Over the past 15 years, a large number of papers have estimated ILUC emissions for various biofuel pathways, including ethanol and biodiesel produced from food crops (grains, sugar crops, and edible vegetable oils) and lignocellulosic feedstocks (some examples are: Searchinger et al., 2008; Al-Riffai et al., 2010; Hertel et al., 2010; Tyner et al., 2010; and Taheripour et al., 2017). However, these papers have exclusively estimated ILUC values for road transportation biofuels. More recently, Zhao et al. (2021) have estimated ILUC values for 17 Sustainable Aviation Fuel (SAF) pathways also produced from food crops and lignocellulosic energy crops. The results of this paper show that ILUC values for aviation biofuels that can be produced from food crops are all positives (e.g., 20, 22.5, 34.6 gCO₂e/MJ for United States soybean oil, Brazilian soybean oil, and palm oil produced in Malaysia and Indonesia, respectively). To the best of our knowledge, no major effort has been made to estimate ILUC values for SAF pathways that can be produced from oilseed cover crops cultivated in rotation with other crops in a double-cropping system. Here, we evaluate ILUC values for three varieties of oilseed cover crops in the United States: carinata, camelina, and pennycress.

We follow Zhao et al. (2021) to assess ILUC values for the cover crops mentioned above. These authors have introduced SAF pathways in the Global Trade Analysis Project-Biofuel (GTAP-BIO) model, which has been frequently used to assess the ILUC emissions due to biofuel production (e.g., Hertel et al., 2010; Tyner et al., 2010; and; Taheripour et al., 2017). GTAP-BIO is a Computable General Equilibrium (CGE) model that tracks production, consumption, and trade of all goods and services produced around the world. In particular, this model is designed to assess the land use consequences of alternative biofuel pathways, while it takes into account interactions between agricultural and energy markets and their links to other economic activities. We further developed this model by introducing SAF pathways for carinata, camelina, and pennycress.

Producing oilseed cover crops in rotation with other crops in a double-cropping system provides an opportunity to produce two products with no additional demand for land: Usable oils that can be converted to biofuels (SAF and biodiesel) and meal that can be used by livestock producers as a source of protein in animal feed diets. For example, the oil and meal contents of carinata are about 40 and 60%, respectively.

While production of an oilseed cover crop in rotation with other crops does not need additional land, its production as the second crop, could generate some savings in land use due to providing meal for the livestock industry. The additional meal produced from expansion in production of oilseed cover crops could help the livestock industry to produce more products using less land. The oilseed cover crops meals can be used as substitute for soybean meal or other oilseeds meals and drop the demand for cropland indirectly. Here we examine the extent to which producing SAF and its biodiesel co-products from oilseed cover crops produced in the United States affect land use changes at the global scale.

To accomplish this task, we added several new sectors into the GTAP-BIO model reported by Zhao et al. (2021) to produce carinata, camelina, and pennycress, their oils, and meals. Following these authors, we also considered the Hydroprocessed Esters Fatty Acids (HEFA) technology to convert the oils of these crops to SAF. This technology has been briefly described in S.M.

The oilseed cover crop sectors use intermediate inputs, including seeds, chemicals, energy, services, and other intermediate inputs. They demand primary inputs, including labor and capital (including profits gained above the operating costs) as well. The cost structures of these crops were determined using the techno-economic analyses used in the lifecycle assessments of these crops (Prussi et al., 2021) and in consultation with experts from CoverCress Inc.

For each new seed sector, a processing sector is included in the model. The new processing sectors purchase seeds, crush the seeds, convert the extracted oils to SAF (and biodiesel co-product), sell the fuels to the blending sector, produce meals, and sell the meals to livestock producers. The livestock industry uses the meal as feed. The cost structures of these industries follow the HEFA cost structure defined in Zhao et al. (2021).

To construct the benchmark database, the conservative yields of 2.4, 1.1, and 1.7 metric tons per hectare were used for carinata, camelina, and pennycress, respectively. In addition, the oil crushing rates of 0.42 for carinata, 0.39 for camelina, and 0.32 for pennycress were assumed. The model database is modified to include small amounts of outputs for each new pathway. Seeding the industry with a small quantity of the product is necessary to permit simulations in GTAP-BIO.

