Land use for biofuels vs. solar¶

Introduction and summary¶

In our article “Putting solar panels on land used for biofuels would produce enough electricity for all cars and trucks to go electric”, we explore how much land is currently used to grow crops for biofuels, and the opportunity cost of that land compared with using it for solar photovoltaic (PV).

This technical document serves as a companion to that article. Here, we explain how our estimates are derived and compare our assumptions with alternative ones in the literature.

In case you are short of time, the main conclusions of this document are as follows:

• The net land used to grow feedstocks for biofuels is around 32 million hectares, as estimated by Cerulogy. This is similar to the land area of Cote d'Ivoire, and roughly between Poland's and Germany's. With that amount of cropland, biofuels produce 1.42 PWh of energy, according to the Energy Institute's Statistical Review of World Energy.

• If we could hypothetically cover the same area with solar PV farms, under reasonable assumptions on modern solar energy density and capacity factors (from IRENA, LBNL, and other sources), we could produce around 32 PWh of electricity. This surpasses the world's total electricity generation in 2024. This also implies that, using the same amount of land, we could produce ~22 times more energy with solar panels than growing crops for biofuels.

• If we replaced vehicles currently powered by biofuels with electric vehicles (under reasonable assumptions on fuel-to-energy ratio for combustion and electric engines), we could power them using just ~2% of that land with solar farms. Globally, around 26 PWh of fuel energy is consumed for road transport (e.g. cars and trucks), according to the IEA's World Energy Outlook. If we replaced them with electric vehicles, they'd need less than 10 PWh of electricity, which could be produced with solar panels in less than a third of the land currently used for biofuel feedstock.

How much area is globally used for biofuel production?¶

Our estimate based on Cerulogy¶

This report by Cerulogy estimates that there were 61.3 million hectares of feedstock in 2023.

However, once they take away biofuel co-products, they estimate a net area of 32.0 million hectares of cropland attributed to first-generation biofuel feedstock.

Their analysis includes nine regions (United States, Brazil, EU + UK, Indonesia, China, India, Argentina, Canada, and Thailand) which cover 95% of the global biofuel production. However, note that 95% of production from those countries doesn't translate into 95% of land; in fact those countries may import feedstock from other countries. They do point out this caveat:

Global trade in biofuel feedstocks (rather than finished fuel), which is not captured by this data, is also sizeable; but it is not trivial to determine from general trade data which feedstocks are used in the biofuels industry versus other industries (cf. Malins & Sandford, 2022).

Despite this caveat, we will take Cerulogy's estimate at face value for the remainder of the document, and assume that global biofuel feedstock takes 32 million hectares of cropland.

Alternative estimate by UFOP¶

The UFOP Report on Global Market Supply 2022/2023 says that:

Crop plants were grown on more than 1.4 billion hectares worldwide in 2021. [...] Only around 8 per cent of the area was used in biofuels production.

FAOSTAT's Land Use dataset estimates that global cropland area in 2021 was 1.57 billion hectares (and hasn't changed much in the last decade). They don't provide more specific numbers in the cited report, so it's unclear what the estimated area of 1.4 billion hectares refers to exactly.

Their estimates come from the Agricultural Market Information Company (AMI), based on different sources: OECD, USDA, Oil World. However, the relevant AMI data seems to be under a paywall.

Similar values between 7% and 8% have been quoted multiple times elsewhere (e.g. Oils & Fat International, Food Unfolded, Biofuels international or Advanced Biofuels USA). They all cite UFOP, even though the origin of those estimates is unclear.

Taking the UFOP values at face value, 8% of 1.4 billion hectares would correspond to 112 million hectares. And applying that same percentage to the actual global cropland estimated by FAOSTAT (1.57 billion hectares), would lead to 126 million hectares.

Their estimates don't seem to adjust for the land use of co-products. But even assuming that the adjustment would cause around a 50% reduction, the resulting value would still be significantly larger than Cerulogy's estimate.

Given that we couldn't find a clear origin for UFOP's large estimate on the extent of biofuel croplands, we prefer to rely on Cerulogy's estimate.

