A Review of Environmental Impacts of Cereal Grain Supply Chains

,


Introduction
The Earth is facing ever-growing problems, including increasing population, decreasing resources, and an environmental crisis.By working to improve the last 2 problems, only then can we better support the Earth's increasing population.This means taking measures to improve the sustainability of our natural resources and our essential products, without sacrificing the quantity needed for an increasing population.
One of these essential products is cereal grains, which is the primary focus of this paper.Cereal grains have been a staple crop since the dawn of agriculture, and continue to be staples in modern human diets.The overall diversity of cereal grains allows for food in a variety of populations around the globe.It also creates a market for not only domestic consumption, but also global export.Cereal grains are also used as feed for a majority of livestock, including, beef, dairy, swine, poultry, and other animal species.Another newer use with cereal grains is in the production of biofuels, bioplastics, and other industrial products.With grains' ability to contribute to both food and energy supplies, the importance of making these crops more sustainable and abundant is clear.marine toxicity potential, freshwater toxicity potential, soil toxicity potential, ozone layer depletion potential, and photochemical oxidation.When looking at all fourteen of these metrics together, a more complete picture can be formed regarding the environmental impacts of cereal grains.
While compiling this database, many articles were found that contributed to the creation of this meta-analysis.There were many studies conducted that would take either a single grain or a few grains and compared them based on a single LCA metric.Other studies were also found where a single grain was studied, but many different metrics were analyzed.The difference between those articles and this is the scope of comparisons.While other studies compared a few grains or metrics, this study compares many grains, metrics, supply chain stages, and region of growth and production.
Sustainability is not a one-size fits all solution, so geographic diversity was of importance while compiling this analysis.Instead of focusing on a single country or region, it was decided that looking for studies internationally would bring a better perspective to the study and the environmental crises across the globe.This allows for more trends to form and to draw conclusions of which regions have obtained more efficient crop production when it comes to minimizing pollutants and resource use.If one region has consistently better findings than other regions, further analysis into specific practices or products can be applied to other regions with hopes to improve environmental sustainability.This analysis gives an overarching idea about how various grains compare when looking at differentiating LCA metrics.During the compilation phase, there were not any articles that were found comparing the number of grains, the number of metrics, or regional areas as this analysis does.This study will be helpful for looking at trends that would be otherwise impossible to see otherwise due to the scope of information presented.
With agriculture being a staple for cultures around the world, numerous studies have been done examining all aspects of crop production.Publications and journals were searched to find articles that focused upon different life cycle assessment (LCA) metrics for various crops.These crops included common ones, such as barley, maize, millet, oats, rice, rye, sorghum, and wheat, as well as less conventional grains, such as amaranth, buckwheat, quinoa, and triticale.Searches also included simple foods made almost exclusively of one grain, such as bread, tortillas, and pasta.Once a database was amassed from the various research articles collected, they were then categorized by the grain type and then again by what they were being used for.These categories included food production, biofuel creation, and animal feed.
Once the articles were categorized accordingly, the process of extracting information from the articles began.The use of an Excel spreadsheet was used to hold the information that was extracted.The data were separated into qualitative information and quantitative information.Qualitative data extracted included the grain that was studied, the region the study took place in, what the crop was being used for, the stage in the supply chain that was being examined, and a description of differences in certain measurements.The quantitative data extracted was the measurements the study found.In total there were 14 different LCA metrics that were found to be the most common.The spreadsheet accommodated all of the data, as well as an addition column beside each metric so that a unit could be associated with each data value.
One element that should be noted was the inclusion of greenhouse gas emissions (GHG) and global warming potential (GWP).Most of the time, these two categories would be combined into one since they are very similar, measuring carbon dioxide release into the atmosphere.The difference arrives in the totality of the measurement.GHG measures only CO 2 emissions, while GWP measures CO 2 emissions as well as other emissions that make up the global warming crisis.For the purpose of comparing all of the studies, it was decided to keep the two categories separate.The biggest reason for this was that there were multiple studies that reported the emissions as GHG only, multiple studies that reported the emissions as GWP only, and a distinct few that reported both separately.
From all the various research articles collected, one main discrepancy was the establishment of Functional Units (FU).The research collected fell into one of three functional units: land, mass, and energy.While efforts could have been made to combine these three groups into a massive comparison, ultimately the data were easier to digest and demonstrated less anomalies if separated by functional unit.
Even with increasing standardization of units, there was still a need for conversion to a single set of units for easier comparison.For the land functional unit, the standard unit was meters squared (m 2).The mass functional unit has a standard unit of kilograms (kg), and energy functional unit was measured in joules (J).Units that needed to be converted included tonnes, tons, pounds, acres, hectares, and kilowatt hours.Here are the formulas used to convert into standardized units: With the data compiled into uniform units and functional units, the decision was made to compare the data using bar graphs.Each functional unit has its own category of graphs, and these graphs were based on LCAs.Some FUs did not have any data on some LCAs so there were holes in the graphs based on these weaknesses in pre-existing research.The resulting graphs were separated first by grain, then by supply chain stage.Each grain has a color assigned to it in order to easily differentiate between the grains.In addition, there were two shades for each color to denote differences in authorship and supply chain stages.With the mass graphs, supply chain stage is denoted on the x-axis, however the land graphs did not see enough variance from the supply chain stage to warrant an additional comparison on the x-axis.The numerical values were displayed above the data point for easier access as well as the key describing the authorship for the various studies.Units and scales range on each graph, which was why the decision was made to keep the graphs separate based on LCAs and functional units.We were comparing each value to the other values on the same graph so that a trend can be determined based on the findings of the studies collected in each of the various metrics.

