Food-land system
The BAU scenario leads to increased production-based agricultural emissions by 2050, as expected. In contrast, the NCNC scenario shifts productivity levers for crop and livestock, dropping GHG emissions by 2050 by 50% (3 MtCO2e) compared to the BAU scenario (Fig. 1a, b). This represents a 29% reduction from 2020 levels and a dramatic 73.4% decline from Greece’s agricultural emissions in 2050. This reduction is primarily achieved through livestock-related emissions dropping to 2.47 MtCO2e in 2050, and land use changes leading to increased emission withdrawals of 3.28 MtCO2e in 2050. This improvement in land-use efficiency stems from the assumed shift to more sustainable diets managing the demand-side, and the higher and more efficient agricultural productivity, which enables greater yields without requiring additional inputs or extensive land expansion, as evidenced by declining pastureland areas (Fig. 1e, f). Additionally, emissions from crop production show notable improvements in the NCNC scenario, with a marked divergence from the BAU projections becoming apparent after 2035. Enhanced agricultural productivity is beneficial both in terms of climate change mitigation (GHG emissions), and of domestic agriculture’s competitiveness. This is particularly relevant following the 2023 extreme weather events and 2023–27 CAP implementation, with Devot et al.47 highlighting how climate change-intensified weather events impact EU agricultural production and costs.
The NCNC scenario dramatically cuts emissions (−58%) and production costs (−46%) compared to the BAU by 2050, largely due to healthier diets (less meat), higher productivity, and reduced reliance on inputs. Land allocation shifts strongly: pasture declines, while “other land” expands, reflecting land sparing and improved efficiency. Cropland remains roughly constant in both scenarios, while forests are protected in both.
Under the NCNC scenario, total costs are projected to decrease from 828 million € in 2025 to less than 630 million € by 2050. This reduction is largely attributed to declining pesticide expenses, which constitute the majority of total costs. Most notably, producers’ pesticide expenditures decrease by 27.5% between 2025–2050 in the high productivity scenario, amounting to just 40% of the costs projected in the BAU. While fertilizer costs show a more modest decline of 14.8% over the same 25-year period, the contrast with the BAU’s upward trend still results in significant cost savings of nearly 40%. This demonstrates how improved productivity can strengthen competitiveness while adapting to climate challenges.
The two main messages of the national land-use policy frameworks (Table 1) are to discourage urban sprawl in agricultural and natural areas, and increase forest area, ensuring that agricultural expansion does not come at the expense of woodland health. Under the BAU scenario, none of those two goals is being achieved (Fig. 1e). While urban land grows only modestly, much of the land-use change is driven by cropland and harvested area, which means agricultural expansion is the dominant driver of sprawl into rural and natural areas (directly at odds with discouraging land-take outside existing settlements). Also, there is no meaningful increase in forest area, but a slightly downward trend. Under the NCNC scenario, only one of these two land-use objectives is met (Fig. 1f): Urban land remains essentially flat through 2050, showing that new housing and infrastructure are indeed being held within existing settlements. However, the pronounced increase in agricultural productivity and the shift to healthier diets, associated with reduced red meat consumption, leads to the marked drop in pastureland (by 29% in the 2020–2030 period, and 78% in 2050, compared to the 2020 levels; The respective decreases for the BAU scenario are 17.3% and 31.2%, respectively). Given the lack of national commitment for a quantitative afforestation target and the marginal reduction in cropland, this leads to a significant surge in the area described as “Other” Land in the FABLE Calculator. “Other land” increase reflects areas under restoration, set-aside, or low-intensity uses, but these gains do not translate into designated, fully protected forests. So, the aspiration to increase forest area and strengthen woodland health remains unmet under the NCNC scenario.
Cross-sectoral energy-emissions analysis
The energy-emission simulation of all sectors was performed for the BAU scenario, continuing current accounts’ trends and assumptions, and the NCNC scenario, which is in essence the Greek NECP. The parameters that are changing according to the specific NECP recommendations include the fuel mix shares serving the demand (increasing the share of cleaner fuels), and improvements in energy efficiencies per sector and use.
The results project a significant reduction in energy consumption and emissions under the NCNC scenario, in contrast to the BAU (Fig. 2a, b). An overall reduction in energy demand of ~20% is observed, with the most drastic reductions achieved in industry (~48%), passengers and freight transportation (~52%), including international aviation and maritime (~49% overall). Improvements in energy efficiency are mainly driving these trends. The decreasing trend of residential energy consumption over time is primarily driven by the country’s shrinking population (AL). The services sector, including public buildings, hotels, hospitals, exhibits a ~18% increase in energy consumption, following increased future needs for services. Agriculture’s energy consumption increases by ~12%, following the increased productivity requirements simulated in the FABLE Calculator. Overall, the level of energy consumption is estimated to remain significantly high in 2050, under both scenarios.
