Freeing land from biofuel production through microalgal cultivation in the Neotropical region

Biofuel production is a key strategy for reducing CO2 emissions globally and is expected to increase substantially in the coming decades, particularly in tropical developing countries. The adoption of sustainable biofuel production technologies that do not place large demands on agricultural or forested lands, has the potential to make a substantial contribution to decreasing greenhouse gas emissions while reducing biodiversity losses and degradation of native ecosystems resulting from high demand for land. With their high productivity per unit area and ability to grow on non-arable lands, microalgal biofuel production systems could become a major sustainable alternative to biofuel production from food crops (first-generation biofuels). However, the potential impacts of microalgal biofuels on food production, biodiversity, and carbon storage, compared to other biofuel production alternatives, are largely unknown. In the present study, the most suitable areas for siting microalgae production farms to fulfill 30% of future transport energy demands were determined within four Neotropical countries with high population densities and high importance for agricultural expansion and biodiversity conservation globally (Colombia, Ecuador, Panama, and Venezuela). These results were contrasted with the best areas for siting oil palm and sugarcane crops to fulfill the same target in future transport energy demands. Microalgal production systems offer the most sustainable alternative for future biofuel production within the Neotropics. Meeting 30% of future transport energy demands with microalgal biofuels reduced land area requirements by at least 52% compared to oil palm and sugarcane. Furthermore, microalgal biofuel production reduced direct competition with agricultural lands, biodiverse areas, and carbon-rich systems within countries, with little overlap with the biodiverse and carbon-rich rainforests. This study can guide decision making towards the identification and adoption of more sustainable biofuel production alternatives in the Neotropics, helping in avoiding unnecessary environmental impacts from biofuel expansion in the region.


Introduction
Renewable energy sources have strong potential to replace fossil fuels and meet future energy demands (Jacobson and Delucchi 2011, IEA 2017, Obama 2017, thereby mitigating global warming and its negative socioeconomic and environmental impacts (Pecl et al 2017, Nunez et al 2019. Among renewable energy sources, liquid biofuels could offset substantial amounts of fossil fuels (Smith et al 2014, Creutzig et al 2015, particularly in the transport sector, which will continue to need liquid fuels in ships, airplanes, and long-haul trucks (Fulton et al 2015). Biofuel production could increase from 1.7 to 8.1 million Barrels of Oil Equivalent (BOE) day −1 between 2016 and 2040, under the implementation of policies that favor the adoption of renewable energy sources for limiting global warming well below 2 • C in comparison to pre-industrial levels (IEA 2017) and under the Paris Agreement (IPCC 2015).
Biofuel production is currently based on food crops (i.e. first-generation biofuels) that compete with suitable land for agriculture (Lambin and Meyfroidt 2011). This drives direct and indirect land-use changes, often in biodiverse areas (Immerzeel et al 2014, Correa et al 2017. As a result, first-generation biofuels have been linked to habitat losses for native species (Danielsen et al 2009, Immerzeel et al 2014, Elshout et al 2019 and increases in greenhouse gas emissions when carbon-rich ecosystems (e.g. tropical forests and savannas) are transformed into monocultures for biofuel production (Fargione et al 2008, Searchinger et al 2008, 2015. Despite the high biodiversity (Antonelli et al 2018, Rull 2020) and carbon contents found in tropical areas of the world (Avitabile et al 2016), future biofuel production will expand within the tropical region, where large tracts of undeveloped and highly suitable lands for agriculture remain (Laurance et al 2014, Laurance 2015. The expansion of first-generation biofuels is likely to exacerbate competition with food production within the tropics (Smith et al 2014, Correa et al 