Greenhouse Gas Emissions (GHG)
The studies with the functional unit of land use had measurements in fourteen different metrics, with some studies having data for multiple grains across multiple metrics.The most diverse selection of data was when greenhouse gas emissions were compared.These measurements can be found in Figure 1 with Table 1 giving more information about the studies, including region and farming practices used.The highest measurement was found in oats with 8.50 x 10 7 kg of CO 2 emitted per FU and the lowest was found in maize with 4.9 x 10 -2 of CO 2 emitted per FU.In this selection of data, there were many different studies that consist of multiple entries looking at one specific grain, such as buckwheat, maize, oats, and wheat.Where these studies differed was the agriculture practices that were used as variables to test certain agriculture practices.These variables mainly consisted of the amount of tillage done to the land, and whether a cover crop was used or not.Some anomalies that were present were grains that had studies showing them both near the higher and lower end of the spectrum, such as sorghum.This could be because they were different studies and may have had other factors that impacted the results that were not presented in their respective articles.Some trends that that were present show that for each grain being compared, there were studies that show the grain on the low and as well on the high end of the spectrum.
Figure 1.(Top) Greenhouse gas emissions for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 1.(Bottom) Global warming potential for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 2 3.1.2Global Warming Potential (GWP) The measurements for global warming potential metric for land use functional unit is found in Figure 1, with Table 2 giving more information about the studies, including region and farming practices used.The highest measurement was found in sorghum with 1.81 x 10 8 kg CO 2 per FU and the lowest was found in maize with 2.71 x 10 3 kg CO 2 per FU.In this selection of data, every grain with the exception of bread had multiple measurements done comparing different variables.These variables were the amount of tillage, the farming techniques used, irrigation systems, and the grains planted before the measurement crop.There was a high outlier with one sorghum measurement, which was over three times higher than any of the previous measurements in the set.This outlier could be caused by that particular study being done at a much smaller scale than the other studies, so when the measurements were scaled up to match each other, smaller differences were exaggerated.This was most likely the case since the majority of the time this study was included in a dataset, it was consistently the highest.

Water Use (H 2 O)
Figure 2 and Table 3 refers to water use for the land based functional unit.The highest recorded value was oats with 2.71 x 10 7 kg of water per FU, and the lowest was barley with 2.50 x 10 7 kg of water per FU.Only two studies were found that measured water use with this particular functional unit, and only three grains were studied.But between the two studies the results were of the grains showed that all three were very similar water usages.
Figure 2. (Top) Water use for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 3. (Bottom) Land use for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 4 3.1.4Land Use (Land) Land use was another unique category that focused on the efficiency of farms given the choice to measure land use while also having the functional unit also be land used.The data for these studies are found in Figure 2 as well as additional information in Table 4.The highest measurement was from oats with 5.6 x 10 7 m 3 of land used per FU and the lowest measurement being 1.00 x 10 4 m 3 of land used per FU.The biggest outlier was with bread, and that as mostly due to the fact that multiple plants were produced to make enough grain for a single loaf of bread, bring the overall efficiency down.

