single crop coefficient method for computing crop water requirements (Allen et al. 1998), using meteorological data and crop-specific information. The actual crop evapotranspiration (ETa) corresponds to an adjusted crop evapotranspiration value (ETcadj) obtained from the multiplication of the reference evapotranspiration (ET0, computed with the Penman–Monteith equation), by the single crop coefficients (Kc) through the cycle of grapevine for wine production, and by a stress coefficient (Ks) of approximately 0.5, compatible with a sustained deficit irrigation strategy, which is the common strategy followed by the winegrowers of the Alentejo region. The MOGRA model also provided the values of effective precipitation (Pef), and irrigation requirements (IR), while effective irrigation (Ief) data were provided by the farmers.

Water footprint calculations

Each of the WF components (expressed in m3 t−1) was calculated following the tier 1 approach of the Water Footprint Network (Hoekstra et al. 2011) and described in Tomaz et al. (2021), using the following equations:

Green Water Footprint (WFG):

$${WF}_{G}=\frac{{ET}_{G}}{Y},$$
(1)

where ETG corresponds to the green component in crop water use (m3 ha−1), obtained as the minimum between Pef and ETa (Hoekstra et al. 2011), and Y is the crop yield (ton ha−1).

Blue Water Footprint (WFB):

$${WF}_{B}=\frac{{ET}_{B}}{Y} ,$$
(2)

where ETB is the amount of irrigation water available for plants (m3 ha−1), defined as the minimum between IR and Ief (Aldaya et al. 2010). In case Pef is equal to or higher than the crop water requirements (CWR), IR equals 0, otherwise, it corresponds to the difference between CWR and Pef.

Grey Water Footprint (WFGr):

$${WF}_{Grey}=\frac{\frac{Appl \times \alpha }{{c}_{max}-{c}_{nat}}}{Y}\times 1000,$$
(3)

where Appl is the chemical application rate, in this case, the N and P application rate, α is the leaching-runoff fraction, i.e., the fraction of applied N or P that reaches the freshwater bodies, cmax is the maximum acceptable concentration of the contaminant in the aquatic environment (kg m−3), and cnat is the natural background concentration of contaminant in the aquatic environment (kg m−3).

In a large number of published studies about water footprints, nitrogen is considered the only contaminant in WFGr calculation but, according to the tier 1 approach described in Franke et al. (2013), when assessing the WFGr, the value for each contaminant of concern must be calculated separately and the overall WFGr will be equal to the largest value found (Franke et al. 2013). Notwithstanding the importance of pesticides’ impact on soil and water resources, given the lack of consistent standards concerning diffuse sources of water pollution in Portuguese agricultural systems, in this study, we considered N and P as the major contaminants of concern. We used a value of α = 10% for nitrogen fertilizers and α = 3% for phosphorus fertilizers (Franke et al. 2013). For the ambient water quality standard, cmax, we used 50 mg N-NO3 L−1 and 5 mg P2O5 L−1, the maximum allowable values, respectively, for nitrogen and phosphorus, concerning the protection of waters against pollution from agricultural sources according to Portuguese legislation (Diário da República, 1ª SÉRIE 1998). In the case of cnat, many previous studies assumed the value 0 which could lead to underestimates. We used the values 0.1 mg N-NO3 L−1 and 0.01 mg P2O5 L−1, respectively, for N and P (Franke et al. 2013). Another assumption that must be considered in WFGR calculation is that, in these vineyards, irrigation does not contribute to the leaching of nitrogen and phosphorus, since drip irrigation systems are characterized by low losses due to percolation or runoff. Furthermore, the applied irrigation volumes correspond to deficit irrigation. Thereby, the leaching of N and P should be mainly due to precipitation immediately before or throughout the grapevine’s cycle (Saraiva et al. 2020).

The total water footprint (WF) of the grapevines’ crop corresponds to the sum:

$$WF={WF}_{G}+{WF}_{B}+{WF}_{Gr}$$
(4)

Carbon footprint calculations

Irrigation powered by electricity, fertilizers and pesticides manufacturing and application, and mechanical field operations were the stages considered within the system boundary, that is, a “cradle to gate” assessment was performed (Marras et al. 2015; Nemecek et al. 2015; Novara et al. 2019; PAS 2050a, b, 2011) (Fig. 1).