2019a and trigger biodiversity losses and CO 2 emissions as forests and savannas are replaced (Searchinger et al 2015, Elshout et al 2019. Biofuel production alternatives that avoid direct or indirect land-use changes in agricultural areas and biodiverse lands can increase the sustainability in biofuel production while achieving future targets in energy demands (Tilman et al 2009, Correa et al 2019a. Microalgal biofuel production systems (i.e. third-generation biofuels) constitute a more sustainable alternative to first-generation biofuels, as they do not require fertile soils, and are thus not expected to directly compete with food production (Schenk et al 2008, Correa et al 2017. Microalgal cultivation requires sun, water (fresh, brackish, or saltwater), and nutrients (Schenk et al 2008, Anto et al 2020. Cultivation systems consist on open ponds or enclosed culture devices (i.e. plates, tubes, bags, columns, or domes that are arranged to maximize sunlight uptake denominated photobioreactors) (Anto et al 2020) set over flat lands to reduce capital costs (Lundquist et al 2010). Open raceway ponds, in which algae and growth media are mixed by paddlewheels in closed recirculation channels with a depth between 15 and 30 cm (Schenk et al 2008, Anto et al 2020, are considered the most cost-effective cultivation system to date (Slade andBauen 2013, Fasaei et al 2018); nevertheless, microalgae can be initially grown in photobioreactors to avoid contamination from other microorganisms and then cultivated in open ponds (i.e. hybrid microalgal cultivation systems) (Schenk et al 2008, Acién et al 2017b, Dickinson et al 2017. Microalgal cultivation does not require rainfall, which would decrease competition with areas more suited for agricultural production or biodiversity conservation (Correa et al 2017(Correa et al , 2019b. Furthermore, microalgal production systems have lower freshwater footprint per unit of energy produced compared to food crops, particularly when recycling water from microalgal cultivation (Gerbens-Leenes 2018). Additionally, microalgae can be co-located with wastewater systems (Acién et al 2018, Jacob et al 2020 and CO 2 pollution sources (e.g. industries) (Razzak et al 2017, Collotta et al 2018, helping in water remediation and CO 2 uptake while production costs decrease (Judd et al 2017).
Yet, little is known about the most suitable areas for siting microalgal biofuels production within countries to achieve high production and profits while decreasing environmental impacts, particularly in the tropics (Sharma et al 2015). Furthermore, spatial analyses comparing microalgal biofuel production systems with major biofuel production alternatives, are missing to date. These analyses can reduce unnecessary environmental impacts resulting from unplanned biofuel expansion in tropical areas of the world (Williams et al 2019) and before microalgal biofuels are implemented at larger scales (Khan et al 2018, Correa et al 2019a, provided reductions in production costs compared to fossil fuels and firstgeneration biofuels (Acién et al 2017a, Chen et al 2018. The Neotropical region is considered promising to meet increasing global food demands (Laurance et al 2014), as a result of the presence of vast areas of undeveloped lands with high agricultural potential (Flachsbarth et al 2015). The Neotropics also harbors several of the most important areas for biodiversity conservation (Antonelli et al 2018, Rull 2020), containing several of the best-protected ecoregions with high conservation priority (Dinerstein et al 2017) and one of the last global wilderness areas globally (i.e. the Amazon rainforest) (Venter et al 2016, Watson et al 2018. Increased interest in replacing fossil fuels with low-cost and sustained production of energy has led to higher governmental support for biofuel production within Neotropical countries (Saravia-Matus et al 2018, Trindade et al 2019, even within countries with large fossil fuel reserves (e.g. Venezuela) (Ansari and Holz 2020). Nevertheless, poor spatial planning could intensify conflicts with agricultural lands, biodiversity, and carbon storage, driving further competition with food production, biodiversity losses, and increased CO 2 emissions if native ecosystems are replaced by biofuel crops in the region.