Acidification Potential (AP)
Acidification potential is the measurement of various emissions that are released into the atmosphere that lowers the pH of air and water bodies.The measurements are found in Figure 3 and Table 5.The highest value was sorghum with 5.41 x 10 5 kg of emissions per FU and the lowest value was wheat with 1.78 x 10 5 kg of emissions per FU.In this collection of measurements sorghum was a slight outlier, being about twice as much as the next leading value.Since this was the same study that shows sorghum surpassing most other grains in multiple other graphs, this may suggest that there is not as much variance in acidification potential between grains as there were with other metrics.
Figure 3. (Top) Acidification Potential for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 5. (Middle) Abiotic depletion for rice and sorghum using Land Functional Units.Descriptive information about the studies can be found in Table 6.(Bottom) Eutrophication Potential for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 7 3.1.6Abiotic Depletion (AD) Figure 3 and Table 6 refer to the abiotic depletion and the amount of antimony in a given area.The highest value was sorghum with 6.57 x 10 5 kg of Sb/FU while the lowest value was rice with 3.58 x 10 5 kg Sb/ FU.Out of the three measurements taken, two of them were the same study which looked at the different farming techniques.Within this study, it showed that conventional based farming methods have a better abiotic depletion rate than traditional based farming methods.

Eutrophication Potential (EP)
For eutrophication potential, there were two different ways that this metric was measured.The most common way was with the amount of phosphate in an area, and the other was the amount of nitrogen in an area.This distinction was apparent in Figure 3 and Table 7.The highest value was sorghum with 2.45 x 10 5 kg PO4/FU and the lowest was 4.80 x 10 3 kg N/ FU, or 2.77 x 10 4 kg PO4/FU.There was a high outlier with sorghum, being over seven times higher than the next largest grain.This could be because it was part of the particular study that has high sorghum measurements in nearly every chart.There were also two rice measurements that were from the same study, which suggest that conventional farming practices were better than traditional farming practices.

Human Toxicity Potential (HTP)
Human toxicity potential was one of the more diverse graphs with the land use functional unit.It has a larger spread of grains than most of the graphs with very diverse values between points, as seen on Figure 4 and Table 8.It was also unique because it has two different ways it was measured, with a majority of the research being done in kg of 1,4 DB and only one study being done in kg of Pb.All of this was shown in Figure 4 and Table 8.The highest value was sorghum with 5.98 x 10 7 kg of 1,4 DB while the lowest was maize with 3.14 x 10 2 kg of 1,4 DB.There are two ranges present in the graph: a low range with maize, oats, and barley, and a middle range with rice and wheat.There was also an outlier with sorghum, which was over 3 times higher than the next highest value.This was not as unreasonable as it was in other graphs since this range as quite a bit bigger than most of the other land use graphs.
Figure 4. (Top left) Human toxicity potential for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 8. (Top right) Non-renewable energy depletion for sorghum using Land Functional Units.Descriptive information about the studies can be found in Table 9. (Bottom left) Marine aquatic toxicity potential for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 10.(Bottom right) Freshwater aquatic toxicity potential for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 11 3.1.9Non-Renewable Energy Depletion (NRED) There was only one study that was done to find the nonrenewable energy depletion for any of the grainsthis was sorghum, and Figure 4 and Table 9 show the one point of data that was collected.The reason that there was only one point was because it is a very uncommon metric.It is typically only applicable when measuring grains when used as an energy source and there is a separate functional unit dedicated to measuring different aspects of grains when used as a fuel source in terms of energy used.Since there was only one measurement, no trends could be inferred for this metric.

Marine Toxicity Potential (MTP)
When water toxicity is being compared it is usually separated into two separate categories, freshwater toxicity and marine toxicity.Figure 4 and Table 10 shows the marine toxicity we found for a variety of grains.The highest grain was sorghum with 3.53 x 10 10 kg 1,4 DB/FU and the lowest value was barley with 3.57 x 10 4 kg 1,4 DB/FU.There was a large disparity between the high end of the data for grains such as sorghum and rice being many magnitudes higher than barley, maize, and oats at the low end of the data.The freshwater toxicity measurements were similar to these results, but with much less of a divergence between the ends of the data.