Fig. 1
figure 1

System boundary and agricultural inputs considered in carbon footprint estimations

Full size image

Therefore, our scope was the total GHG emissions at the vineyard stage, and the functional unit considered was 1 ha of area in the studied vineyards. The GHG emission rates were estimated using an emission factor approach which, combined with the agricultural input, generates an emission for a given period (IPCC 2006, 2019; PAS 2050a, b, 2011):

$$Emission=Agricultural Input \times Emission Factor,$$
(5)

where Emission is the GHG emission, expressed in kg GHG ha−1, the Agricultural Input is expressed in unit ha−1, and the Emission Factor in kg unit−1. The emission factors used in the study are described in Table 5.

Table 5 GHG emission factors for the vineyard stage used in carbon footprint calculations
Full size table

The GHG emissions and CF calculations considered in the study were made according to the tier 1 approach of the IPCC, also described in Litskas et al. (2017) and Cool Farm Alliance (2023), following the IPCC guidelines (IPCC 2006, 2019):

Diesel fuel carbon emissions (CED):

$${CE}_{D}={V}_{D} \times {EF}_{D},$$
(6)

where VD is the volume of diesel consumed due to vineyard operations (data provided by the farmers; L ha−1), and EFD is the emission factor for diesel fuel (kg CO2e L−1).

Irrigation carbon emissions (CEI):

$${CE}_{I}={E}_{I} \times {EF}_{I},$$
(7)

where EI is the energy associated with irrigation (kWh ha−1) and EFI is the emission factor of electricity for irrigation (kg CO2e kWh−1). The EI was estimated using:

$${E}_{I}={D}_{I} \times {E}_{I;M},$$
(8)

where DI is the irrigation depth (mm) and EI;M is the energy requirements, dependent on the irrigation method (kWh mm−1 ha−1), for which we used the value 2.0 for drip irrigation (Cool Farm Alliance 2023).

Nitrogen fertilizer direct (NEN-Direct) and indirect (NEN-Indirect) emissions:

The NEN-Direct are the equivalent carbon emissions related to nitrous oxide gas (N2O) produced during nitrification and denitrification processes. The N2O is a gaseous intermediate in denitrification and a by-product of nitrification. One of the main controlling factors in these processes is the availability of inorganic N in the soil, derived from human-induced net N additions to soils and/or management practices that mineralize soil organic N (IPCC 2006, 2019). We used the following for the calculation of the direct fertilizer-induce emissions approach (Cool Farm Alliance 2023):

$${NE}_{N-Direct}={\frac{44}{28} \times (F}_{MN}+{F}_{ON}) \times {EF}_{F},$$
(9)

where FMN is the annual amount of mineral N fertilizer applied (kg N), FON is the annual amount of organic N fertilizer applied (kg N), EFF is the emission factor for N2O emissions from N inputs, and 44/28 is the conversion factor of N2O-N to N2O.

Emissions of N2O also take place through two indirect pathways, namely, (i) the volatilization of N as NH3 and oxides of N (NOx), and the deposition of these gases and their products NH4+ and NO3−, and (ii) the leaching and runoff from the land of N from synthetic and organic fertilizer additions, which can be negligible in dry climates (IPCC 2006, 2019). Thereby:

$${NE}_{N-Indirect}={\frac{44}{28} \times (F}_{MN} \times a+{F}_{ON} \times b) \times {EF}_{F},$$
(10)

where a and b are, respectively, the fraction of mineral and organic N fertilizer that volatilize as NH3 and NOx. The default values for a and b are 10% and 20%, respectively (IPCC 2006, 2019).

