The effect of non-terrestrial versus terrestrial like soil on growth of wild-occurring, nitrogen-fixing and crop plants, in relation to Mars and Moon colonization.
The possibility of human space colonisation raises the importance of food availability and thereby the possibility of agriculture on non-terrestrial soil. Some of the challenges regarding plant growth in space concern the nutrient availability and its influence on the growth of the plants. The effect of Moon and Mars soil simulants on different plant types was investigated by growing 14 different plant species (either crop, nitrogen-fixers or wild type) in 840 pods for 50 days under same conditions. The seed germination number, as well as total biomass, was recorded and investigated. The results indicated an effect of the soil type on the plant´s germination (Kruskal-Wallis test p-value <0.001) and biomass (two-way ANOVA test p-value <0.003), with crops performing best of all of the plants and Martian soil giving highest yield. Therefore, additional investigation of crops alone was performed which showed that overall plants germinate best in Martian soil and S. cereale gives highest biomass yields of all crop species investigated (two-way ANOVA test p-value <0.001). However, further research is needed in the effect of non-terrestrial soil on plant yield.
In recent years space exploration has been studying the solar system in hope to find life and/or suitable conditions for human colonization. Agriculture is a major issue in this topic to create a self-sustainable station (Kozyrovska et al., 2006; Finetto, Lobascio and Rapisarda, 2010).
Previous studies performed that investigated the effect of lunar soil on plant growth did not show any harmful side effects when the plants were exposed to some lunar materials gathered during Apollo 11 and 12 missions (Ferl and Paul, 2010). However, this did not test the plant´s growth in lunar simulated soil which could have affected the findings. An important factor could be the presence of nitrogen as it affects both plant growth as well as the environment surrounding it by affecting the pH (Stevens et al., 2011). This could be potentially solved through the growth of nitrogen-fixing plants as well as crop plants (Wamelink et al., 2014). However, the investigations of both lunar and Martian soil samples did suggest presence of carbon as well as reactive nitrogen, which are of major importance for plant growth (Palomba et al., 2009; Ferl and Paul, 2010; Stevens et al., 2011). Additionally, studies performed on plants grown in micro and hypo gravity have shown that it is possible to culture plant species in those conditions (Maggi, 2010; Kamal et al., 2018). Therefore, the remaining question is whether the suitable nutrients in the soil from Moon and Mars are available to be taken up by plants and whether there are substances such as aluminium in the soil that would have a negative impact on plant growth.
This study aims to investigate the possibility to grow plants on the two celestial objects closest to Earth; Moon and Mars using the available soil there. The focus of this study was to see which of the plant types would perform best when it comes to germination and the total biomass of the plant. The independent variables were the plant type and species as well as the soil simulants. The dependent variable was the seed germination and the total biomass recorded. During the study it was assumed that the atmospheric conditions would be same as on Earth, to make the station suitable for human use. The gravity difference was not taken into consideration in this study.
H0 = The different soil simulants do not affect any of the plant´s germination and total biomass when recorded under controlled conditions.
The data used were collected by Wamelink et al., 2014.
2.1 Soil preparation
The moon and Mars soil have been mimicked using earth volcanic soil as their regoliths are strongly similar to the volcanic soil (Rickman, McLemore and Fikes, 2007; Carlton et al., 2014). As the control nutrient-poor coarse river Rhine sand was used. It was taken from 10 meters below the sediment as it is nutrient-poor and contains no organic matter. The earth, moon and mars simulants were tested for nutrients, pH, water and organic nutrients such as nitrogen and phosphorus as well as organic matter and other minerals listed in Table 1. The soils prepared were not sterilised as this could have caused a change to the properties previously investigated.
Table 1. The results of soil samples analysis of earth, moon and mars-like soil (Wamelink et al., 2014).
2.2 Species ascertainment
For this experiment three plant types were chosen; crops, nitrogen fixers and naturally occurring in Netherlands wild plants as shown in Table 2 (Wamelink et al., 2005). The species had to have small seeds for the purpose of this experiment to enhance the dependence on the nutrients available in the soil, rather the ones from the nutrient stock within the seed. The seeds were also not sterilised to not affect the seed structure and therefore, bacteria could be present on the seeds. The nitrogen fixers and crop seeds were bough at Welkoop, Wageningen whereas the wild species were bought from Cruydt Hoeck (Nijeberkoop) (Wamelink et al., 2014).
