Feasibility of biodiesel production and CO2 emission reduction by Monoraphidium dybowskii LB50 under semi-continuous culture with open raceway ponds in the desert area

Background Compared with other general energy crops, microalgae are more compatible with desert conditions. In addition, microalgae cultivated in desert regions can be used to develop biodiesel. Therefore, screening oil-rich microalgae, and researching the algae growth, CO2 fixation and oil yield in desert areas not only effectively utilize the idle desertification lands and other resources, but also reduce CO2 emission. Results Monoraphidium dybowskii LB50 can be efficiently cultured in the desert area using light resources, and lipid yield can be effectively improved using two-stage induction and semi-continuous culture modes in open raceway ponds (ORPs). Lipid content (LC) and lipid productivity (LP) were increased by 20% under two-stage industrial salt induction, whereas biomass productivity (BP) increased by 80% to enhance LP under semi-continuous mode in 5 m2 ORPs. After 3 years of operation, M. dybowskii LB50 was successfully and stably cultivated under semi-continuous mode for a month during five cycles of repeated culture in a 200 m2 ORP in the desert area. This culture mode reduced the supply of the original species. The BP and CO2 fixation rate were maintained at 18 and 33 g m−2 day−1, respectively. Moreover, LC decreased only during the fifth cycle of repeated culture. Evaporation occurred at 0.9–1.8 L m−2 day−1, which corresponded to 6.5–13% of evaporation loss rate. Semi-continuous and two-stage salt induction culture modes can reduce energy consumption and increase energy balance through the energy consumption analysis of life cycle. Conclusion This study demonstrates the feasibility of combining biodiesel production and CO2 fixation using microalgae grown as feedstock under culture modes with ORPs by using the resources in the desert area. The understanding of evaporation loss and the sustainability of semi-continuous culture render this approach practically viable. The novel strategy may be a promising alternative to existing technology for CO2 emission reduction and biofuel production. Electronic supplementary material The online version of this article (10.1186/s13068-018-1068-1) contains supplementary material, which is available to authorized users.


Background
Renewable and environmentally friendly alternative fuels are urgently needed for future industrial development, because of the diminishing world oil reserves and the environmental deterioration associated with fossil fuel consumption [1,2]. Microalgae are increasingly considered as feedstock for next-generation biofuel production because of their many excellent characteristics, such as broad environmental adaptability, short growth period, high photosynthetic efficiency, and high-quality lipid [3,4]. However, the commercial feasibility of microalgal biodiesel is limited because only few microalgal strains can be grown reliably with high lipid content (LC) outdoors. Lipid productivity (LP) under outdoor conditions is significantly lower than that in the laboratory due to pollution from other microorganisms and fluctuations in environmental parameters [5][6][7]. Large-scale outdoor cultivation using sunlight is the only solution for the sustainable industrial production of microalgal biofuel [8]. Therefore, an essential prerequisite to achieve the industrial-scale application of microalgal biofuel is the selection of robust and highly productive microalgal strains with relatively high LC outdoors.
Two cultivation systems are commonly used for largescale outdoor microalgal cultivation: the open system (e.g., open raceway ponds, ORPs) and the closed photobioreactor system (e.g., tubular, flat plate or column photobioreactors) [1,[9][10][11]. Compared with closed photobioreactors, ORPs consume less energy and require lower investment and production costs for microalgal cultivation [12]. Although microalgal cultivation in ORPs offers many advantages, the high cost of cultivation systems impedes the commercialization for lipid production. Thus, developing an economically feasible culture mode to increase the lipid production and thereby reduce cultivation costs is necessary [7,13].
Increasing LP via culture modes can reduce the costs and enhance the economic feasibility of microalgal biodiesel production. The photoautotrophic two-stage cultivation mode is a highly promising approach to increase lipid production in photobioreactors by improving LCs [14][15][16]. However, only few studies have employed this mode in ORPs [17]. The semi-continuous mode is a simple and efficient strategy to increase lipid production in microalgal biomass by continuously increasing biomass [7,18]. This mode can avoid a low cell division rate at the early exponential stage and light limitation at the late stationary stage. Furthermore, it maintains the microalgal culture under exponential growth conditions, resulting in enhanced biodiesel production [7]. However, the number of cycles for semi-continuous culture is limited because of the different nutrient consumption rates of algal cells. Thus, exploring the adequate cultivation time of microalgal cells under whole semi-continuous cultivation mode is important to evaluate their survivability.
The large-scale cultivation of microalgae requires large areas of land and water resources. Arid and semiarid regions account for 41% of the global land area [19]. Thus, cultivating microalgae in desertification areas avoids competition with food crops for arable land and water. In addition, the unique climatic conditions (strong solar radiation, long sunshine duration, and large day and night temperature difference) in deserts are beneficial to the accumulation of dry weight (DW) in the cells. Compared with other crops, microalgae are more compatible with desert conditions. Furthermore, cyanobacteria and green microalgae can be stably and efficiently cultivated in desert areas [13,20,21]. Therefore, microalgal cultivation is an effective means to utilize desert lands and sunshine.
In the current work, the ability of three microalgae to produce high lipid indoors was determined in a 5 m 2 (1000 L) ORP to select for high environmental adaptability and lipid accumulation capability in the desert area. The influences of two-stage cultivation mode and semi-continuous mode on cell growth, CO 2 fixation rate, and evaporation rate were first investigated under 5 m 2 (1000 L) ORP. Algal strain growth was scaled up to a 200 m 2 (40,000 L) ORP with semi-continuous mode to determine cycle times. Outdoor cultivation test at different times was conducted to assess the stability of the algal strain in long-term semi-continuous operations. Finally, the energy consumption of life cycle was analyzed to assess the feasibility of biodiesel production and CO 2 mitigation in desert area.