In the Neotropics, oil palm and sugarcane are considered the most cost-effective and productive crops for biodiesel and bioethanol production, respectively, and as a result, future biofuel demands have been projected to be met based on their expansion (Furumo and Aide 2017, Jaiswal et al 2017, de Andrade Junior et al 2019. In the present study, a comparative assessment of best areas for siting three different biofuel production alternatives (i.e. microalgae, oil palm, and sugarcane) to meet a 30% target in domestic transport energy demands by 2050, was developed within four Neotropical countries (Colombia, Ecuador, Panama, and Venezuela), aiming at minimizing direct impacts on high-value agricultural lands and biodiverse areas. These countries represent around 15% of transport energy demands in the Americas (after excluding Canada and the USA) and three of them are considered megadiverse in terms of total and endemic number of species (Mittermeier 1997). In contrast to Correa et al (2019b), biofuel feedstocks were assumed to be produced within each country's boundaries to guarantee national energy security targets without imports and favor the creation of local jobs (Fragkos and Paroussos 2018, Uría-Martínez et al 2018). Then, how each production alternative would compete with agricultural lands, biodiversity, and areas of high contents of aboveground biomass among countries, was analyzed. This novel spatially-explicit analysis can help in understanding how biofuel production alternatives compare to each other in the Neotropical region, in terms of impacts on food production, biodiversity, and carbon storage. The analysis can guide decision making towards the implementation of more sustainable biofuel production alternatives in a priority region for agricultural expansion and biodiversity conservation globally.

Development of optimization model for biofuel production alternatives
This study was developed in four densely populated Neotropical countries with high priority for agricultural expansion and biodiversity conservation (Colombia, Ecuador, Panama, and Venezuela). For each biofuel feedstock (microalgae, oil palm, and sugarcane) the best production areas to satisfy a 30% target of each country's future transport energy demands (i.e. 7.4, 3.9, 1.1, and 9.5 million tonnes of oil equivalent for Colombia, Ecuador, Panama, and Venezuela, respectively, by 2050), while avoiding areas of high agricultural and biodiversity value, were determined. The analysis modified the framework developed by Correa et al (2019b) to select best areas for microalgal biofuel production globally based on integer linear programming (Beyer et al 2016), aiming at maximizing microalgal biofuel profitability and minimizing the overlap with pixels of high agricultural and biodiversity value with a spatial resolution of 5 × 5 km. In the present study, the objective function was applied to each country assuming that biofuel feedstocks (i.e. microalgae, oil palm, and sugarcane) should be cultivated within each country's boundaries to fulfill 30% of their future domestic transport energy demands. This led to the development of eight models on best microalgal biofuel production areas based on two microalgal cultivation scenarios within each of the four countries, and the development of eight models on best oil palm and sugarcane biofuel production areas based on the cultivation of these two biofuel feedstocks within each country. The maximum percentage of each pixel that could be converted to microalgal cultivation was set to 80%, assuming that 20% must be reserved for associated infrastructure (Wigmosta et al 2011). For oil palm and sugarcane, which can be more densely produced, the maximum percentage of conversion was set to 90%. Following Correa et al (2019b) these percentages were set after excluding water bodies (Lehner and Döll 2004), protected areas (UNEP-WCMC 2016), Key Biodiversity Areas (KBA) (BirdLife International 2016), and urban areas (Schneider et al 2009).

Development of profitability model under climate change scenarios
For microalgae, two cultivation scenarios were considered. Scenario 1: Cultivation using fresh, brackish, or saltwater sources, and Scenario 2: Cultivation using seawater that does not compete with freshwater (assuming the cultivation of microalgal strains tolerant to a wide range of salinity conditions, thereby avoiding the use of freshwater to maintain salinity as water evaporates) (Ishika et al 2017). The profitability was estimated by overlaying water availability, microalgal lipid productivity, availability of flat lands, and proximity to main transport networks (i.e. main roads and railroads) (See Supplementary Information for details (stacks.iop.org/ERL/15/094094/ mmedia)). Future potential mean annual precipitation, mean annual potential evapotranspiration and mean annual temperature, which affect water availability and lipid productivities, were calculated for 2050, based on ensembles models for the climate change Representative Concentration Pathway (RCP) 8.5 (i.e. high emissions climate change scenario) (Riahi et al 2011). These ensemble models were constructed by averaging mean annual and monthly temperatures, minimum monthly temperatures, maximum monthly temperatures, and mean annual precipitation values among the General Circulation Models Oil palm and sugarcane profitability was estimated by overlaying water availability, agro-climatically attainable yield in dry weight by 2050 (IIASA/FAO 2012), the availability of flat lands, the proximity  . The agro-climatically attainable yields by 2050 were obtained from the Global Agro-Ecological Zones GAEZ database (IIASA/FAO 2012) by averaging the yields among the following climate change models for 2050: CCCma CGCM2 A2, CSIRO Mk2 A2, Hadley CM3 A2, and MPI ECHAM4 A2, which are representative of the high-emissions RCP 8.5 scenario used for future microalgal cultivation (Riahi et al 2011). It was assumed that freshwater can be readily obtained for crop cultivation and that a high level of intensification for crop production will be achieved.