There is an overall energy demand reduction by ~20% under NCNC versus BAU, driven primarily by industry and transportation, while GHG emissions follow analogous and substantial downward trends, under the NCNC scenario.
With respect to the supply side (energy generation), Fig. 3a, b illustrate the total energy generated per feedstock fuel type, which is then used to cover the consumption. As expected, there is a substantial decline in oil refining products under NCNC, almost by 3 times in 2050. Conversely, electricity production is expected to rise significantly, by 3.7 Mtoe in 2050. New contributions to energy production include hydrogen and synthetic fuels, reaching in total 1.1 Mtoe and 0.571 Mtoe, respectively. The shift in energy production types, highlighted by the reduced reliance on conventional petroleum products and fossil-based electricity, contributes to further GHG emission reductions. Emissions are projected to decrease from 26 MtCO₂e in 2022 to 5.2 MtCO₂e by 2050. These changes are attributed to the evolving energy mix and technologies introduced under the NCNC scenario.
Under NCNC the supply mix shifts by 2050, with oil refining falling by ~68%, electricity generation increasing by ~3.6 Mtoe, and new shifts (hydrogen + synthetic fuels) contributing about 1.7 Mtoe in total. These shifts drive a reduction in emissions from energy generation of ~20 MtCO₂e.
Note that both energy consumption and fuels supply results were also validated, cross-checking with data from NCNC’s assumptions, EUROSTAT48, and the IEA49.
The NECP-projected energy sources, particularly for electricity, indicate a complete phase-out of lignite for electricity production, a 77% reduction in natural gas use, and substantial increases in clean energies (renewables, hydrogen and synthetic fuels). These have been also simulated in detail. Indicatively for the significant changes that are projected, we mentioned that wind and solar power deployment are about to increase by 540% by 2050, while the hydroelectric power output is projected to rise by 120%.
The implementation of the NECP would lead to a dramatic reduction of GHG emissions by 2050 compared to the BAU scenario, decreasing by 91.7% (Fig. 2c, d, Fig. 3c, d). These emissions are calculated using the 100-year Global Warming Potential (GWP) of direct GHG emissions and are predominantly composed of Carbon Dioxide (CO₂), with smaller contributions from Methane (CH₄), Nitrous Oxide (N₂O), and Carbon Monoxide (CO). By 2050, the NCNC scenario achieves near-complete decarbonization, whereas the BAU emission reductions are more minor. At this stage, it is worth commenting again on the key difference between the FABLE Calculator’s production-based agricultural GHG emissions and LEAP’s energy-based agricultural GHG emissions. FABLE Calculator estimates agricultural GHG emissions by simulating food and livestock production processes, including land use changes, agronomic practices, and non-energy-related processes (such as enteric fermentation, manure management, and fertilizer application), capturing thus a broader range of emissions associated with agricultural production. Hence the term “production-based emissions” for FABLE Calculator. Complementarily, the LEAP model calculates the emissions based solely on the energy use in production processes (per unit of final products). Hence the term “energy-based emissions” for LEAP.
In general, regarding the total GHG emissions, the primary driver of the reductions in the total emissions (both from energy consumption and energy generation) is the significant decrease in fossil fuel use across the residential, industrial, and transportation sectors, one of the core recommendations of the NECP. Additionally, the adoption of renewable energy sources in electricity production—coupled with the introduction of hydrogen and synthetic fuels, particularly in the transportation sector—further contributes to these reductions. Figure 4a, b show the flows of feedstock fuels into energy transformation processes to produce fuels that cover different energy demand uses, indicatively for 2050. The transition to cleaner fuels is obvious, as mentioned. Both Sankey diagrams indicate that the energy production-transformation-consumption balance is “confirmed” throughout the simulation period.
Under NCNC most emissions cuts come from a sharp drop in fossil fuel use, with renewables and minor contributions from cleaner fuels replacing a large share of fossil fuels and oil products.
Biofuel production potential
As mentioned, the agricultural output results of the FABLE Calculator are analyzed through the BiofuelGCH Calculator, to account for the residues available for biofuel production (without affecting food production), and estimate this potential. This refers to the amount of bioethanol (produced from corn, sugarbeets, and wheat residuals), and the amount of biodiesel (produced from sunflower and olive residuals). So, it does not take into account the wooden and pellet potential production, which is however the major use of biomass for residential heating and cooking.
The results indicate that there is a significant potential to produce biofuels domestically, ranging from 208–435 ktoe in 2022 to 268–519 ktoe in 2050. This production can fully cover the biofuel demand from uses such as agriculture, transportation, energy production and transformation processes (Fig. 5a), and the excess amount can be used for exports (Fig. 5b).
By 2050 the production potential (min-max) ranges from 268–519 ktoe, and it is able to cover the domestic demand (excluding wood and pellet products), providing thus the option for exports up to 253–503 ktoe under the NCNC demand case.