Organism
Monoraphidium dybowskii LB50 and Micractinium sp. XJ-2 were provided by Prof. Xudong Xu of the Institute of Hydrobiology, the Chinese Academy of Sciences. Podohedriella falcata XJ-176 was isolated from Xinjiang Taxi River Reservoir (Additional file 1: Figure S1). The stock cultures were maintained indoors in a sterilized BG11 medium containing 1.5 g NaNO 3 Figure S1). Two scales of ORPs at 5 and 200 m 2 were utilized. The length, width, and maximum depth were 4.80, 1.05, and 0.60 m and 34.50, 5.80, and 0.60 m in 5 and 200 m 2 illuminated areas of ORP, respectively (Additional file 1: Figure S1). The culture depth in raceway ponds was set to 20 cm, with 1000 and 40,000 L culture volumes. A stainless steel paddlewheel, 0.80 m in diameter, was used for the circulation of the cultures in 5 and 200 m 2 ORPs at 0.35 and 0.25 m s −1 , respectively. Microalgae were cultivated using a modified BG11 medium containing 0.25 g L −1 urea, but 0.1 M NaHCO 3 was added to the medium used for M. dybowskii LB50. The medium was thoroughly compounded with groundwater. A series of scale-up pre-cultivation was employed (Additional file 1: Figure S1). Water in the system was replenished every day to prevent serious evaporative losses in the open raceway system. Cell concentration measured as an OD 680 of 0.1 was inoculated into the culture in 5 and 200 m 2 ORPs.
After pre-cultivation, the batch culture was conducted with three microalgae in 5 m 2 ORP (1000 L) to select the optimal stain for lipid production.
For two-stage salt induction culture in 5 m 2 ORPs, M. dybowskii LB50 was cultivated in 5 m 2 ORPs outdoors. On the 10th day, which is at the late-exponential growth phase, NaCl and industrial salts (Hubei Guangyan Lantioan salt chemical co., Ltd, China. Additional file 2: Table S1) were added at final concentrations of 0 and 20 g L −1 . Industrial salts, often referred in China to NaCl, NaOH (caustic soda), and Na 2 CO 3 are widely used in the industry. In the current study, the main component of industrial salt was NaCl. Industrial salt can be inexpensive and is easily produced because of the low purity. Day 0 was assumed as the time of salt addition.
For semi-continuous cultivation, further experiments were conducted with semi-continuous mode in two ORP scales. Two-thirds of the culture was harvested, and the remaining culture was used as the seed for subsequent batches and replaced by the same volume of nutritionrich growth media containing half of the urea concentration. The algal culture was harvested every 3 or 4 days. The semi-continuous experiment was carried out in a 200 m 2 ORP for a month.
The water used for algal cultivation was pumped from the ground and contained 89.39 ppm Na + , 62.92 ppm SO 4