Freshwater Toxicity Potential (FTP)
Freshwater toxicity shows mainly the same story as MTP, as seen in Figure 4 and Table 11.The highest value was sorghum with 1.61 x 10 7 kg 1,4 DB/FU while the lowest value was maize with 73.9 kg 1,4 DB/FU.A difference between these measurements and ones found in the marine toxicity section was that there was a wheat value that was measured in kg of Zinc instead of the 1,4 DB.This section was similar to the marine toxicity in that sorghum and rice were quite a bit higher than the other grains, although not as extreme.With both of these datasets taken into consideration, these data strongly suggest sorghum and rice should be promoted when talking about aquatic toxicity.
3.1.12Soil Toxicity Potential (STP) Figure 5 and Table 12 refer to the soil toxicity potential in a given area.The highest value was sorghum with 4.02 x 10 5 kg 1,4 DB per FU while the lowest value was rice with 1.80 x 10 5 kg 1,4 DB per FU.Out of the three measurements taken, two of them were the same study which looked at the different farming techniques.Within this study, it showed that conventional based farming methods have a better soil toxicity rating than traditional based farming methods.
Figure 5. (Top left) Soil toxicity potential for rice and sorghum using Land Functional Units.Descriptive information about the studies can be found in Table 12.(Top right) Ozone layer depletion for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 13.(Bottom left) Photochemical oxidation for various grains using Land Functional Units.Descriptive information about the studies can be found in Table 14 3.1.13Ozone Layer Depletion Potential (ODP) Ozone layer depletion was interesting because it had the largest outlier out of all the datasets for the functional unit of land use.It is seen in Figure 5 and Table 13 where only one datapoint was visible because it was so much higher than the rest.This goes to sorghum with 5.89 x 10 8 kg 1,4 DB per FU with the next highest grain being rice with 9 kg 1,4 DB per FU.Along with the graph, nearly every other graph that features this study of sorghum, with some exception, was always at the top of every graph.

Photochemical Oxidation (PO)
The last dataset for the land use function unit was photochemical oxidation.This dataset was different from the others because one third of the datapoints were measured in kg of Non-Methane Volatile Organic Compound (or NMVOC) while the rest was measured in kg of ethylene.These data are available on Figure 5 and on Table 14.
This highest value was 2.20 x 10 5 kg NMVOC per FU while the lowest value was 0.291 kg C2H4 per FU.There was a large disparity between the NMVOC measurements and the C 2 H 4 measurements, with all NMVOC measurements being consistently higher than any C 2 H 4 measurements.This could mean that all grains emit more NMVOC's than ethylene, but more studies should be done to test this claim.

Greenhouse Gas Emissions (GHG)
With the functional unit of land completed, the focus of the study now turns to the mass functional unit.Within the graph (Figure 6) and the supplementary table (Table 15), information about how various cereal grains can affect greenhouse gas emissions can be found.The highest greenhouse gas emission was the production of pasta with 1.6 kg of CO 2 per kg of product.Meanwhile, the lowest emissions came from the growing of wheat with 0.026 kg of CO 2 per FU.There were multiple comparison studies contained within this graph.The first comparison was the production of bread.Differences come from different stages of the overall production process.Namely preparation, processing, and producing.Secondly, the comparison of maize growing practices compares different geographic regions of the United States, taking into account differences in fields, rainfall, and fertilization techniques.The comparison of the growing of oats follows the same comparison as the aforementioned bread.Finally, the production of rice differences follows the same comparison as the first comparison listed in the different stages of the overall production processes.Overall conclusions that may be drawn from this graph was the low greenhouse gas emissions of wheat, bread, and barley as well as the higher greenhouse gas emissions of millet, pasta, and sorghum.
Figure 6.(Top) Greenhouse gas emissions for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 15.(Bottom) Water use for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 16 3.2.2Water Use (H 2 O) When comparing the graph of Figure 6, and the corresponding table (Table 16), two things become obvious.The water usage of oats and rice were both greater than the rest of the grains compared within the water use graph.
Oats has a value of .649m 3 of water used per FU and rice has 0.701 m 3 per FU.Meanwhile, the production of bread has the lowest values at 0.0125 m 3 and .0166m 3 per FU.There were multiple comparison studies contained within this graph.The first comparison was the production of bread.The difference in data accommodates for the differences in water usage according to green, blue, and gray water respectively.The rest of the comparisons contained within this graph are attributed to geographic differences and more information can be found in Table 16.One possible outlier was the fourth rice value.These data originate from Uzbekistan and could possibly explain the high-water usage than the other countries in that study.Overall trends demonstrate that maize, bread, and pasta have low water usage, while rice and oats have higher water usage.This follows what previous research has concluded due to both of those crops high water intensity per FU.