Table 2. Detailed information on the plant species used for the experiment. The table contains its abbreviation as well as its Latin and English name, the group it was assorted to and a description of the species soil preferences (Wamelink et al., 2005, 2011, 2012, 2014).
2.3 Experimental design and data collection
Twenty replications of same conditions (same soil and species) were performed resulting in 840 pods (20 replicates of 3 soil types and 14 species types). Each of the pods contained five seeds and same volume of soil, for a similar height, of one of the three soil simulants (100g of either Moon, Earth or 50g of Mars soil as maars soil was heavier). This resulted in 100 seeds being used for each species. Additionally, 25 grams of water was added that was previously demineralized (to simulate water that would be available on Mars or Moon) and could not lead as the pods did not have a cap. This also prevented roots to mix outside of the pod.
The pods were placed randomly (see Figure 1) under a greenhouse (80 µmol of strength) that was switched on if the sunshine gave below 150 watt/m2 of light, to give a similar light availability for each pod. The temperature was kept at 20-22°C the entire duration of the experiment, which started on 8th of April 2013 (mean day time was 16 hours) and continued for 50 days. The average humidity recorded every 24 hours was 65.0±15.5%.
The number of germinated seeds was recorded during the duration of the experiment. After 50 days the total biomass was also recorded by cleaning, harvesting and drying the plant in a stove for 24 hours at the temperature of 70°C. 25 of the plants resulted in total mass below the weighting limit of the apparatus and was therefore assigned a value of 0.5mg of the plant germinated and did not recover or 0.1 for plants that germinated but died directly afterwards.
Figure 1. The experimental design of the pods. The plants were randomly assorted to each of the blocks with 42 pods in each block. Each pod had a label to differentiate the conditions and the plant species grown (Wamelink et al., 2014).
2.4 Statistical analysis
Prior to the statistical analysis, the data were investigated for the presence of outliers that were removed as the residuals with outliers did not show a normal distribution. Test for equal variance was met for both parts of the experiment when investigating the processed raw data.
The data collected for the number of germinated plants were discrete quantitative as well as not normally distributed (p-value <0.005 in probability plot) and, therefore, the nonparametric test Kruskal-Wallis test was performed for the number of germinated plants versus the plant type as well as versus the soil type.
The biomass raw data were processed using the square root, to fit the normal distribution more precisely and thereby be able to perform a parametric two-way ANOVA test. This was crucial as it was important to compare the interaction between soil and plant type. When the raw data for all of the species types were processed using the square root, the histogram obtained did show a trend towards normal distribution, although the Anderson Darling and p-value was 13.862 and <0.005 respectively. When the raw data for the crop plants were processed using the square root the, histogram obtained did show a strong trend towards normal distribution, although the p-value was <0.005 for that dataset, however, the Anderson Darling value was 1.372 suggesting a trend towards normal distribution. Therefore, two-way ANOVA analysis was performed as it was crucial for this experiment to investigate the effect between the two independent variables.
3.1 Germination analysis
In the Kruskal-Wallis test where the effect of plant type was investigated on the number of germinated seeds the p-value obtained was <0.001 with two degrees of freedom. The sample size varied between each plant type as the outliers were removed prior to the statistical analysis. Additionally, the H-value obtained was 188.99 when adjusted for ties suggesting a significant statistical difference between the medians of the groups.
When it comes to the Kruskal-Wallis test where the effect of soil type was investigated on the number of germinated seeds the p-value obtained was <0.001 with two degrees of freedom. The sample size varied between each soil type as the outliers were removed prior to the statistical analysis. The H-value obtained was 29.67 when adjusted for ties also suggesting a statistical difference between the medians of the groups, but not as strong as in the plant type investigation.
The results presented in Graph 1suggest that crops had the highest median with the lowers range of data spread. Additionally, the crops performed best in the Mars soil simulants.