Biomass measurement
Biomass productivity (BP, mg L −1 day −1 ) was calculated according to Eq. (1): where B2 and B1 represent the DW biomass density at time t (days) and at the start of the experiment, respectively. Algal density was determined by measuring the OD 680 -the optical density of algae at 680 nm. The relationships between the DW (g L −1 ) and the OD 680 values of the algae were described using Eqs. (2)(3)(4): The cells were harvested by centrifugation and baked in an oven.

Lipid analysis
Total lipid was extracted from approximately 80-100 mg of the dried algae (w 1 ) using a Soxhlet apparatus, with chloroform-methanol (1:2, v/v) as the solvent. Total lipid was transferred into a pre-weighed beaker (w 2 ) and blowdried in a fume cupboard. The lipid was dried to a constant weight in an oven at 10 °C and weighed (w 3 ). LC (%) and the LP (mg L −1 day −1 ) were determined according to Eqs. (5,6):

Determination of urea concentration
Urea concentration was determined following the protocol outlined by Beale and Croft [22]. The liquid sample collected from the raceway pond was filtered using a 0.22 μm-pore filter and then diluted 60-fold with deionized water for each sample. The sample was collected and mixed with 1 volume of diacetylmonoxime-phenylanthranilic acid reagent (1 volume of 1% w/v diacetylmonoxime in 0.02% acetic acid and 1 volume of phenylanthranilic acid in 20% v/v ethanol with 120 mM NaCO 3 ). Exactly, 1 mL of activated acid phosphate (1.3 M NaH 2 PO 4 , 10 mM MnCl 2 , 0.4 mM NaNO 3 , 0.2 M HCl in 31% v/v H 2 SO 4 ) was added before incubation in boiling water for 15 min. The tubes were left to cool, and their OD 520 were determined using a UV/Vis spectrophotometer.

Determination of pH, irradiance, conductivity, and evaporation
The temperature, conductivity and pH of the culture medium were determined daily by utilizing respective sampling probes (YSI Instruments, Yellow Springs, Ohio, USA). Irradiance was measured with a luxmeter (Hansatech Instruments, Norfolk, UK). The depth at four fixed positions was determined in the raceway ponds every day, and evaporation (L m −2 day −1 ) was calculated according to Eq. (7): where h2 and h1 represent the average depth at time t (days) and at the start of the experiment, respectively. S represents the area of the raceway ponds.

Determination of CO 2 fixation rate
According to the mass balance of microalgae, the fixation rate of CO 2 (mg L −1 day −1 , g m −2 day −1 ) was calculated from the relationship between the carbon content and volumetric growth rate of the microalgal cell, as indicated in Eq. (8): where BP is in mg L −1 day −1 or g m −2 day −1 ; C carbon is the carbon content of the biomass (g g −1 ), as determined by an elemental analyzer (Elementar Vario EL cube); M CO 2 is the molar mass of CO 2 ; and M C is the molar mass of carbon (Additional file 3: Table S2).

Net energy ratio (NER) and energy balances
NER is defined as the ratio of the energy produced over primary energy input as represented in Eq. (9): On the basis of the data obtained in the 200 m 2 ORP for cultivating M. dybowskii LB50 for 1 year, NER is estimated using the method discussed by Jorquera et al. [23].
Energy balance is defined as the difference between energy produced and primary energy input, as represented Eq. (10): Energy balance = Energy produced (lipid or biomass) − Energy requirements.