In the absence of differentiated EF for all types of fertilizers used in the studied vineyards, we used the default values for dry cropland in the IPCC tier 1 approach, both in Eq. 9 and Eq. 10. For the calculation of total nitrogen fertilizer carbon equivalent emissions (CEN), it should be noted that each GHG makes a different contribution to global warming. According to the IPCC 5th Assessment Report (IPCC 2014), N2O has a global warming potential (GWP) of 265, that is, compared to CO2, it is 265 times higher in terms of the 100-year global warming potential, therefore:

$${{CE}_{N}=(NE}_{N-Direct}+ {NE}_{N-Indirect}) \times 265$$
(11)

Urea fertilizer carbon emissions (CEU):

Adding urea (CO(NH2)2) to soils leads to a loss of CO2 that was fixed in the industrial production process. To estimate the amount of CO2 emission that results from the addition of this type of fertilizer, Eq. (11) was used:

$${CE}_{U}=\frac{44}{12}{ \times F}_{U} \times {EF}_{U},$$
(12)

where FU is the annual amount of Urea fertilizer applied (kg Urea), EFU is the emission factor for CO2 emissions from Urea inputs, and 44/12 is the conversion factor of CO2-C to CO2.

Fertilizer production carbon emissions (CEPF):

The fertilizer production and transport related carbon emissions were calculated separately for N, P2O5, and K2O using:

$${CE}_{PF}=F \times {EF}_{PF},$$
(13)

where F is the amount of fertilizer applied (kg N, kg P2O5, or kg K2O) and EFPF is the default emission factor for the European region of N, P2O5, or K2O fertilizer production (includes raw material extraction, energy supply, the manufacturing process to the product storage at the production site) (Cool Farm Alliance 2023).

Pesticides production carbon emissions (CEPP):

The pesticide production related carbon emissions were calculated separately for herbicides, fungicides, and insecticides using:

$${CE}_{PF}=P \times {EF}_{PP},$$
(14)

where P is the amount of pesticide applied (kg herbicide, kg fungicide, or kg insecticide) and EFPP is the correspondent emission factor of herbicides, fungicides, or insecticides production and post-production (West and Marland 2002).

Carbon footprint (CF):

The GWP per unit grain production is the CF, expressed in kg CO2e ton−1. The CF components, \({{\text{CF}}}_{{\text{i}}}\), were obtained by the ratio between the input-related carbon equivalent emissions, \({{\text{CE}}}_{{\text{i}}}\) and yield (ton ha−1):

$${CF}_{i}=\frac{{CE}_{i}}{Y},$$
(15)

Where i corresponds to the type of emission from Eq. 6 to Eq. 14, and CEi = CED, CEI, CEN, CEU, CEPF, CEPP. The total CE, \(\sum {CE}_{i}\), represents the total GWP per unit area, within the system boundary on each year, that is, the absolute value for GHG emission.

Statistical analyses

Data were analyzed using Statistica 7 (StatSoft, Inc. 2004). Factor Analyses (FA) were conducted to examine latent (unobserved) common characteristics of the water and carbon footprints and explore relationships between footprint components. The factors were extracted using the Principal Components Analysis (PCA) method and the matrix of factor loadings was submitted to varimax rotation to yield a factor structure simpler to interpret (Jagadamma et al. 2008; Tomaz et al. 2021). Factors were retained when presenting eigenvalues > 1 and a contribution for the proportion of variance > 10%. Footprint components with large absolute value factor loadings are more likely to represent a common factor, thereby, they were considered highly correlated whenever factor loadings were > 0.60 and moderately correlated when loadings were between 0.40 and 0.60 (Jagadamma et al. 2008; Lee et al. 2005; Tomaz et al. 2021).

The total WF and CF were plotted on yield, separately, on a Cartesian plane to distinguish the effects of vineyard management practices and to reveal de relationships between the estimated footprints and yield, following the methodology described by Zhang et al. (2018). These plots facilitate visualization of the relationships between WF and CF and above-average yield (Win–Win and Lose-Win) and below-average yield (Win-Lose and Lose-Lose), identifying groups of vineyards/year in each quadrant (Fig. 2). The inputs and outputs for each group of vineyards/year were summarized, through the calculation of the mean and standard error (SE) values.