Graph 1. The boxplot shows the median of the number of germinated seeds for each plant (crop: 236, nitrogen fixer: 240, wild type: 360) and soil type (earth: 278, mars: 279, moon: 279). The median was used to shows a better representation of the raw number of germinated plants per plant type due to the non-normal distribution of the data. The box shows the interquartile range and the whiskers extend from both sides of the box.
3.2 Total biomass analysis
The two-way ANOVA test for the plant species versus the soil type performed in Minitab showed a p-value of <0.001 when the variance of the total biomass was compared with plant type (2 degrees of freedom) and soil (2 degrees of freedom) with F-value of 116.19 and 126.77 respectively. When the two independent variables were compared (4 degrees of freedom) the p-value was <0.003 and F-value was 15.34 which indicates a statistical significance. Tukey Pairwise Comparison indicated that the data show a statistical difference with crops grown in Martian soil showing a significant difference from the other variables.
The data presented in Graph 2show that crops grown in Mars simulated soil had the highest biomass gained and showed a statistically significant difference from the other groups. The results for the other soil and plant species did not show a strong difference from each other, however when comparing the soil types then crop performed significantly best compared to other plant species.
Graph 2. The mean total dry biomass in milligrams for each of the tree soil and plant types (crop N = 236, nitrogen fixers N = 204, wild type N = 281, mars N = 257, earth N = 242, moon N = 222). Error bars are represented with standard error of the mean. The letters represent the Tukey comparison results obtained in Minitab after performing a two-way ANOVA test of variance.
3.3 Crop plant analysis
The data in Graph 1and Graph 2suggests that crop plants performed best and, therefore, this plant type was investigated further. There were four species of crop plants investigated.
In the Kruskal-Wallis test for both the species and soil comparison to the number of germinated seeds the p-value obtained was <0.001 with three degrees of freedom with the outliers removed prior to the statistical analysis. Additionally, the H-value obtained was 28.55 for species independent variable and 23.81 for soil independent variable when adjusted for ties suggesting a significant statistical difference between the medians of the groups.
The data presented in Graph 3suggest that the Martian soil was best for all the different crop types with median of 5 for each of them. However, this is not the case for the other soil types as there L. sativum could have performed best. However, the range of the data does not strongly support that.
Graph 3. The boxplot shows the median of the number of germinated seeds for each crop species (D. carota N = 58, L. sativum N = 60, S. cereal N = 60, S. lycopersicum N = 60) and soil type (earth N = 80, mars N = 80, moon N = 78). The median was used to shows a better representation of the raw number of germinated plants per plant type due to the non-normal distribution of the data. The box shows the interquartile range and the whiskers extend from both sides of the box.
The two-way ANOVA test for biomass of the crop species versus soil type performed in Minitab showed a p-value of <0.001 when the variance of the total biomass was compared with crop species (2 degrees of freedom) and soil (2 degrees of freedom) and with F-value of 524.20 and 275.65 respectively. When the two independent variables were compared (4 degrees of freedom) the p-value was <0.001 and F-value was 9.12 which indicates a statistical significance. Tukey Pairwise Comparison indicated that the data that show a statistical difference was the S. cereale as shown in Graph 4. The other crop plants do not show a statistical difference between each other when soil type is taken into account.
Graph 4.The mean total dry biomass in milligrams for each of the four crop species (D. carota N = 58, L. sativum N = 60, S. cereal N = 60, S. lycopersicum N = 60) and soil types (earth N = 80, mars N = 80, moon N = 78). Error bars are represented with standard error of the mean. The letters represent the Tukey comparison results obtained in Minitab after performing two-way ANOVA test of variance.
4.1 Plant type investigation
The results indicate that plants do germinate and grow in both terrestrial as well as Moon and mars simulated soil. An effect between soil and plant type and the total biomass can be observed (see Graph 2) which indicates that the null hypothesis could be rejected. However, although the germination of plants was better for crop plants, due to the high spread of data no definite correlation can be drawn from the results (see Graph 1). Several factors could have influenced the germination of the seeds which could have affected the data results. Additionally, the germination data were investigated with a non-parametric statistical test which could affect its credibility. The issue of normal distribution could also have influenced the final data.