Land requirements
The implementation of the NCNC, as simulated in LEAP, requires in total 35051 MW of solar energy, and 24780 MW of wind power in 2050. This corresponds to an additional capacity of 28051 MW and 16280 MW, respectively, compared to the current (2025) solar and wind power. Moreover, the NCNC projects that 52.46% of the wind power will be onshore, while the rest should be offshore. So, this results in 8541 MW.
The LandReqGCH model, based on these figures, uses typical values from the literature to convert these additional required capacities in solar and wind power into land requirements (km2) for the installation of additional solar panels and wind farms (onshore). These values from the literature are used as land conversion coefficients (km2/MW), taking into account the types of land uses and the types of projects, and considering a range of options, according to Denholm et al.50 and Ong et al.51.
So, for solar panels that would range from 670 km2 (min) to 846 km2 (average) and to 1022 km2 (max). The onshore wind farms would require from 19 km2 (min) to 25 km2 (average) and to 35 km2 (max). These magnitudes, even at their high end, are under 1100 km2 or about 0.8% of Greece’s total land area, but it is non-negligible when overlaid on a landscape already under competing demands. Specifically, in Fig. 1e, f we observe rapid growth in “Other land” (mosaic, low-intensity uses) and continued pressure on cropland and pasture, under both the BAU and NCNC scenarios. Installing solar parks and wind farms will most likely encroach on these less-intensive zones or marginal agricultural lands, rather than pristine forests. If this required renewables infrastructure ends up replacing set-aside fields or semi-natural grasslands, it may conflict with the objective to preserve permanent grasslands and pastures, which will challenge traditional grazing economies and even biodiversity.
The LandReqGCH model also provides estimates of the expected costs for the installation of these projects, considering their typical costs52,53. Regarding the solar panels, the cost would range (min-average-max) from 669.8 million € to 845.7 million € and to 1021.6 million €. The respective costs for wind farms would range from 28.2 million € to 38 million € and to 52.5 million €.
Water Requirements
The WaterReqGCH model was applied for all sectors and years of the studied period, providing also estimates for monthly distributions, accounting thus for seasonality in water requirements. The water sector faces the highest uncertainties, as the consumption is affected by various socio-economic, infrastructure, and hydro-climatological factors that are inherently uncertain. Moreover, there are no specific demand management measures per sector, according to the Greek RBMPs.
Urban water use, encompassing residential and service sectors, represents the 7–8% of total consumption. This comparatively modest share is indicative of more efficient urban water management, for a lower population-driven demand relative to agricultural needs. Urban water consumption decreases from an average of 725.19 hm³ in 2020 to 630.31 hm³ in 2050, driven by Greece’s declining population. The NCNC scenario assumed a reduction in water network losses, so they reach 20% in total. This measure would further reduce the urban water requirements to 578 hm³ in 2050, which is within the estimated range area plotted in Fig. 6.
Urban (residential and services) (a), agricultural (irrigation and livestock) (b), and industrial water requirements (c). The monthly water requirements plot (d) shows the monthly allocation of the total consumption. Agriculture remains dominant (~88–89% of water use to 2050) with strong seasonality, while other uses stay minor, with residential and industry being the most significant ones.
Agriculture is the dominant consumer of water resources, consistently accounting for 88–89% of the total consumption over the period 2020–2050. This is indicative of the sector’s reliance on irrigation and water-intensive practices, which reflect Greece’s Mediterranean climate and the importance of agriculture in its economy. Agricultural water consumption follows a slight increase after 2025 and reaches an average consumption of 8041.12 hm³ by 2050, with only minor fluctuations. The NCNC scenario for agriculture, as defined within the FABLE Calculator, assumes that the number of livestock population and the amount of irrigated areas will remain stable, aiming to higher productivity outputs while using the same input resources. Based on this assumption, the livestock and irrigation water requirements will not vary outside of the plotted uncertainty range for agriculture, as shown in Fig. 6. Another key factor here is the assumption that demand remains stable, driving this relatively stable behavior, which is largely uncertain though.
Industrial use remains the smallest contributor at 3–4%, aligning with Greece’s economic structure, where industrial activity is less dominant compared to agriculture and services. Its water consumption remains relatively stable, with slight increases from 328 hm³ in 2020 to 331.61 hm³ in 2050. The ranges of minimum-maximum values are larger for agriculture and reflect various data and computational uncertainties. The NCNC does not assume any specific measures per industry types’ water use.
The monthly distribution of the total water requirements is shown indicatively for 2020, and follows the same pattern until 2050. It reveals a sharp increase during the prolonged Greek summer period (May–October), reflecting peak irrigation needs and heightened urban water use during the tourist season, and due to increased temperatures. For instance, the average monthly water requirement in July (1866.6 hm³) is more than eight times higher than in December (134.55 hm³). This pronounced seasonality underscores the pressure on water resources during the dry season and the importance of adequate storage and distribution infrastructure34.