Estimation of future transport energy demands
and overlapping with areas of high biodiversity, agricultural value, and soil carbon contents Future transport energy demands were calculated for each country based on current transport energy demands (IEA 2018) and an estimated annual increase in transport energy consumption at 2.5% for non-OECD Americas (EIA 2016). Land-covers potentially replaced by microalgae, oil palm, and sugarcane (Channan et al 2014) were estimated by overlapping them with proposed cultivation areas.
Non-parametric Dunn's tests with Bonferroni corrections (Dino 2017) were performed to detect test differences among biofuel production alternatives per country, considering agricultural and biodiversity value, and aboveground biomass. Potential conflicts among each biofuel production alternative, agricultural value, biodiversity value, and aboveground biomass, were mapped.

Results
To fulfill 30% of each country's transport energy demands by 2050 would require three times as much land (based on sugarcane) and twice as much land (based on oil palm) as microalgal systems (table  1). Best areas for microalgal biofuel production at the lowest direct competition with high-value agricultural lands and biodiversity would correspond to drier lowlands (figure 1). For Scenarios 1 (use of fresh, brackish, or saltwater) and 2 (use of seawater), main cultivation areas would be located in the Colombian Caribbean region (i.e. mostly in the Guajira desert) and inter-Andean valleys (i.e. mainly in the Alto Magdalena Valley), in the Ecuadorian coastal region (i.e. mainly in the provinces of Santa Elena, Manabí, and Guayas), in the Panamanian Pacific coast (i.e. mainly in the provinces of Veraguas and Coclé), and in the Venezuelan Caribbean coastline (i.e. around the Gulf of Venezuela in the state of Falcón and around the Lake Maracaibo in the state of Zulia). Based on these cultivation scenarios, estimated replaced land-covers would correspond to cropland/natural vegetation mosaics (24.9-28.5%), followed by savannas (19.5-20.5%), grasslands (10.5-10.8%), croplands (10.6-13.1%), open shrublands (7.7-7.9%), and woody savannas (7.1-7.3%) (table 1). Best areas for oil palm production at the lowest direct competition with high-value agricultural lands and biodiversity would be mainly located in humid areas within the Colombian Caribbean and Catatumbo regions and Middle Magdalena Valley, the Ecuadorian Pacific lowlands (i.e. mainly in the provinces of Esmeraldas, Pichincha, Manabí, Los Ríos, and Guayas), the Panamanian Pacific coast (i.e. mainly in the province of Chiriquí), and the Venezuelan northern lowlands (i.e. mainly the states of Zulia, Táchira, Monagas, and Delta Amacuro). Areas for sugarcane production at the lowest direct competition with high-value agricultural lands and biodiversity would be located in humid foothills, mainly in the Colombian Cauca and Magdalena valleys, the Ecuadorian Pacific lowlands (i.e. mainly in the provinces of Pichincha, Los Ríos, Manabí, and Guayas), the Panamanian Pacific coast (i.e. mainly in the provinces of Chiriquí and Veraguas), and the Venezuelan northern foothills (i.e. mainly in the states of Lara, Falcón, Guárico, and Sucre). Land-covers potentially replaced by oil palm and sugarcane include cropland/natural vegetation mosaics (35% and 35.6% for oil palm and sugarcane, respectively), followed by evergreen broadleaf forests (32.4% and 30.5% for oil palm and sugarcane, respectively), and savannas (19.9% and 18.3% for oil palm and sugarcane, respectively).