When it comes to the total biomass recorded, the Moon simulant performed worst out of the three which is not in line with previous studies of plants exposed to lunar samples (Ferl and Paul, 2010). Additionally, most of the plants performed best in the Martian soil simulants, instead of the earth-like soil as predicted with the highest difference obtained in crop plants. A statistical difference was also observed for the wild type plants as they performed best in Mars soil, compared to the other (see Graph 2).
What is interesting is that the nitrogen-fixing plants did not perform better in the non-terrestrial soil simulants (see Graph 2) as nitrogen availability is of main concern for space agriculture (Palomba et al., 2009; Ferl and Paul, 2010; Stevens et al., 2011).
4.2 Crop plant investigation
Further investigations of the raw data focused on the crop species as this plant type performed best (see Plant type investigation). The results indicate that there is a difference in the biomass favouring S. cereale in both the three different soils with the Tukey test of comparison clearly presenting that S. cereale performed best in Martian soil simulants (see Graph 4). When it comes to the germination, all crop species performed similarly in Martian simulated soil (see Graph 3). For the other soils, L. sativum performed best, however due to the large spread of data these results are not statistically significant. Nonetheless, the data do suggest that S. cereale did perform best as although it had overall a lower median for germinated seeds it still had the highest biomass yield. Therefore, the null hypothesis can be rejected to some extent and more research is needed.
The fact that the data suggest that crop plants performed in Martian soil simulants could be of importance for future agriculture as this, volcanic soil could be used instead of normal terrestrial soil and or fertilisers.
Since the experiment used simulated Martian and Moon soil, the results cannot be considered a precise representation of the non-terrestrial soil effect on plant growth. Additionally, the plants were grown in pods which could have an effect on the growth patterns (Wamelink et al., 2014). Therefore, for future studies, it would be beneficial to use soil samples delivered directly from Moon and Mars as well as try growing the plants in full soil cultivation.
Additionally, half of the investigated naturally occurring plants perform better in nutrient-poor soil whereas the other three perform better in nutrient-poor soil (see Table 2). This can be crucial when the plant is being grown in nutrient-poor soil (see Table 1) such as the one of Mars and Moon (Ferl and Paul, 2010; Maggi, 2010). This could have affected the findings and therefore, should be investigated separately in further studies.
Moreover, the difference in the performance of wild plants could have affected the normal distribution of the data resulting in a slightly non-normal distribution which affected the results obtained from the raw data. To additionally reassure that normal distribution was obtained and the assumptions for a two-way ANOVA met, the sample size could have been increased as well as the number of species investigated. There were four species of both crops and nitrogen fixers and six species of wild type. This difference in the sample size could have had an effect on the normal distribution and, therefore, it would be advantageous for further studies to increase the number of species to at least five per plant type. It would be also beneficial to separate the wild plants to the ones that benefit from nutrient-poor from the plants that benefit nutrient-rich soil and have at least five different species of each.
It would also be interesting to investigate the performance of plants when grown in a multicultural garden as some plant perform better when there is a mix of plants. Additionally, to minimise the number of fertilizers needed for agricultural growth on Mars or Moon the idea of benefits from plants and fungi interactions should be further investigated. The presence of mycorrhizae has been found to be beneficial for agriculture in nutrient-poor soils on in the Mediterranean (Abbott, Robson and Scheltema, 1995). Additionally, the effects of symbiosis of plants with bacteria could be investigated, however, this could be rejected to keep more sterile conditions in the space colonisation.
Another important factor to consider is the conditions in which the plants would be grown. For this experiment, it was assumed that the atmosphere would be the same as on earth for humans to be able to access the crops. However, this might not be the case as machines could be used for agricultural use instead. Therefore, atmosphere could be mechanically adjusted in the agricultural chambers to maximise crop yield and so differ from this experiment’s assumptions. Additionally, the change in the gravitation was not considered in this experiment and that is another factor that could affect the plant growth (Maggi, 2010; Kamal et al., 2018).
To conclude, the results cannot be used to reject the null hypothesis due to the variability of the germination data as well as due to the low reliability of the statistical test used. Nevertheless, these results suggest that crops perform best of the different plant species with S. cereale having best biomass results in all of the three soils however, L. sativum had overall most germinations. Further studies are needed with increased sample size. Since this field of research is relatively new much more investigations are needed before humanity could begin space colonisation.