Fig. 2
figure 2

Plot to visualize the relationships between environmental footprints and yield (methodology adopted from Zhang et al. (2018))

Full size image

Results and discussion

Water and carbon footprints

The total WF mean value for the three vineyards, over the study period 2018–2020, was 223 m3 ton−1, varying from 491 m3 ton−1 in 2018 to 773 m3 ton−1 in 2019. (Fig. 3). Therefore, values presented considerable variability that can be explained with different meteorological conditions over the three years. The WFB was, on every year and vineyard, the most important component, with the higher proportion occurring in 2019 (75%), which is explained by the dry climatic conditions of the year, leading to higher crop water requirements. The WFB/WFG ratio was consistently > 1, varying from 1.2 (V2-2020) to 4.4 (V2-2019).

Fig. 3
figure 3

WF components in the studied vineyards from 2018 to 2020. WFG green water footprint, WFB blue water footprint, WFGr grey water footprint

Full size image

More than the WFB absolute values, these ratios show the dependence on irrigation for the crops to attain sustainable yields in the dry Mediterranean conditions of South Portugal. Another important observation is that the WF values found were lower than the benchmark of 609 m3 t−1 for grapevines worldwide (Mekonnen and Hoekstra 2011) or generally lower than 300 m3 t−1, found by Steenwerth et al. (2015) for irrigated grapevines in California.

The CF components estimated are presented in Fig. 4. The average total CF was 98 kg CO2e ton−1. Average values each year were 86 kg CO2e ton−1 in 2018, 111 kg CO2e ton−1 in 2019, and 97 kg CO2e ton−1 in 2020. The major components were CFD (35%-47%), followed by CFFe (18%-26%) and CFI (14%-15%).

Fig. 4
figure 4

CF components in the studied vineyards from 2018 to 2020. CFD diesel carbon footprint, CFI irrigation carbon footprint, CFFe fertilizer emissions carbon footprint, CFNF nitrogen fertilizer carbon footprint, CFPF phosphorus fertilizer carbon footprint, CFKF potassium fertilizer carbon footprint, CFHb herbicide carbon footprint, CFFg fungicide carbon footprint, CFIs insecticide carbon footprint

Full size image

Average CFD increased in 2020 when favorable conditions for the development of pests and diseases in the Spring led to more insecticide and fungicide spraying operations. The use of nitrogen fertilizer is an important source of CO2 and nitrous oxide gas (N2O) emissions. To reduce direct and indirect N2O emissions, it is important to improve N use efficiency, by minimizing losses caused by erosion, leaching and/or volatilization and identifying alternate sources including nitrogen fixation and carbon sequestration by cover crops in the row, alternative fertilizer sources, and recycling nutrients contained in crop residue (Barão et al. 2019; Cataldo et al. 2020; Freibauer et al. 2004; Lal 2004; Novara et al. 2019, 2020; Pacheco et al. 2023).

As expected, a higher proportion of CFI occurred in 2019, the driest year, when an average emission of 16 kg CO2e ton−1 was due to irrigation. The absolute maximum value of CFI was observed in V2 in 2019 (28 kg CO2e ton−1), but the highest proportion of CFI to total CF occurred in V3 (on average, 20%).

Overall, the estimated CF is lower than the findings in previous studies with irrigated grapevines: 87–584 kg CO2e ton−1 (Steenwerth et al. 2015); 280–850 kg CO2e ton−1 (Litskas et al. 2017); 317–346 kg CO2e ton−1 (Hefler and Kissinger 2023).

The relationships observed among footprint components were explored through factor analysis and translated into a three-factor model which explained 91.7% of the total variance (Table 6). Factor 1 explained 63.5% of the variance and presented high positive loadings (> 0.60) of WFB (0.97), CFD (0.83), CFI (0.90), CFHb (0.69), and CFFg (0.85). Thereby, given the high weights (> 0.90) of the blue components of footprints, WFB, and CFI, this factor can be described as an “Irrigation latent variable”. Factor 2 was responsible for explaining 16.9% of the total variance and it was mainly influenced by WFG (0.94), CFIs (0.87), and CFNf (0.47). Lastly, Factor 3 accounted for 11.4% of the total variance and presented high correlations with WFGr (0.84), CFFe (0.80), CFNf (0.72), CFPf (0.84), and CFKf (0.96), therefore, it can be denoted as a “Grey latent variable”. The structure of the three-factor model showed how the WF and CF components indicators correlate and interact (Galli et al. 2012). Safeguarding differences in scope, this allows for a joint interpretation and the exploratory analysis of multivariate relationships in further agricultural plot- or farm-level studies.