Statistical analysis
The values were expressed as mean ± standard deviation. The data were analyzed by one-way ANOVA using SPSS (version 19.0). Statistically significant difference was considered at p < 0.05.

Results and discussion
Growth, lipid accumulation, and CO 2 fixation rate of the three microalgae in 5 m 2 ORPs outdoors Three strains of potential microalgae (Additional file 4: Table S3) were grown in 5 m 2 ORPS to evaluate their lipid accumulation and CO 2 fixation potential. As shown in Fig. 1  respectively. During the time course of culture, CO 2 fixation rate was low at the beginning and stable stage and was the highest at the exponential growth stage, reaching 163 mg L −1 day −1 . At the late growth stage, CO 2 fixation rate was negative, indicating that the microalgal cells did not grow or died, releasing large amounts of CO 2 possibly through respiratory metabolism.
Microbial contamination during large-scale algal cultivation can significantly and consistently reduce biomass production. In this context, eukaryotic contaminants, such as amoebae, ciliates, and rotifers, and clusters of cells based on microscopy were found to cause biomass deterioration in P. falcata XJ-176 cultivation. In the current study, this phenomenon was rarely observed during the cultivation of M. dybowskii LB50 and Micractinium sp. XJ-2. These results showed that the two species demonstrate high environmental tolerance, especially to the high light intensity in the desert (Additional file 6: Figure S2), and could inhibit the excessive growth of bacteria [16,24]. Consequently, M. dybowskii LB50 exhibited improved lipid accumulation potential outdoors, particularly during cultivation in the desert.

Two-stage induction culture of microalgae
In addition to selecting a fast-growing strain with high LC, improving the LC or biomass to increase lipid yield is also necessary to enhance the economic feasibility of microalgae-based CO 2 removal and biodiesel production [13,16,25]. LC can be improved through many ways [7,26], among which two-stage salt induction is very effective [27]. In our previous study, the LC of M. dybowskii LB50 was increased by 10% through NaCl induction in 140 L photobioreactors outdoors [16]. However, few studies on NaCl induction in ORPs have been conducted [17]. Figure 2 shows that the biomass was not significantly decreased on the first day of NaCl and industrial salt induction (p > 0.05), but was significantly reduced on the third day (p < 0.05). The effect of industrial salt induction on LC was similar to that of NaCl. LC increased by 7% on day 1 of induction and by 10% on day 2 of induction. Thus, LP was 3.3 g m −2 day −1 without significant difference within 1 or 2 days of induction. Only 1 day was required for induction to shorten the culture period. Meanwhile, CO 2 fixation rate was 78 mg L −1 day −1 at the time course of induction ( Table 1). The pH of the culture liquid did not significantly change, after adding NaCl or industrial salt, but the conductivity increased by five times after adding salt ions (Additional file 2: Table S1). Consequently, the two-stage industrial salt induction culture mode in ORPs favorably increased the LC and reduced the costs.
Two-stage cultivation has been performed in closed photobioreactors outdoors. Tetraselmis sp. and Chlorella sp. were cultured in 120 L closed photobioreactors, and lipid productivities of microalgae were increased by suitable CO 2 concentration [11,28]. Moreover, NaCl induction in the column photobioreactors was favorable [16]. However, these reports have not been verified in ORPs. Kelley [29] reported LC can be increased by using a twostep method involving N deficiency and light conversion in 3 m 2 ORPs. LP can also be increased by NaCl induction during dual mode cultivation of mixotrophic microalga in culture tubes [17]. In this study, we confirmed that LC was significantly increased not only in the open runway pool (1000 L), but also with industrial salt induction.