Alternative estimate by Popp et al. (2018)¶

Popp et al. (2018) estimated that 2% of global cropland is devoted to biofuels (30–35 million gross hectares). After removing co-products, the area would be smaller. They say:

If we include co-products [...] then the land that is needed to grow feedstocks reduces [...] to about 1.5%.

This would imply around 23 million hectares.

It's unclear which year their estimates are based on, but they quote explicitly:

At present, around 2% of the 1.515 billion ha which makes up the total global crop area (FAO, 2013).

So, what they refer to as present may be based on 2013 FAO data (which could include data up to 2011 or 2012). Since then, biofuel production has increased by ~86%, so it would be reasonable to expect a significantly larger net area today, closer to Cerulogy's estimate.

How much electricity can be produced by solar PV in that area?¶

Our estimate based on various sources¶

If we were to produce solar energy on biofuel cropland now, we'd install solar panels with modern, utility-scale capacity factors.

We can find representative capacity factors in the latest IRENA's Renewable Power Generation Costs report. In Table 3.2 they show global weighted average capacity factors (AC/DC) for utility-scale solar PV systems by year of commissioning (page 99).

The AC/DC factors reported for 2024 (converted to AC/AC assuming an inverter loading ratio of 1.25) are:

• 5th percentile: 11.5% (AC/DC) x 1.25 ~ 14% (AC/AC)

• Weighted average: 17.4% (AC/DC) x 1.25 ~ 22% (AC/AC)

• 95th percentile: 22.6% (AC/DC) x 1.25 ~ 28% (AC/AC)

After speaking with various experts in the field, we concluded that a typical modern installation has a power density of 1 MWp per 1.5 hectares (in terms of DC power at peak). In terms of AC power, this would correspond to ~0.53 MW(AC) / ha.

We can convert power density to an average energy density using IRENA's average DC conversion factor:

( 1 MWp / 1.5 ha ) x ( 1 GWh / 1000 MWh ) x ( 17.4% of 8760 h / 1 year ) = 1.02 GWh / ha / year

We will therefore assume a fiducial value for energy density of 1 GWh / ha / year.

A conservative range of power densities would be somewhere between 0.3 MW(AC) / ha and 0.7 MW(AC) / ha, around the midpoint value of 0.53 MW(AC) / ha chosen before.

Combining the ranges of capacity factors and power densities, we can calculate a lower limit, a midpoint, and an upper limit for the amount of electricity produced in a year in 32 million hectares of land:

• Lower limit: 32 · 10⁶ ha x ( 0.3 MW / ha ) x ( 1 PWh / 10⁹ MWh ) x ( 14% of 8760 h / 1 year ) = 12 PWh / year

• Midpoint: 32 · 10⁶ ha x ( 1 GWh / ha / year ) x ( 1 PWh / 10⁶ GWh ) = 32 PWh / year

• Upper: 32 · 10⁶ ha x ( 0.7 MW / ha ) x ( 1 PWh / 10⁹ MWh ) x ( 28% of 8760 h / 1 year ) = 55 PWh / year

Our fiducial value of 1 GWh / ha / year could be considered conservative, given that:

• A significant fraction of the biofuel cropland is in the US and Brazil, which have good conditions for solar generation.

• The hypothetical transition from biofuel cropland to solar panels would take place in the coming years, so efficiencies would improve slightly with respect to the present.

Therefore, if we filled the entire cropland of biofuel feedstocks with solar panels, we could produce ~32 PWh per year. This value is just above the world's total electricity production, which was 31 PWh in 2024, according to Ember's Yearly Electricity Data. Currently, ~18 PWh of that electricity comes from fossil fuels, and only ~2 PWh from solar.

According to the Energy Institute's Statistical Review of World Energy, the energy that was actually produced by biofuels in 2024 was 1.42 PWh. Therefore, if we used solar farms, we could produce ~22 times more energy (in electricity) in the area currently devoted to biofuel crops.