Table 6 Factor loadings, eigenvalues, and percentage of total variance in a three-factor model for 12 variables (footprints components)
Full size table

Trade-off and win–win relationships

The analysis of the trade-off and win–win relationships between yield and WF showed that vineyards/years were grouped in the Win–Win quadrant mostly due to a “year effect” (Fig. 5a). In fact, all vineyards in 2018 were present in this group, indicating how the higher annual and Spring rain and ensuing lower irrigation requirements during the 2018 season, led to decreased total WF, WFB, and higher yields (Fig. 5b and Table 7).

Fig. 5
figure 5

Relationships between WF and yield (a) and between WF components and groups of vineyards (b). The vertical and horizontal lines in (a) indicate, respectively, the mean yield and mean WF for all vineyards. Win–Win (green diamonds)—WF win-yield win; Win-Lose (blue triangles)—WF win-yield lose; Lose-Lose (red circles)—WF lose-yield lose

Full size image
Table 7 Relevant inputs of WF and yield for four groups of vineyards/year (means ± SE)
Full size table

A trade-off relationship (Win-Lose) was found for V3, in 2019, mostly because of low applied irrigation and low levels of phosphorous fertilization, influencing blue and grey components of WF. In the Lose-Lose quadrant, we found again, a “year effect” since all vineyards in 2020, plus V2 in 2019 were grouped. In this group, neither irrigation volumes nor fertilization doses were the highest, so it was primarily the reduced yields that led to this result.

While the relationships between yield and WF were mostly “year-controlled”, the same pattern did not apply to yield and CF relations. The Win–Win quadrant includes two vineyards (V1, V3) in 2018 but also V1 in 2019 (Fig. 6a). This group was characterized by lower CF components resulting from fuel consumption, N and P2O5 fertilization, as well as reduced fertilizer and pesticide use (Fig. 6b). In general, and as expected, the average values of inputs were lower than the ones found for the vineyards in the Lose-Lose quadrant (Table 8).

Fig. 6
figure 6

Relationships between CF and yield (a) and between CF components and groups of vineyards (b). The vertical and horizontal lines in (a) indicate, respectively, the mean yield and mean CF for all vineyards. Win–Win (green diamonds)—CF win-yield win; Win-Lose (blue triangles)—CF win-yield lose; Lose-Win (yellow squares)—CF lose-yield win; Lose-Lose (red circles)—CF lose-yield lose

Full size image
Table 8 Relevant inputs of CF and yield for four groups of vineyards/year (means ± SE)
Full size table

Notwithstanding these observations, it is clear that the higher average yield (18.3 ton ha−1) played an important role in grouping the vineyards in the Win–Win quadrant. The V3 vineyard in 2019 and 2020 was located in the trade-off quadrant of lower CF versus lower yield (Win-Lose), a result that was mainly related to very low energy inputs (fuel and irrigation), as well as very reduced use of fertilizers and pesticides. The trade-off of high CF versus high yield (Lose-Win), found for V2 in 2018, relates to moderate yield (16.5 ton ha−1) coupled with moderate use of fuel (203.3 L) and energy for irrigation (520 kWh), but higher fertilization rates.

Together with the latent variables found in factor analysis, relating different components of WF and CF, it is worth noting the overlapping of groups of vineyards/year from Figs. 5a and 6a, namely: V1-18, V1-19, and V3-18, in the Win–Win quadrant; V3-19, in the Win-Lose quadrant; V1-20 and V2-20 in the Lose-Lose quadrant. Thereby, we can consider that these correspondences translate (i) the relationships between the two footprint indicators, (ii) the interconnection of agricultural inputs versus yields in affecting the environmental impacts and production results, and (iii) the farming practices most influential on these environmental indicators in irrigated grapevines, which we identified as being:

  • Irrigation volumes, and energy for pressurized irrigation systems, dependent on crop water requirements, given the annual climate conditions;

  • Fuel consumption, dependent mostly, on the number of pesticides spraying operations and, indirectly, on favorable climate conditions for the development of pests and diseases;

  • Nitrogen and phosphorus fertilization rates, depending on the productive potential of the vineyards, which is also, indirectly related to climatic conditions.