Semi-continuous culture in 5 m 2 ORPs
Given its convenient operation and cost-effectiveness, semi-continuous cultivation is also a good choice [30]. Semi-continuous cultivation has attracted considerable attention in energy microalgae [7,18,31]. Unfortunately, the culture medium used in semi-continuous cultivation cannot be reused for an unlimited number of times because of the difference in nutrients consumption rate of cells. Portions of the nutrient concentration excessively increase with culture time and eventually inhibit cell growth.
In the 1000 L ORP, the BP increased from 44.86 to 74.16 mg L −1 day −1 after repeated culture, and the LC remained stable at 30% in M. dybowskii LB50 (Fig. 3). Finally, areal LP (ALP) increased from 2.73 to 4.58 g m −2 day −1 (Table 2), and the CO 2 fixation rate increased from 16.1 to 26.7 g m −2 day −1 after repeated culture. During the whole semi-continuous culture, the CO 2 fixation rate reached 23 g m −2 day −1 (114 mg L −1 day −1 ). The pH of the culture medium did not significantly change (9.14-9.52, Fig. 3c), indicating that the growth consistently improved throughout the semi-continuous culture. However, the fluctuations in light intensity and temperature were large. Increased illumination and prolonged periods of light exposure were favorable factors for microalgal culture in desert areas, but high evaporation due to increased illumination was unfavorable. Evaporation occurred at 1.62 L m −2 day −1 (Fig. 3). The minimum amount of evaporation was 0.68 L m −2 day −1 at low temperature and light intensity (day 6, rainy day), whereas the highest evaporation rate was 2.26 L m −2 day −1 at high temperature and light intensity in the 5 m 2 ORP.
The two-stage induction culture exhibited slightly higher LP than the semi-continuous culture in the same culture time in a 5 m 2 ORP. However, the semi-continuous culture was more favorable for CO 2 emission reduction than the two-stage induction culture. The semi-continuous culture prolonged culture period to reduce the supply of the original species. Figure 4 shows the semi-continuous culture of M. dybowskii LB50 in a 200 m 2 ORP (40,000 L) for a month. BP was 15.2 g m −2 day −1 during the initial growth (0-7 days). The highest BP was 26.8 g m −2 day −1 during the first cycle of semi-continuous culture, but was decreased at the second cycle, because of the rainy days (11-12 days, Fig. 4c). The average biomass productivity  Table 3) after 1 month of semi-continuous culture at five cycles of replacement. The LC did not significantly change during the four cycles, but significantly decreased at the fifth passage. Therefore, the LP also decreased during fifth passage. The change in CO 2 fixation rate was the same as that during biomass production. The average CO 2 fixation rate was 30.8 or 33.9 g m −2 day −1 at 0-26 or 0-20 days (Table 3).