Abiotic Depletion (AD)
In Figure 7 and the corresponding Table 17, the abiotic depletion of rice was shown.Due to lack of previously conducted research on abiotic depletion within the mass functional unit, only rice was able to be compared.The differences between the two values can be attributed to the differences in farming practices in Iran.The first value uses consolidated or modern farming practices, while the second value uses traditional farming practices.The higher value of abiotic depletion for traditional farming practices suggests that improvements in farming techniques for rice has reduced abiotic depletion.

Eutrophication Potential (EP)
For the mass FU graphs of the eutrophication potential, see Figure 7.The corresponding tables were also included in the analysis.Current differences in the unit used to conduct eutrophication potential analyses dictated the separation of graphs.For Figure 7, the unit was kg of Nitrogen emissions per FU.The highest value was 2.8x10 -3 kg of N for the growing of wheat.Meanwhile, the lowest value was the growing of barley with 1x10 -5 kg of N. For the first two comparisons within this figure, the differences were differences in regional growing practices.For the last comparison, the difference can be attributed to the differences in production of biofuels.
The first point uses straw gasification, while the second point uses straw direct combustion.Outliers within this graph include the low value of barley and one study of wheat.For Figure 7, the unit was kg of PO 4 emissions per FU.The highest value was the growing of barley with 5.62x10 -4 kg of phosphate per FU.The lowest was the growing of rice with 5.2x10 -5 kg of phosphate per FU.Two comparisons take place within this graph.The first comparison can be attributed to the differences in farming practices in Iran.The first value uses consolidated or modern farming practices, while the second value uses traditional farming practices.The second comparison concerns the type of rice grown.The first value reflects conventional rice, while the second value uses long-term organic rice.Overall trends to be noted within this graph was the high value of barley, and the similarity between wheat and rice.
Figure 7. (Top left) Eutrophication Potential of Nitrogen for various grains using Mass Functional Units.
Descriptive information about the studies can be found in Table 17.(Top right) Eutrophication potential of Phosphate for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 18.(Bottom left) Abiotic depletion for rice using Mass Functional Units.Descriptive information about the studies can be found in Table 19.(bottom right) Human toxicity potential for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 20 3.2.5 Human Toxicity Potential (HTP) Figure 7 and Table 20 refer to the human toxicity potential of various grains in the mass FU.The highest value was the production of bread at 3x10 -1 kg of 1,4 DB per FU.Meanwhile the lowest value was the growing of maize at 1x10 -2 kg of 1,4 DB per FU.For the comparisons contained within the human toxicity potential graph, the first and last comparison can be attributed to differences in geographic practices.For the comparison of rice growing practices, the comparison was the differences in growing practices in Iran.The first value uses consolidated or modern farming practices, while the second value uses traditional farming practices.Trends to be noted within the graph show that there was not great variance of the human toxicity potential for various grains and their various supply chain stages.

Marine Toxicity Potential (MTP)
The marine toxicity potential graph (Figure 8 and Table 21) details the affect that cereal grains have upon marine health.Measured in kg of 1,4 DB per FU the highest value was the growing of rice with the value of 1.66x10 -2 .The lowest value was 4.8x10 -5 for the growing of triticale.Only one comparison takes place on this graph and that comparison was between the growing practices of rice.The first value uses consolidated or modern farming practices, while the second value uses traditional farming practices.One takeaway from this analysis was that the marine toxicity potential for rice was higher than the other grains.This can be attributed to the large water usage for rice in comparison to the other grains found in this graph.
Figure 8. (Top left) Marine toxicity potential for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 21.(Top right) Freshwater toxicity potential for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 22.(Bottom left) Soil toxicity potential for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 23.(Bottom right) Ozone layer depletion for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 24 3.2.7 Freshwater Toxicity Potential (FTP) When analyzing the graph of the freshwater toxicity potential (Figure 8), and its corresponding table (Table 22), the impact that rice has was clear.Rice has the largest impact with 7.48x10 -3 kg 1,4 DB per FU.Meanwhile, triticale has the lowest with 4.4x10 -5 kg 1,4 DB per FU.The first comparison found was the comparison between growing practices of rice.These two practices were consolidated and traditional farming practices.The other two comparisons were both geographic differences within France.One takeaway from this analysis was that the freshwater toxicity potential for rice was higher than the other grains.This can be attributed to the large water usage for rice in comparison to the other grains found in this graph.