Abbott, L. K., Robson, A. D. and Scheltema, M. A. (1995) ‘Managing Soils to Enhance Mycorrhizal Benefits in Mediterranean Agriculture’, Critical Reviews in Biotechnology. Taylor & Francis, 15(3–4), pp. 213–228. doi: 10.3109/07388559509147409.
Carlton, C. et al. (2014) JSC MARS-1: Martian regolith simulant, Orbitec website. Available at: http://www.orbitec.com/store/JSC_Mars_1_Characterization.pdf (Accessed: 28 July 2014).
Ferl, R. J. and Paul, A.-L. (2010) ‘Lunar Plant Biology—A Review of the Apollo Era’, Astrobiology. Mary Ann Liebert, Inc., publishers, 10(3), pp. 261–274. doi: 10.1089/ast.2009.0417.
Finetto, C., Lobascio, C. and Rapisarda, A. (2010) ‘Concept of a Lunar FARM: Food and revitalization module’, Acta Astronautica, 66(9), pp. 1329–1340. doi: https://doi.org/10.1016/j.actaastro.2009.10.027.
Kamal, K. Y. et al. (2018) ‘Simulated microgravity, Mars gravity, and 2g hypergravity affect cell cycle regulation, ribosome biogenesis, and epigenetics in Arabidopsis cell cultures’, Scientific Reports, 8(1), p. 6424. doi: 10.1038/s41598-018-24942-7.
Kozyrovska, N. O. et al. (2006) ‘Growing pioneer plants for a lunar base’, 37, pp. 93–99. doi: 10.1016/j.asr.2005.03.005.
Maggi, F. (2010) ‘Space agriculture in micro- and hypo-gravity : A comparative study of soil hydraulics and biogeochemistry in a cropping unit on Earth , Mars , the Moon and the space station’, 58, pp. 1996–2007. doi: 10.1016/j.pss.2010.09.025.
Palomba, E. et al. (2009) ‘Evidence for Mg-rich carbonates on Mars from a 3.9μm absorption feature’, Icarus, 203(1), pp. 58–65. doi: https://doi.org/10.1016/j.icarus.2009.04.013.
Rickman, D., McLemore, C. and Fikes, J. (2007) Characterization summary of JSC-1a bulk lunar mare regolith simulant, Orbitec website. Available at: http://www.orbitec.com/store/JSC-1A_Bulk_Data_Characterization.pdf; http://www.orbitec.com/store/JSC-1AF_Characterization.pdf (Accessed: 28 July 2014).
Stevens, C. J. et al. (2011) ‘Ecosystem responses to reduced and oxidised nitrogen inputs in European terrestrial habitats’, Environmental Pollution, 159(3), pp. 665–676. doi: https://doi.org/10.1016/j.envpol.2010.12.008.
Wamelink, G. W. W. et al. (2005) ‘Plant species as predictors of soil pH: Replacing expert judgement with measurements’, Journal of Vegetation Science. John Wiley & Sons, Ltd (10.1111), 16(4), pp. 461–470. doi: 10.1111/j.1654-1103.2005.tb02386.x.
Wamelink, G. W. W. et al. (2011) ‘Ecological ranges for the pH and NO3 of syntaxa: a new basis for the estimation of critical loads for acid and nitrogen deposition’, Journal of Vegetation Science. John Wiley & Sons, Ltd (10.1111), 22(4), pp. 741–749. doi: 10.1111/j.1654-1103.2011.01286.x.
Wamelink, G. W. W. et al. (2012) ‘Vegetation relevés and soil measurements in the Netherlands : the Ecological Conditions Database ( EC )’, pp. 125–132. doi: 10.7809/b-e.00067.125.
Wamelink, G. W. W. et al. (2014) ‘Can Plants Grow on Mars and the Moon: A Growth Experiment on Mars and Moon Soil Simulants’, PLOS ONE. Public Library of Science, 9(8), p. e103138. Available at: https://doi.org/10.1371/journal.pone.0103138.