Scaled up semi-continuous cultivation in 200 m 2 ORP
Evaporation occurred at 0.88 ± 0.31 L m −2 day −1 in the 200 m 2 ORP, and the maximal evaporation rate was 1.44 L m −2 day −1 under high light intensity (128-1568 μmol m −2 s −1 ). Even during a rainy day, minimal evaporation loss of 0.39 m −2 day −1 , which included the leakages and washout of the ORP, was found. Therefore, the average daily evaporation loss rate was 0.44%, and evaporation loss rate was 8.8-11.44% during the whole semi-continuous culture. Figure 4d shows that a small  amount of urea can accumulate after each cycle of replacement. The accumulation of urea in the medium reached 0.05 g L −1 until the fourth cycle of semi-continuous culture. These results suggested that the growth and lipid of cells were affected by the accumulation of partial nutrients and the remaining death cells in the media as cycle times increased. Therefore, five cycles of repeated culture were conducted in this study. However, further scalable work can be continued for long-term cultivation with additional repeated times,   (Fig. 5). M. dybowskii LB50 could exhibit stable growth for a month with semicontinuous culture. The biomass and LC were maintained at 18-20 g m −2 day −1 and 30%, respectively. The CO 2 fixation rate remained at 33 g m −2 day −1 , but the evaporation exhibited increased difference in various months. The evaporation rates were 0.39-1.44 L m −2 day −1 ( x = 0.9 L m −2 day −1 ), 0.56-3.29 L m −2 day −1 ( x = 1.6 L m −2 day −1 ), and 0.74-3.72 L m −2 day −1 ( x = 1.8 L m −2 day −1 ) in September 2014, July 2015, and August 2016. The evaporation loss rate of a semicontinuous culture is 6.5-13%. Water resources are a potential limitation for microalgal culture, but evaporation affects its scale and sustainability [32]. Furthermore, regions with high BP receive high solar irradiance and thus result in high evaporation rates [33]. Evaporation of the ponds was assumed to occur at a rate of 0.4 cm day −1 (0.4 L m −2 day −1 ) [34]. In this case, further work on the water cyclic utilization and evaporation reduction can be conducted for sustainable cultivation because of the increased evaporation.
Replacement ratio or dilution ratio, the volume ratio of new medium to total culture, is an important parameter in semi-continuous culture because it influences microalgae growth and cell the biochemical components. Ho et al. [35] reported that BP increases with replacement ratio, but lipid causes the opposite effect. The 90% replacement group exhibited the highest overall LP among five replacement ratios (10, 30, 50, 70, and 90%). Some studies reported that a semi-batch process with a 50% medium replacement ratio is suitable for microalgal biomass production and CO 2 fixation [13,36]. In the current study, the LC was unaffected by the 2/3 replacement ratio mainly because of the high and long duration of light in the desert. Although the microalgal concentration in the reactor was not high, cells could grow rapidly.
Cycle time is another parameter affecting the continuity of semi-continuous culture. Previously, five to six cycles of repeated semi-continuous culture were conducted and resulted in inhibited growth or decreased LC [35,37]. The LC of Desmodesmus sp. F2 significantly decreased at the sixth repeated cycle when five replacement ratios were adopted for semi-continuous cultivation for six repeated cycles [35]. In the 2/3 replacement test, the LC remained high throughout the five-cycle repeated course in 200 m 2 ORPs. Table 4 shows the LP of microalgae in large-scale culture outdoors. The largest scale was implemented in the cultivation of N. salina in the USA, and LP was 10.7 m 3 ha −1 year −1 [38], followed by the cultivation of M. dybowskii LB50, Graesiella sp. WBG-1, and M. dybowskii Y2 in 200 m 2 ORPs (40,000 L). The LPs (5.3 g m −2 day −1 ) of M. dybowskii LB50 and M. dybowskii Y2 were higher than those of Graesiella sp. WBG-1 (2.9 g m −2 day −1 ) and the others in ORPs and tubular photobioreactors. Increased CO 2 fixation ability (CO 2 fixation rate of 34 g m −2 day −1 ) was obtained under semicontinuous modes with ORPs in the desert area (Table 4). These results indicated that high biomass production was obtained and CO 2 mitigation was feasible by microalgal culture in the desert. The volumetric LP (VLP) in ORPs was lower than that in photobioreactors (Table 4). Finally, all types of bioreactors must focus on the ALP in microalgae industry applications. In brief, the semi-continuous mode in ORPs is more practical than other operation modes in other bioreactors for long-term cultivation. Thus, it is suitable for oleaginous microalgae industry applications because it is economic, convenient, and demonstrates high ALP.