Soil Toxicity Potential (STP)
The soil toxicity potential is analyzed in Figure 8 and Table 23.The largest value was the growing of oats with 6.4x10 -4 kg 1,4 DB per FU.The smallest was the production of maize with 9x10 -5 kg 1,4 DB per FU.Three comparisons are found on this graph.The first comparison found was the comparison between growing practices of rice.These two practices were consolidated and traditional farming practices in Iran.The second comparison were various geographic differences.The final comparison was between the years 2010 and 2012, thus accounting for differences in rainfall and soil productivity.Overall, the graph does not display a large difference between any particular grain and the toxicity of the soil, however oats and rice have a larger value than the other grains found in this analysis.

Ozone Layer Depletion Potential (ODP)
Within the ozone layer depletion graph and table (Figure 8 and Table 24), the affect that the various cereal grains have upon the ozone layer was measured in kg of CFC-11 emitted during the supply chain stages.The highest value was the production of maize with 7.25x10 -8 kg of CFC-11 per FU.Meanwhile the lowest was the growing of bread at 5x10 -10 kg of CFC-11 per FU.Possible outliers within this graph include a value of bread growing and maize production.These outliers could be attributed to different regional practices or methods of collecting data between studies.Overall trends show that the growing of grains has less impact than the production of grains upon the ozone layer.

Photochemical Oxidation (PO)
Concluding the mass functional unit, analysis was conducted of the photochemical oxidation measured in kg of NMVOC per FU (Figure 9 and Table 25).The highest value was the production of oats with 5.75x10 -3 kg of NMVOC per FU.The lowest value was the growing of maize with 1.30x10 -4 kg of NMVOC per FU.Overall, the production of grains has a higher impact on photochemical oxidation than the growing of said grains.
Figure 9. (Top left) Photochemical oxidization for various grains using Mass Functional Units.Descriptive information about the studies can be found in Table 25.(Top right) Greenhouse gas emission for maize and rice for Energy Functional Units.Descriptive information about the studies can be found in Table 26.(Bottom left) Global warming potential for various grains using Energy Functional Units.Descriptive information about the studies can be found in Table 27.(Bottom right) Water use for various grains using Energy Functional Units.
Descriptive information about the studies can be found in Table 28 3.3 Energy FU

Greenhouse Gas Emissions (GHG)
The final functional unit was the energy functional unit, measured in Joules.There has not been extensive research conducted on the effect that various grains have upon LCAs in terms of biofuels.There was some literature though, which has been analyzed in the following figures and tables.For greenhouse gas emissions, there were two data points (Figure 9 and Table 26).Rice has the highest emissions at 1.17x10 12 kg of CO 2 /J of energy produced.The lowest point was maize with 1.55x10 5 kg of CO 2 /J of energy.Overall trends suggest that rice causes a greater greenhouse gas impact than maize.

Global Warming Potential (GWP)
In a similar vein to greenhouse gas emissions, the global warming potential has been measured for rice, rye and wheat (Figure 9 and Table 27).The largest impact was rye with 3.69x10 16 kg of CO 2 .The smallest impact was wheat with 4.10x10 7 kg of CO 2 .There was one comparison conducted between wheat.The difference between the data points can be attributed to confidence intervals in the original study.One possible outlier was the large impact of rye which could be caused by geographic differences or insufficient data being collected to accurately compare and draw conclusions.

Water Use (H 2 O)
The water usage needed for biofuels was estimated for barley, maize, rice, rye, sorghum and wheat (Figure 9 and Table 28).The highest value was 4.19x10 11 m 3 of water used for sorghum.The lowest value was maize with a value of 1.1x10 11 m 3 of water used.A possible outlier was sorghum, however insufficient data was collected to determine possible outliers.Overall, the water usage graph demonstrates the general consistency of water usage for use in biofuels.