Microalgal cultivation would lead to lower competition with high-value agricultural lands in Ecuador (compared to sugarcane), Panama and Venezuela (compared to oil palm and sugarcane) (figure 2). Microalgal cultivation would consistently lead to lower competition with biodiversity and aboveground biomass (figures 3, 4). Higher targets in biofuel demands would increase competition with areas of higher agricultural and biodiversity value and larger aboveground biomass within countries (figures 5, 6, 7). 5  Boxplots showing statistically significant differences among biofuel production alternatives (microalgae, oil palm, sugarcane) considering the biodiversity (B) value of lands needed for fulfilling each country's 30% of transport energy demands by 2050. Letters represent homogeneous groups based on post hoc Dunn's tests with Bonferroni corrections. Asterisks represent the maximum and minimum outliers (defined after multiplying the interquartile range by 1.5). Microalgal cultivation scenario 1 (use of fresh, brackish, or saltwater sources), microalgal cultivation scenario 2 (use of seawater).

Discussion
Based on our analysis, microalgal biofuel production systems would require between 48% and 33% of the land area compared to oil palm and sugarcane, respectively, to fulfill 30% of transport energy demands by 2050. This arises because of the higher  to approach the production efficiency of microalgal production systems, with estimated maximum oil yields at 136 900 l ha −1 (Chisti 2007). Furthermore, an increase in microalgal productivity can be achieved through technological improvements that maximize biomass and lipid production, including the use of photobioreactors, nutrient starvation techniques (Leong et  Microalgal production systems would additionally offer a more adaptable alternative for maintaining biofuel productivity while coping with increasing temperatures and changes in rainfall patterns in the Neotropics. This is because many microalgal strains thrive in areas with higher mean annual temperatures (Moody et al 2014), microalgal cultivation does not depend on rainfall (particularly if cultivating seawater species) (Ishika et al 2017), and there are a wide range of microalgal strains that could be rapidly developed to cope with evolving cultivation conditions (Aslam et al 2017, Lim and Schenk 2017). Thus, increases in mean annual temperature and changes in precipitation patterns within the study region would not significantly alter best areas for siting microalgal production farms. In contrast, oil palm and sugarcane can be cultivated within a more limited temperature range and as rainfall is usually required for their production, the best areas for microalgal cultivation are expected to change with global warming. Unless new varieties tolerant to higher mean annual temperatures

Competition with high-value agricultural lands, biodiversity, and aboveground biomass among biofuel production systems
As in Correa et al (2019b) our analyses indicate that microalgal cultivation should be prioritized in dry lowlands (i.e. areas with median rainfall values equal than 547 and 498 mm for Scenarios 1 and 2 in the present study, respectively, which correspond to semi-arid lands) while oil palm and sugarcane would require humid areas (i.e. areas with median rainfall values equal to 1607 and 1382 mm for oil palm and sugarcane, respectively). In general, dry lands are considered less suited for agriculture (Li et al 2017, Taylor 2018, and additionally hold lower biodiversity values (Neves et al 2020) and lower aboveground biomass contents (Álvarez-Dávila et al 2017) than more humid areas. As a result, competition with food production, biodiversity, and aboveground biomass among countries would, in general, decrease if biofuel production is based on microalgae, compared to oil palm and sugarcane. Notwithstanding, dry ecosystems hold unique species (Brito et al 2014), provide valuable ecosystem services (e.g. tourism and desertification control) (Dudley et al 2014), and are considered at risk in the Neotropics (Etter Furthermore, less than 5.3% of microalgal production areas (i.e. around 1267 km 2 ) would occur within the highly biodiverse and carbon-rich rainforests, in contrast to oil palm and sugarcane that would occur within rainforests in between 30.5-32.4% of their estimated cultivation areas (i.e. 16 373 and 22 017 km 2 , respectively). Zoning the cultivation in areas of lower biodiversity value and lower carbon contents, such as pastures and , could reduce potential deforestation if expanding food crops for biofuel production. However, in doing so biofuel productivities per unit area would also be reduced (as pixels with lower potential productivities would need to be selected), thereby increasing the land footprint required to meet energy targets.