Finally, to assess the ILUC value for each oilseed cover crop pathway (i.e., carinata oil HEFA, camelina oil HEFA, and pennycress oil HEFA) an exogenous shock by 212.9 Million Gallons of Gasoline Equivalent (MGGE) was introduced to the model. For each pathway, the sock consists of 25% SAF and 75% biodiesel co-product. These shares are embedded in the HEFA technology. Following Zhao et al. (2021), a unique set of model parameters has been used across all examined pathways.

Oilseed cover crops can be produced on marginal land or in rotation with other crops as a second crop in one planting year. In this paper, we only concentrate on the second option. Several papers (Sindelar et al., 2017; Akter et al., 2021) have addressed the plantation of oilseed cover crops in a double-cropping system, and some of the papers estimated the areas that these cover crops could be produced in rotation with other crops in the United States (Embaye et al., 2018; Alam and Dwivedi, 2019). Producing these cover crops in the United States Corn-Belt is an important option. Conventionally, a large portion of cropland in the Corn-Belt has been used to produce corn and soybeans in a 2-year rotation (henceforth: C-S rotation). Therefore, the area that follows this rotation is divided equally between soybean and corn in each crop year. If the land is used for corn production in the first year, that land will be used for soybean production in the second year and vice versa (Table 1). In this crop rotation, the whole area of cropland under corn and soybeans remains fallow in about 5 months from February to April at the beginning and then December and January at the end (Table 1). As mentioned above, some papers have shown that the oilseed cover crops can be introduced into this rotation in a double-cropping system to produce non-food oilseeds used for SAF production. The rotation of corn, oilseed cover crops, and soybean (henceforth: C-SC-S) is presented in the last two lines of Table 1. This table represents a general timeline for the C-SC-S double-cropping rotation. The exact timing of this rotation could vary according to the local agro-ecological conditions, maturity of soybean variety, and the type of cultivated oilseed cover crop. Depending on these conditions, the oilseed cover crops may be cultivated after corn harvest in September or October and harvested in May before soybean cultivation. It is important to note that, as shown in Table 1, with the proposed crop rotation, in between corn and soybean, each piece of cropland land will be cultivated for oilseed cover crops every other year. This means that compared with the conventional corn-soybean rotation, the proposed new rotation drops the area of winter fallow land by half.

To determine the potential area of producing oilseed cover crops in the United States, we rely on the work developed by Sindelar et al. (2017). They estimated the potential area for these crops for the United States Corn-Belt based on the common rotations used during 2009–13 in this region. However, Sindelar et al. (2017) missed two important facts: 1) areas of corn and soybean are growing over time while other crops (including Alfalfa) are declining, and 2) multiple cropping is possible outside the Corn-Belt as well. To remove these deficiencies, we considered area expansion for corn and soybean over time. In addition, we included potential areas outside the Corn-Belt. We worked with the average area of 2010–2020 based on USDA and estimated the potential area for producing oilseed cover crops for 2021 to 2035. The S.M. represents details of this estimation.

Finally, it is important to note that carinata, camelina, and pennycress can be produced on marginal cropland or even productive cropland as the main crop in late spring and summer. However, these alternatives may compete with other food crops and cause induced land use changes. The CORSIA framework has not considered these practices in providing feedstock for SAF production. Hence, within the CORSIA framework, oilseed cover crops should not displace other crops. The areas of land that we considered in this paper satisfies this limitation.

The annual emission savings due to the use of a given oilseed cover crop ( E S t j measured in a million metric tons of CO₂e) for SAF production could be gained from various channels, including:

1) Savings in emissions due to replacement of SAF with conventional jet fuel, R t j

Therefore E S t j =   R t j + L t j + S O C t j   (1) Where: R t j = ( E c o n v − E S A F ) ∗ Q t j (2) L t j = − ( I L U C j ) ∗ Q t j (3) S O C t j = θ j ∗ A t j (4) In these formulas, Q t j represents total energy content of produced SAF and its related energy co-products measured in Mega Joules (MJ), E c o n v and E S A F show life cycle emissions for conventional jet fuel and SAF, both measured in a million metric tons of CO₂/MJ, I L U C j measures land-use change emissions due to cultivation of seed j measured in a million metric tons of CO₂e/MJ, θ j captures improvements in SOC due to cultivation of seed j measured in gCO₂e/ha, and A t j shows hectares of cultivated seed j for SAF production.