Acidification Potential (AP)
The acidification potential data (Figure 10 and Table 29) have been quantified for rice, rye, and wheat in units of SO 2 .Rye has the largest value at 5.22x10 14 kg of SO 2 and wheat has the smallest value at 3.24x10 3 kg of SO 2 /J.A possible outlier was rye, however insufficient data was collected to determine outliers.Overall, the acidification graph details the similarity between rice and wheat in terms of acidification potential.
Figure 10.(Top left) Acidification Potential for various grains using Energy Functional Units.Descriptive information about the studies can be found in Table 29.(Top right) Eutrophication Potential for various grains using Energy Functional Units.Descriptive information about the studies can be found in Table 30.(Bottom left) Human toxicity potential for rye using Energy Functional Units.Descriptive information about the studies can be found in Table 1.(Bottom right) Marine aquatic toxicity potential for rye using Energy Functional Units.
Descriptive information about the studies can be found in Table 32 3.3.5Eutrophication Potential (EP) Eutrophication potential (Figure 10 and Table 30) shows the impact that rice, rye, and wheat have upon phosphate emissions.Rye has the highest value at 6.11x10 14 kg PO.Meanwhile, rice has 1.35x10 11 kg of PO 4 /FU.One thing to note was that the wheat values mentioned in the graph were the data for the amount of nitrate absorbed, so it has a negative LCA measurement.31-35).These graphs all display the values for rye in each corresponding LCA.No comparisons or conclusions can be drawn due to the limited scope of the research.

Photochemical Oxidation (PO)
The final environmental impact was photochemical oxidation and can be found in Figure 11 and Table 36.This graph compares rice and rye.Rye has the largest value with 1.87x10 13 kg of C 2 H 4 /J and rice has the smallest with 1.22x10 9 kg of C 2 H 4 /J.Possible outliers could be either point, however insufficient data has been collected to make that determination.Based on the presumption that the data were accurate, rice has a significantly lower impact on photochemical oxidation than rice.
Figure 11.(Top left) Freshwater aquatic toxicity potential of rye using Energy Functional Units.Descriptive information about the studies can be found in Table 33.(Top right) Soil toxicity potential for rye using Energy Functional Units.Descriptive information about the studies can be found in Table 34.(Bottom left) Ozone layer depletion for rye using Energy Functional Units.Descriptive information about the studies can be found in Table 35.(Bottom right) Photochemical oxidation for rice and rye using Energy Functional Units.Descriptive information about the studies can be found in Table 36 One of the most notable differences when comparing the graphs was that there were many graphs where sorghum was many magnitudes higher than the rest of the grains.Once this trend was noted, further analysis of the paper done by Caffery et al. revealed that their study was done using sweet sorghum instead of grain sorghum.The biggest difference between the two was that grain sorghum was harvested by collecting the grains, like nearly every other comparison that was done in this meta-analysis, while sweet sorghum was harvested for the entire stalks and is predominantly used for syrup.This accounts for the massive disparity in many of the charts that measure sweet sorghum.

Implications
This review identifies and discusses specific areas for agricultural improvements throughout cereal grain supply chains.While some environmental aspects may not be applicable to all production, due to geographic differences, by further examining various agricultural practices around the globe, these practices can be studied and tested in varying agricultural regions.For example, considering the land functional unit and the freshwater toxicity potential for wheat, there was a substantial difference in environmental impacts.The first data set originates from France and the other from Italy.While these two regions may not differ substantially environmentally, the higher values from France should indicate the potential for alterations of agricultural practices based on the results of this paper.For a more severe geographic difference, once again considering the land functional unit for the eutrophication potential of wheat, there was an apparent difference between wheat grown in France and wheat grown in Iran.While geographical differences play a role here, future studies could be conducted to see if any of the practices used in France could be applied with positive environmental impacts in Iran.There were several more examples of this throughout the study where geographic differences can be applied to work on general improvements.
Another implication of this research is general improvements needed during specific stages of supply chains.Overall, the growing stage performed significantly worse than either the transportation or processing stages of the supply chain.This is something that almost all crops in almost all regions need to work to improve on, and future research needs to be conducted in order to determine what steps need to be taken to minimize the difference between these stages.

Conclusions
The biggest aspect found during the analysis of the articles was that there was no one grain that was primarily better than another in terms of environmental performance.A particular grain can do well for some environmental parameters, and in another it can do poorly.Each grain has its own specific strengths and weaknesses that account for the emissions that they release to the environment.For best practices going forward, these data suggest that we should continue growing a variety of grains so that we don't create an excess of one particular pollutant.Another suggestion is that we determine which emissions we want to minimize and which ones we can give more leeway to.Currently, lower emissions such as ozone and greenhouse gasses have a high priority, and we can work towards growing fewer crops that have these emissions or making those grain supply chains less environmentally impactful.Liu, et al. 2018 Australia Conventional farming practices used in Australia.Rajaniemi & Mikkola, 2011 Finland Conventional farming practices used.Rajaniemi & Mikkola, 2011 Finland Reduced tillage on the soil.Charles, et al. 2005 No Region Listed Good agricultural practices, as stated by Swiss regulations.Yan, et al. 2014 China Average taken from 123 farms.Charles, et al. 2005 No Region Listed Good agricultural practices used, based on Swiss regulations.Charles, et al. 2005 No Region Listed Good agricultural practices used, based on Swiss regulations.