Alternative estimate based on Cerulogy¶

Cerulogy assumes an average power density of 0.08 kWp / m² for a solar farm, and a farm productivity of 3.9 kWh / kWp / day. This leads to an energy density of:

( 0.08 kWp / m² ) x ( 3.9 kWh / kWp / day ) x ( 365 days / 1 year ) x ( 10⁴ m² / 1 ha) x ( 1 GWh / 10⁶ kWh ) =  1.14 GWh / ha / year

This energy density is 14% larger than the one we assumed. Therefore, the estimated amount of electricity that solar PV could produce on 32 million hectares is also larger:

32 · 10⁶ ha x ( 0.64 MW / ha ) x ( 20% of 8760 h / year ) x ( 1 PWh / 10⁹ MWh ) = 36 PWh

Even though we take Cerulogy's estimate of biofuel cropland, we prefer to use our more conservative value of energy density.

Alternative estimate based on UNECE's lifecycle assessment¶

We can find alternative estimates of solar energy densities in this UNECE report (and this corrigendum), which is also used in this OWID article.

UNECE does a lifecycle assessment of the amount of land required per year to produce 1 MWh of electricity. This is estimated over the entire production process (including the extraction of materials, refining processes, etc.).

Their estimated land use depends on whether they are installed on-ground, or on roofs. Given that we want to consider the land currently used for biofuel production, we will consider only on-ground PV, specifically poly-silicon. Therefore, the minimum, average, and maximum (lifecycle) land used per MWh is 12, 19, and 37 square meters-annum.

If we used 32 million hectares of land for solar electricity, they would generate:

• Minimum: 32 · 10⁶ ha x ( 10⁴ m² / 1 ha ) x ( 1 MWh / 37 m² ) x ( 1 PWh / 10⁹ MWh ) = 8.6 PWh

• Midpoint: 32 · 10⁶ ha x ( 10⁴ m² / 1 ha ) x ( 1 MWh / 19 m² ) x ( 1 PWh / 10⁹ MWh ) = 17 PWh

• Maximum: 32 · 10⁶ ha x ( 10⁴ m² / 1 ha ) x ( 1 MWh / 12 m² ) x ( 1 PWh / 10⁹ MWh ) = 27 PWh

Note that these lifecycle values are not directly comparable to Cerulogy's cropland estimate, since they do not represent the physical land footprint of solar farms. But these values can provide a rough lower bound on solar land intensity.

Alternative estimate based on LBNL¶

Bolinger & Bolinger (2022) (Lawrence Berkeley National Laboratory, LBNL) estimate the following values for power and energy densities in the US:

• Power density:
  • Fixed-tilt: 0.87 MW(DC) / ha -> 0.69 MW(AC) / ha
  • Tracking plants: 0.59 MW(DC) / ha -> 0.45 MW(AC) / ha
• Energy density:
  • Fixed-tilt: 1.10 GWh / ha / year
  • Tracking plants: 0.97 GWh / ha / year

Assuming tracking plants:

32 · 10⁶ x ( 1 PWh / 10⁶ GWh ) x ( 0.97 GWh / ha / year ) =  31 PWh / year

This is in very good agreement with our estimate, given that their energy density for tracking plants is very close to our fiducial value of 1 GWh / ha / year.

Alternative estimate based on Victoria et al. (2021)¶

An alternative estimate of solar energy density comes from Victoria et al. (2021). In their supplementary Note S3, they say:

[...] Assuming 17% efficiency and that only 30% of the land is covered by solar panels in large-scale installations, the capacity density results in 51 W / m², which is in agreement with values reported in Ong et al. (2013). For an average annual generation for solar PV of 1370 kWh / kW, 38 million ha would be needed. The land area of the world is 13,003 million ha OWID (2019). Hence, our current electricity consumption could be supplied by solar PV covering 0.3% of the available land.

We can translate these values into:

( 51 Wp / m² ) x ( 1 kWp / 10³ Wp ) x ( 10⁴ m² / ha ) x ( 1370 kWh / kWp / year ) x ( 1 GWh / 10⁶ kWh ) = 0.70 GWh / ha / year

This would lead to a total solar production in the current biofuel cropland of:

32 · 10⁶ x ( 1 PWh / 10⁶ GWh ) x ( 0.70 GWh / ha / year ) = 22 PWh

They say that their values are consistent with NREL's Ong et al. (2013). The power density of 0.41 MW / ha is on the low side, and the capacity factor of 20% is just a bit below our fiducial value. We consider that these relatively old US estimates may not be representative of today's achievable global energy density.