The use of seawater for microalgal cultivation could reduce competition with freshwater-provided that microalgal strains that tolerate a wide range of salinity conditions are cultivated (Ishika et al 2017)and become the best alternative in dry areas with compromised water security (Vorosmarty et al 2010). However, a trade-off among seawater availability, agricultural value, and biodiversity value, would Figure 6. Overlapping between profitability and biodiversity value (B value ranging from 0 to 1) for (a) microalgal cultivation scenario 1: use of fresh, brackish, or saltwater sources, (b) microalgal cultivation scenario 2: use of seawater, (c) oil palm, and (d) sugarcane. Red colors show areas with high potential profitabilities and high potential conflicts with biodiversity. Cultivation areas for fulfilling each country's 30% transport energy demands by 2050 are delineated in black. occur in the Colombian Caribbean region, where suitable areas for food production and with high biodiversity are located near the coasts. Reducing future targets in microalgal biofuel production (e.g. from 30% to 20% of total transport energy demands) would become a more sustainable option to limit potential direct land-use change of high-value lands for agriculture or biodiversity conservation in the Colombian Caribbean.
While microalgal cultivation would be the preferable alternative for biofuel production within the study region, the co-location of microalgal production systems with existing oil palm (Cheah et al 2018) and sugarcane crops (Santana et al 2017) has been recently explored as a mean to increase the overall sustainability and profitability of biofuel production. Oil palm and sugarcane industrial operations can offer CO 2 and nutrients that facilitate microalgal productivity, as well as electrical energy that can be used during microalgal cultivation and harvesting (Klein et al 2018).

Increases in transport energy demands
Following the success of Brazil in bioethanol production from sugarcane, several Neotropical countries aim to increase mandatory biofuel blends (Gonzalez-Salazar et al 2017, Cutz andNogueira 2018). Furthermore, high demands for biofuels in developed economies can promote biofuel's exports from Neotropical countries (Acharya and Perez-Pena 2020). Increasing targets in biofuel production, either for the domestic production of biofuel blends or for exports, would intensify conflicts with food production, biodiversity, and carbon storage in the region. Even for microalgal production systems, environmental impacts would increase as areas with lower Overlapping between profitability and aboveground biomass (AG biomass in tonnes ha −1 ) for (a) microalgal cultivation scenario 1: use of fresh, brackish, or saltwater sources, (b) microalgal cultivation scenario 2: use of seawater, (c) oil palm, and (d) sugarcane. Red show areas with high potential profitabilities and high potential conflicts with carbon storage. Cultivation areas for fulfilling each country's 30% transport energy demands by 2050 are delineated in black. productivity per unit area, or with higher agricultural, biodiversity, and aboveground biomass values are considered suitable to reach higher biofuel production targets. Limiting targets in biofuel production could thus reduce environmental impacts in the Neotropical region.
Encouraging biofuel production outside priority areas for biodiversity and carbon storage (e.g. tropical rainforests) can reduce the overall environmental impacts of biofuel production in the Neotropical region. This could be achieved through the implementation of policies and frameworks (e.g. sustainability certification schemes) that prevent deforestation and promote sustainable biofuel production practices (De Man and German 2017). The implementation of microalgal production systems can derive from the development of taxation schemes that discourage less sustainable alternatives while subsidies and technological improvements increase their profitability

Conclusions
Even in countries with large fossil fuel reserves (e.g. Venezuela), biofuel production is projected to increase to fulfill future transport demands, decreasing the uncertainties associated with future fossil fuel markets (Ansari and Holz 2020) while climate change is mitigated (Binsted et al 2019). Microalgal production systems offer the most sustainable alternative for future biofuel production within the Neotropical region. Compared to oil palm and sugarcane, microalgal production systems can be implemented in areas with lower value for food production, biodiversity, and carbon storage, as they can overlap drier lowlands within countries. Furthermore, to fulfill 30% of each country's transport energy demands, less than 5.3% of their estimated cultivation areas would overlap rainforests, in contrast to oil palm and sugarcane, which would overlap rainforests in at least 30% of their estimated cultivation areas. Additionally, microalgal production systems could reduce land area requirements by at least 52% compared to oil palm and 67% compared to sugarcane. Decreasing targets in biofuel production (either for domestic production or for exports) would further limit potential environmental impacts associated with the expansion of microalgal biofuels.