Appendix
Table 9. Descriptive information for studies in Fig. 4 Author Geographical Region Summary Patel, et al. 2019 India Conventional farming practices used.Charles, et al. 2005 No Region Listed Good agricultural practices used, based on Swiss regulations.Table 27.Descriptive information for studies in Fig. 9 Authors Geographical Region Summary Chungsangunsit, et al. 2004 Thailand Only rice production used in measurement.Sastre, et al. 2016 Spain Conventional farming practices used.Nguyen, et al. 2013 Denmark Straw underwent gasification.Sastre, et al. 2015 Spain Bottom end of 95% confidence interval.Sastre, et al. 2015 Spain Top end of the 95% confidence interval.Chungsangunsit, et al. 2004 Thailand Only rice production used in calculation.Sastre, et al. 2016 Spain Conventional farming practices used.


Tonnes

3. 3
.6 HTP/ MTP/ FTP/ STP/ ODP Due to lack of previous research using energy as a FU, the graph and tables of human toxicity potential, marine toxicity potential, freshwater toxicity potential, soil toxicity potential, and ozone layer depletion potential only have one data point per graph (Figures 10 and 11 and Tables

Table 1 .
Descriptive information for the studies in Fig.1

Table 2 .
Descriptive information for studies in Fig.1

Table 3 .
Descriptive information for studies in Fig.2

Table 4 .
Descriptive information for studies in Fig.2 Gonzá lez-Garcí a, et al. 2016 No Region Listed Conventional farming practices used.

Table 5 .
Descriptive information for studies in Fig.3

Table 6 .
Descriptive information for studies in Fig.3

Table 7 .
Descriptive information for studies in Fig.3

Table 8 .
Descriptive information for studies in Fig.4

Table 10 .
Descriptive information for studies in Fig.4

Table 11 .
Descriptive information for studies in Fig.4

Table 12 .
Descriptive information for studies in Fig.5

Table 13 .
Descriptive information for studies in Fig.5

Table 14 .
Descriptive information for studies in Fig.5

Table 15 .
Descriptive information for studies in Figure6

Table 16 .
Descriptive information for studies in Figure6

Table 17 .
Descriptive information for studies in Figure7

Table 18 .
Descriptive information for studies in Figure7

Table 20 .
Descriptive information for studies in Figure7

Table 21 .
Descriptive information for studies in Figure8

Table 22 .
Descriptive information for studies in Figure8

Table 23 .
Descriptive information for studies in Figure8

Table 24 .
Descriptive information for studies in Figure 8 Authors Geographical Region Summary Bartzas, et al. 2015 Spain For use in biofuels, conventional growing practices Lechón, et al. 2005 Spain Data obtained from the European Fertilizer Manufacturers Association from the year 2000 for data pertaining raw materials, energy, and fertilizer Gonzá lez-Garcí a, et al. 2016 USA Conventional growing practices Kulak, et al. 2015 France Northern France, bread from integrated crop and livestock production Kulak, et al. 2015 France Southern France, bread from horse farming Guzmá n-Soria, et al. 2019 Mexico 90 kg N/ ha applied w/ 20% at pre-plant, 40% at tillering, 40% at stem elongation Noya, et al. 2015 Italy Conventional growing practices Gonzá lez-Garcí a, et al. 2016 USA

Table 25 .
Descriptive information for studies in Figure9

Table 26 .
Descriptive information for studies in Fig.9

Table 29 .
Nguyen, et al. 2013tion for studies in Fig.10Authors Geographical Region SummaryChungsangunsit, et al. 2004 ThailandOnly rice production used in measurement.Shafie, et al. 2011Malaysia Entire life cycle of rice used in measurement.Sastre, et al. 2016Spain Conventional farming practices used.Nguyen, et al. 2013Denmark Straw underwent gasification.

Table 30 .
Descriptive information for studies in Fig.10

Table 31 .
Descriptive information for studies in Fig.10

Table 32 .
Descriptive information for studies in Fig.10

Table 33 .
Descriptive information for studies in Fig.11

Table 34 .
Descriptive information for studies in Fig.11

Table 35 .
Descriptive information for studies in Fig.11

Table 36 .
Descriptive information for studies in Fig.11