In summary, the estimates of WF and CF cannot be decoupled from grapevine yield and the role that climate conditions and adjusted agronomic options play to meet potential productivity. At the farm level, the capture of variability in the footprints indicators and components, along with the possibility of correlations with features of the cropping systems, of the adjusted technical options and of the local climatic data, makes this assessment more meaningful with a degree of detail required for an accurate assessment, in line with the conclusions by Herath et al. (2013) in their study of the water footprint of New Zealand’s wines.

The carbon stock changes that could arise from the type of biomass residue management (pruning wood cutting and incorporation) and soil management practices (cover crops in the interrow) adopted by the farmers in this study, were not estimated due to lack of data. Mediterranean-type perennial crops such as grapevine exhibit biological, structural, and management features, like the incorporation of pruning debris, and the practice of cover cropping in the interrow, which have the potential to sequester important amounts of CO2 (Vendrame et al. 2019). Examples of studies that report on these estimations are Litskas et al. (2017), Marras et al. (2015), Tezza et al. (2019) or Novara et al. (2020). Although the IPCC Guidelines for national greenhouse gas inventories (IPCC 2006; 2019) include the account of changes in carbon stocks that can result from alterations in land use, management practices (tillage), and biomass (inputs), the increase in soil carbon stocks with the adoption of various improved management practices may not be properly credited (Sanderman and Baldock 2010). Nevertheless, these potential positive effects on carbon budgets should be considered in future research projects, designed to address the problem of insufficient data and inconsistent methodological approaches.

Other inconsistencies regarding water and carbon footprints estimations have been addressed in several studies, e.g.: the variability in cumulative seasonal ET estimations influence on reliable WF assessments (Laan et al. 2019); shortcomings associated with grey water footprint accounting, like the variability of water quality standards or the account of multiple pollutants (Liu et al. 2017); the variability in emission factors from inorganic and organic fertilizer production and use, which often results in high uncertainty on the outcomes of the analyses (Walling and Vaneeckhaute 2020); the need to adjust emission factors to the particularities of the cropping systems of each region regarding water management, crop type, and fertilizer management (Cayuela et al. 2017; Minardi et al. 2022).

Recent policy initiatives, like the European Green Deal, and emerging soil carbon markets, proposed to align with the United Nations Sustainable Development Goals and accelerate management transitions, are promoting new paradigms in different economic sectors, including agriculture (European Comission 2023; United Nations 2023). These policies and regulations require rational methods to assess the impacts of freshwater use in agriculture and measure GHG emissions but also the sequestration of SOC and the carbon sinking potential of agroecosystems (Blasi et al. 2016; Stanley et al. 2023).

Conclusions

The current study provided insights into the water and carbon footprints of irrigated Mediterranean vineyards. The results point to the relationships between different components of these footprints, namely related to energy and agrichemicals use, and interconnection between environmental conditions, agricultural practices, and yield on the footprints of irrigated agroecosystems. The adoption of practices like deficit irrigation strategies, suitable variety rootstock selection, low to no-till or the use of cover crops, can promote an increase in water use efficiency and in soil, water and biodiversity conservation, reducing farm inputs and environmental footprints. Therefore, the potential for carbon sequestration of biomass residue management and soil management practices, like cover crops in the interrow of vineyards should be properly credited in CO2 accounting. Moving away from conventional farming to low-input systems, like integrated, organic or regenerative viticulture is a pathway that is already recognized by the International Organization of Vine and Wine (OIV) and being followed in different viticultural regions. Using environmental footprints to measure the demand for natural resources linked to farm practices can lead to some considerations about the importance of carefully considering the trade-off between productive and environmental consequences of farming practices and driving farmers toward maintaining or increasing sustainable actions. The contribution of different farming options to increase the water and carbon efficiency in grapevine production under the dry Mediterranean conditions of southern Europe is a field of study that can provide farmers, planners, and policy makers with valuable information to an effective green transition of viticulture.