The S.M. further describes these formulas and variables and provides sources of the implemented data for each equation.

Unlike biofuels produced from food crops, producing SAF from oilseed cover crops provides savings in cropland demand for two main reasons. First, producing these crops in a double cropping system in rotation with other crops does not increase demand for cropland. Second, converting these crops to SAF produces some meals that can be used by the livestock industry and that leads to savings in demand for cropland and hence less demand for deforestation.

Figure 1 shows the estimated induced land-use changes for the examined pathways. It shows that producing SAF from oilseed cover crops, regardless of the seed type, generates some savings in demand for cropland (blue bars). That leads to savings in land conversion towards active cropland [positive changes in: forest (dark green bars), pasture (light green bars), and marginal land (orange and yellow bars)]. For the same expansion in SAF, among the examined pathways, carinata and pennycress generate the lowest and highest savings in demand for cropland as shown in Figure 1. Two factors explain this observation. The first factor is seed yield. Carinata has the highest yield. The higher the yield is, the lower the saving in cropland will be for a given expansion in SAF. The second factor is the meal content of seed. Again, carinata has the lower meal content (or the highest oil content). For these two reasons, for a given shock in production of SAF, carinata provides the lowest and pennycress delivers the largest savings in demand for cropland. However, it is important to note that for a given area of land, carinata delivers more fuels than pennycress, for the same reason (higher seed yield and higher oil content).

Figure 2 shows the calculated ILUC values obtained for the examined pathways and their decompositions across sources of ILUC. The estimated ILUC values for carinata, camelina, and pennycress are -12.9, -15.3, and -18.3 gCO₂e/MJ. These negative ILUC values reflect the fact that producing SAF from oilseed cover crops leads to savings in GHG emissions due to savings in demand for cropland. While carinata provided the lowest savings in land use emissions per MJ of produced fuels, it could provide more emissions savings for a given available area of land, as this crop produces more energy per unit of land. To highlight the importance of these savings, consider the estimated ILUC values for two other SAF pathways that could be produced in the United States: Corn ETJ and soybean oil HEFA. The estimated ILUC values for these two pathways are 24.9 and 20 gCO₂e/MJ (Zhao et al., 2021). Therefore, producing SAF from oilseed cover crops has an absolute advantage over other SAF pathways that can be produced from food crops.

Figure 2 shows decompositions of the ILUC values for carinata, camelina, and pennycress by the sources of savings in emissions as well. This figure shows that savings in SOC, improvements in agricultural biomass, savings in foregone forest sequestration, and savings in losses of natural vegetation are the main sources of emissions savings for the examined oilseed cover crops; for details, see S.M. From this perspective, carinata, camelina, and pennycress are very similar.

The results show that there is a potential of 28.7 million hectares of land that we can sow oilseed cover crops such as carinata, camelina, and pennycress in 2021 (Figure 3). From this area, around 25.3 million hectares are in the United States Corn-Belt, and the other 3.4 million hectares are for the rest of the United States (not shown in the Figure). The potential area is expected to increase to 29.3 million hectares in 2035, as shown in Figure 3. Sindelar et al. (2017) also found that there is a potential of 27.1 million ha in the United States Corn Belt for sowing bioenergy crops. They also concluded that several factors affect the adoption rate of the farmers to choose oilseed cover crops in either corn-soybean rotation or corn-corn-soybean rotation.