Other alternative estimates¶

Victoria et al. (2021) (in their supplemental information) concluded that we could power all electricity (27 PWh in 2019) with 38 million hectares of solar PV.

Nøland et al. (2022) (Table 16) estimated an average annual energy density for utility-scale solar PV of 0.087 ± 0.029 TWh / km2 (and a median of 0.080), based on 17 solar farms in different countries. The mean value corresponds to 0.87 GWh / ha / year, which is 13% lower than our fiducial value of energy density.

Franz and Dumke (2025) (Table 6) estimate a mean energy density of 93 kWh per m² per year (and a median of 91), based on 107 free-field PV power plants across Europe from 2006 to 2022. The mean value corresponds to 0.91 GWh / ha / year, which is 9% lower than our fiducial value.

One way to improve our estimates would be to estimate the energy density at the country level. Even within a country, there would be some hotspots where solar farms could be most effective (e.g. cropland close to transmission lines), as shown in Sturchio et al. (2025) for the US. But those improvements fell out of the scope of this analysis.

How much electricity would be needed to power all road transport?¶

Top-down estimate¶

From the IEA's World Energy Outlook 2025 (free dataset), we can extract the total road energy in 2024, which is 93.3 EJ (or 25.8 PWh). We can safely neglect the fraction of vehicles that are currently powered with electricity.

Let's assume an efficiency of around 30% for combustion cars and 80% for electric cars. Then, we can estimate the amount of energy that would be required to power all vehicles, if they were replaced by electric ones:

93 EJ x ( 0.278 PWh / 1 EJ ) x ( 30% / 80% ) = 9.7 PWh

The amount of energy produced by biofuels in 2024 is, as we saw before, about 1.42 PWh. Let's assume, for simplicity, that all this energy goes to power combustion vehicles. Then, if we replaced those vehicles with electric ones, they would require only 0.53 PWh. To power all those vehicles with solar farms, assuming our fiducial energy density, we'd need:

( 0.53 PWh / year ) x ( 10⁶ GWh / 1 PWh ) x ( 1 ha x year / 1 GWh ) =  0.53 million ha

In other words, if instead of growing crops to feed cars with biofuels, which takes 32 million hectares, we produced solar electricity to power electric cars, we'd need less than 2% of the land for the same amount of transport.

Moreover, if all road transport was electric, they could be powered with just 10 million hectares of solar PV, since:

( 9.7 PWh / year ) x ( 10⁶ GWh / 1 PWh ) x ( 1 ha x year / 1 GWh ) = 9.7 million ha

So, if all road transport was electric, we would be able to power them with just 10 million hectares of solar PV, which is less than a third of the area currently used for biofuel feedstock.

Bottom-up estimate¶

The International Energy Agency estimates that, globally, passenger cars covered 25.8 trillion passenger-kilometers in 2024.

First, we need to convert that to kilometers (or car-kilometers). To do that, we can assume that the average car occupancy is 1.5 (so, 1.5 people in a car, on average). Therefore:

25.8 · 10¹² km passenger x ( 1 car / 1.5 passengers ) = 17.2 · 10¹² car km

That gives 17.2 trillion kilometers.

We'll assume that the average electric car uses around 0.2 kilowatt-hours to drive one kilometer. Covering all those kilometers traveled by car would therefore use:

17.2 · 10¹² car km x ( 0.2 kWh / 1 car km ) x ( 1 PWh / 10¹² kWh ) = 3.44 PWh

That means 3.44 PWh of electricity would be needed to power the global car fleet, assuming all cars were electric.

The efficiency of electric cars varies based on model, age, size, and other factors. We can assume a value of 0.2 kWh per kilometer based on several sources. This analysis by Weiss et al. (2024) of the European car fleet found certified and average ratings of 19 kWh and 21 kWh per 100 kilometers. That's around 0.2 kWh per kilometer.