Energy consumption evaluation of outdoor cultivation in different culture modes
The biodiesel production from microalgae involved a course of cultivation, centrifugation, drying, and extraction via a conventional method. We assumed that 100,000 kg dry weight of biomass was produced within  the year (270 days). Other parameters were included in our assessment according to the actual operation. Table 5 shows that the net positive energy for oil production (1.34-2.72) and biomass production (1. 41-2.52) in the two-stage salt induction or semi-continuous culture mode was higher than those in the batch mode in 5 m 2 ORPs. Moreover, in the 200 m 2 ORP, the net positive energy of oil production in the semi-continuous and batch modes was 1.52-2.69, indicating that the semicontinuous culture increased the biomass yield, but not the additional energy consumption. The NER of oil and biomass production increased with a scale-up of the culture system. In addition, the energy demand for producing 1 kg of biodiesel was 14.2-23.3 MJ under semicontinuous mode in 200 m 2 ORP. Figure 6 shows that the energy consumption of cultivation assumed the highest proportion (55-72%) under any culture mode. The energy balance in the two-stage salt induction culture mode was higher than that in the other methods mainly due to the increase of LC by industrial salt induction to increase the energy produced by oil. The energy produced by oil was 1.27 times larger than that under other modes within the same biomass production (100,000 kg), but the energy balance was only about 10% higher than that under semi-continuous mode. These results demonstrate that the energy consumption of the cultivation process was increased and was reduced by scaling up. The energy balance thus increased after scaling up. Moreover, the energy balance under semi-continuous mode was five times higher than that under batch mode in 5 m 2 ORPs and was 1.15 times higher in 200 m 2 ORP. Therefore, reducing energy consumption by intermittent agitation or by optimizing mixing, mixing velocity, and paddlewheel must be prioritized to reduce the energy consumption of the entire industrial chain [39].

Table 5 Comparative energy analyses for biomass or bio-oil production based on 1 year of cultivating M. dybowskii LB50 via different culture modes under OPRs
The assumed annual biomass production is 100,000 kg a Data were based on this study b Determined by dividing the illuminated area actual by production the volume of each unit c 3.72 W m −3 from Jorquera et al. [23]. 12.5 W m −3 from the actual date for the 200 m 2 raceway pond d Includes 8 h of daily pumping e Stepan et al. [52]. 539 kWh ton −1 biomass f Stephenson et al. [53]; Gao et al. [54]. 345.34 kWh ton −1 biomass g Energy content of net oil yield (assumed value of 39.04 MJ kg −1 ); Jorquera et al. [23] h Energy content of net biomass yield (assumed value of 31.55 MJ kg −1 ); Jorquera et al. [23] i NER would be above 1 if including coproduct allocation [55]  NER is associated with the type of culture system, and the NER of oil is generally less than 1 in tubular photobioreactors and greater than 1 in ORPs [23]. Ponnusamy et al. [40] reported that the energy demand for producing 1 kg of biodiesel is 28.23 MJ. Only 14-23 MJ was required for 1 kg of biodiesel in this study, which significantly decreased the energy consumption. He et al. [13] reported that the semi-continuous mode reduces the total costs (14.18 and 13.31$ gal −1 ) by 14.27 and 36.62% compared with the costs of batch mode in M. dybowskii Y2 and Chlorella sp. L1 in the desert area. Therefore, using semi-continuous culture mode with ORPs in the desert area can result in higher biomass, lower energy consumption, and lower costs compared with other culture modes.

Conclusion
Three microalgae were investigated for their environmental tolerances and lipid production potential in ORP outdoors, and M. dybowskii LB50 can be efficiently cultivated using resources in the desert. Lipid production can be improved by using two-stage salt induction and semi-continuous culture modes in ORPs. After 3 years of operation, M. dybowskii LB50 was successfully and stably cultivated under semi-continuous mode for a month (five cycles of repeated culture) in 200 m 2 ORPs in the desert, reducing the supply of the original species. The BP and CO 2 fixation rates were maintained at 18 and 33 g m −2 day −1 , respectively. The LC decreased only during the fifth cycle of repeated culture. Evaporation occurred at 0.9-1.8 L m −2 day −1 (6.5-13% of evaporation loss rate). Finally, using the semi-continuous and two-stage salt induction modes for cultivating M. dybowskii, LB50 can reduce energy consumption and increase energy balance via energy analysis of life cycle. Therefore, M. dybowskii LB50 is a promising candidate for the large-scale, outdoor production of biodiesel feedstock in desert areas. The outdoor ORP cultivation system together with the semi-continuous culture method in desert areas is a suitable strategy to further decrease the cultivation cost and increase the biomass/oil production and CO 2 emission potential of M. dybowskii LB50.