In practice, the extent to which this area of land will be used for SAF production depends on various factors. The overall national policy and the plan of the United States aviation industry on using SAF are critical factors. A national mitigation policy that supports SAF production (e.g., financial support, mandates, and low carbon fuel standard) could generate major demand for oilseed cover crops encouraging farmers to adopt the C-CS-S rotation in practice. A major commitment to using SAF by the main operators of the aviation industry could cause the same effects. A secure demand for oilseed cover crops, providing educational and logistic supports for the farm industry, will encourage farmers to adopt the C-CS-S and other types of rotations better and faster to produce more oilseed cover crops. In practice, it depends on the regional climate and agronomy conditions, the costs of producing each type of oilseed cover crops, and provided policy incentives the farmers are expected to adopt and produce a mix of these crops. These factors jointly will determine the penetration rate for oilseed cover crops in each region.

After knowing the potential area for producing oilseed cover crops, we could assess the potential production of oilseed cover crops. As addressed before, the existing literature mainly shows that carinata, camelina, and pennycress are the main cover crops that can be produced in a double-cropping system in rotation with other crops across the United States. As shown in Results and Discussion of the S.M., yields of these crops are not similar and vary by region and location of production. In addition, the extent to which each of these crops will be produced by farmers is uncertain. However, it is straightforward to provide a Production Possibility Frontier (PPF) for producing these crops at a macro level for United States agriculture. To define this PPF, since carinata and camelina are similar crops in terms of productivity and oil content, we concentrate only on carinata and pennycress. Hence, the PPF represents the maximum trade-off between the two selected oilseed cover crops.

To construct the PPF, we considered two extreme scenarios of devoting all available land in winter either to carinata or pennycress. As an example, consider Figure 4, which represents the represents the maximum trade-off between the two selected oilseed cover crops for 2035. In this year, as mentioned in the previous section, the total available area for producing oilseed cover crops will be about 29.3 million hectares. This area could be entirely devoted to the production of carinata. In this option, given the assumed carinata yield of 2.4 metric tons per hectare, the total available carinata seed for SAF production will be about 70.7 million metric tons (point A in Figure 3). On the other hand, if the entire land is devoted to pennycress in 2035, given the assumed yield of 1.68 metric tons per hectare, the available pennycress seed for SAF production will be about 49.5 million metric tons (point B in Figure 4). Any point on or under the PPF for 2035 (line AB in Figure 4) shows a possible option for the mix of carinata and pennycress feedstock for SAF production in 2035. For instance, point C shows the mix of the available land divided between carinata and pennycress half by half. Note that the annual PPF lines for 2021 up to 2035 fall below the PPF for 2035. However, the changes are not huge, as the potential area is expected to increase slowly. Indeed, the PPF for 2035 represents the set of potential mixes that might occur in the future in the production of oilseed cover crops for SAF.

Table 2 provides information about the annual emissions savings that can be achieved for the period of 2021–2035 for the two extreme cases mentioned above. The first scenario devotes the entire potential land for oilseed cover crops to carinata production. The second scenario allocates the entire land to pennycress. For the first scenario, the overall potential emissions savings increase from 91 MMT of CO₂ emissions in 2021 to 92 MMT of CO₂ emissions in 2035. The corresponding figures for pennycress for 2021 and 2035 are 50 and 51 MMT, respectively. Therefore, carinata can make more emissions savings than pennycress if it can be produced all over the potential area for the oilseed cover crops. The higher emission saving for carinata can be due to higher yield per hectare (2.4 versus 1.68 MT) as well as higher oil yield for carinata (43 versus 29%). Finally, Table 2 shows the potential savings in emissions for a scenario that equally divides the potential area for oilseed cover crops between carinata and pennycress. The estimated potential emissions savings for this case is about 70 MMT for 2021. It could increase to 72 MMT in 2035.

Therefore, the total annual emissions savings from SAF obtained from oilseed cover crops in 2035 could be between 52 and 92 MMT. These emissions savings are about 6–11% of the current annual emissions generated by the aviation industry on a global scale. In other words, if we consider oilseed cover crops in the corn-soybean rotation, we will be able to make enough feedstock to decrease the emissions of the aviation industry at the global scale up to 11%.

To better understand where the total emission savings come from, Figure 5 shows the decomposition of the total emissions saving for the third scenario mentioned above. The core LCA makes more than 70% of savings which is the highest. In contrast, improving SOC due to the cultivation of oilseed cover crops has the lowest share in total savings, around 11%. Savings due to ILUC stands in the middle by around 17% of total CO₂ savings.