You can find the efficiency of different models in this EV Database. The average across this huge number of models and brands was 0.19 kWh per kilometer.

Doing the same estimate for trucks is a bit more challenging, because their electric models have not reached market penetration in the way that electric cars have. They are at an earlier stage of development, so numbers on energy efficiency are more difficult to find.

The International Energy Agency estimates that heavy-duty trucks covered 35,000 billion tonne-kilometers in 2024.

To convert to kilometers, we assume that the average truck load is around 12 tonnes. That gives:

35 · 10¹² tonne km x ( 1 truck / 12 tonnes ) = 2.92 · 10¹² truck km

We assume that a medium heavy-duty truck uses around 1.2 kWh per kilometer driven (if electric). Covering 2900 billion kilometers therefore uses:

2.92 · 10¹² truck km x ( 1.2 kWh / 1 truck km ) x ( 1 PWh / 10¹² kWh ) = 3.50 PWh

That's 3.50 PWh of electricity to power the global truck fleet, assuming they were all electric.

Again, the efficiency of electric trucks varies based on a range of factors including size, weight, and design. We assume an efficiency of 1.2 kWh per km, based on published estimates from a number of sources.

The vehicle manufacturer Scania reports that its electric trucks have an efficiency of 1.1 kWh per km. An independent group tested the truck and found a similar result of 1.15 kWh per km.

During an 18-month trial of twenty 19-tonne electric trucks, the average efficiency ranged from 0.8 to 1.1 kWh per km, depending on urban or rural driving.

A recent study conducted by the International Council on Clean Transportation found a mean consumption in the range of 1.0 to 1.2 kWh per km.

Larger trucks tend to require more electricity per kilometer, and some will have an energy consumption higher than 1.2 kWh per km. However, this will tend to average out across an entire truck fleet.

In summary, we estimate that the amount of electricity that would be needed to power all road transport, if fully electrified, would be around 7 PWh per year, comprising 3.5 PWh for cars and a similar amount for trucks.

Bibliography¶

Franz, M., & Dumke, H. (2025). Evolution of patterns of specific land use by free-field photovoltaic power plants in Europe from 2006 to 2022. Energy, Sustainability and Society, 15(1), 12.

Nøland, J. K., Auxepaules, J., Rousset, A., Perney, B., & Falletti, G. (2022). Spatial energy density of large-scale electricity generation from power sources worldwide. Scientific Reports, 12(1), 21280.

Ong, S., Denholm, P., Heath, G., Margolis, R., & Campbell, C. (2013). Land-Use Requirements for Solar Power Plants in the United States. https://doi.org/10.2172/1086349

Popp, J., Kot, S., Lakner, Z., & Oláh, J. (2018). Biofuel Use: Peculiarities and Implications. Journal of Security & Sustainability Issues, 7(3).

Popp, J., Harangi-Rákos, M., Gabnai, Z., Balogh, P., Antal, G., & Bai, A. (2016). Biofuels and their co-products as livestock feed: global economic and environmental implications. Molecules, 21(3), 285.

Sandford, C., Malins, C., Phillips, J. (2024). Diverted harvest. Environmental Risk from Growth in International Biofuel Demand. Cerulogy. https://www.cerulogy.com/diverted-harvest/

Sturchio, M. A., Gallaher, A., & Grodsky, S. M. (2025). Ecologically informed solar enables a sustainable energy transition in US croplands. Proceedings of the National Academy of Sciences, 122(17), e2501605122. https://doi.org/10.1073/pnas.2501605122

Victoria, M., Haegel, N., Peters, I. M., Sinton, R., Jäger-Waldau, A., del Cañizo, C., Breyer, C., Stocks, M., Blakers, A., Kaizuka, I., Komoto, K., Smets, A. (2021). Solar photovoltaics is ready to power a sustainable future. Joule 5, 1041–1056. https://doi.org/10.1016/j.joule.2021.03.005

Weiss, M., Winbush, T., Newman, A., & Helmers, E. (2024). Energy consumption of electric vehicles in Europe. Sustainability, 16(17), 7529.