The Pennsylvania State University the Graduate School the POTENTIAL, SAFETY, and ENVIRONMENTAL IMPACTS of GROWING DUCKWEED in AN

Home , Lemnoideae

The Pennsylvania State University The Graduate School

THE POTENTIAL, SAFETY, AND ENVIRONMENTAL IMPACTS OF GROWING

DUCKWEED IN AN ECOLOGICAL WASTEWATER TREATMENT SYSTEM AS A SOURCE

OF PROTEIN FOR LIVESTOCK OR FISH

A Dissertation in Environmental Engineering by Benjamin J. Roman

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2020

The dissertation of Benjamin J. Roman was reviewed and approved by the following:

Rachel A. Brennan Associate Professor of Environmental Engineering Dissertation Adviser Chair of Committee

John M. Regan Professor of Environmental Engineering

Joshua D. Lambert Professor of Food Science

Charles T. Anderson Associate Professor of Biology

Patrick Fox Department Head of Civil and Environmental Engineering

ABSTRACT

Duckweed (Lemnaceae) are small, floating aquatic plants that have rapid growth rates and high protein contents when grown on nutrient-rich water, making them an ideal candidate to be used as a protein source. When combined with ecological wastewater treatment systems, duckweed can simultaneously remove unwanted nutrients from wastewater and produce a large quantity of protein-rich biomass that can be used for livestock or fish feed. The benefits of ecological wastewater treatment systems and the utilization of duckweed as a protein supplement are individually well documented, but no studies have investigated the potential of combining the two processes to recover duckweed grown in a wastewater treatment system as a protein source. In this dissertation, wastewater-grown duckweed was evaluated for its protein production potential, food safety, environmental impacts, and protein quality using a combination of laboratory-scale experiments and modeling.

Duckweed grown in a pilot-scale ecological wastewater treatment system (the Penn State Eco-

MachineTM) was first evaluated for its growth rate and protein content. It was determined that wastewater- grown duckweed can produce over 10 tonne/ha-yr of protein, which is 5-10 times higher than common land-grown crops (soybean, oats, etc.). In addition, after drying for 48 hours at 40oC or 4 hours at 60oC

(which is required for long-term storage), it was found that that no E. coli remained throughout the biomass. Thus, these studies suggest that wastewater-grown duckweed can be below regulatory limits for microbial food safety when dried at the appropriate temperature and duration.

In addition to evaluating the safety of dried plant biomass, counts of coliform bacteria and E.coli were measured on water samples collected throughout the Penn State Eco-MachineTM and after UV disinfection to determine if the treated effluent from the Eco-MachineTM is of sufficient quality to be effectively UV disinfected and used as irrigation water (≤126 CFU E.coli / 100mL). The results show that the Penn State Eco-MachineTM can reduce 3 logs of E.coli and 4 logs of coliform, and that remaining coliforms and E.coli are below detection after the UV disinfection system, suggesting that these systems are capable of producing effluent that is of sufficient quality to be used for irrigation water.

iii

A life cycle assessment was performed with the aim of quantifying the environmental impacts of the Penn State Eco-MachineTM treating municipal wastewater while concomitantly producing animal feed

(derived from duckweed) and irrigation water (derived by UV disinfection of the treated water). The results show that Eco-MachinesTM that do not require heating (i.e., in tropical/equatorial regions) consume about a third of the energy and produce half of the greenhouse gas emissions as conventional wastewater treatment systems (not including the GHG that are produced during the treatment process, such as NOx,

CH4, etc.). This would also have the added benefit of being in climatic region that is ideal for duckweed growth, further increasing the benefits of the system. In addition, increasing the growth area of duckweed by the use of vertical farming within the Eco-MachineTM improved the overall beneficial impact of the system. These results show that coupling feed/irrigation water production with ecological wastewater treatment systems creates a net benefit to human health and the environment.

Finally, a study measuring the body growth, food intake, and final organ and adipose tissue mass of male CF-1 mice fed a diet where 10% or 25% of the control protein (casein) was replaced with duckweed protein (DWP) was conducted to determine if duckweed protein is similar in quality to casein.

The average growth rates of the mice fed the three diets over a 30 day period were not significantly different: 0.21 g/day for the control diet; 0.24 g/day for the 10% DWP diet; and 0.25 g/day for the 25%

DWP protein diet. The daily food intake of both DWP diets was 6.5 – 8.0% higher than the control diet, but feeding efficiency (body weight gain/food intake) did not differ among diets. The relative weight of the liver, spleen, kidneys, heart, and epidydimal fat, and the colon length were not significantly different between treatment groups. The results from this study indicate that replacement of up to 25% dietary casein with DWP has no adverse effects on the growth rate and final organ and adipose tissue weights of laboratory mice.

TABLE OF CONTENTS LIST OF FIGURES ...... viii LIST OF TABLES ...... x ACKNOWLEDGEMENTS ...... xii Chapter 1 - Introduction ...... 1 Humans and agriculture ...... 1 Global food crisis ...... 1 Combining wastewater treatment with food production ...... 3 Safety concerns ...... 3 Life cycle assessment ...... 4 Duckweed as a protein source ...... 5 References ...... 10 Chapter 2 – A beneficial by-product of ecological wastewater treatment: an evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture ...... 13 Abstract ...... 13 Introduction ...... 14 Materials and Methods ...... 16 Experimental setup ...... 16 Duckweed harvesting, growth rate, and protein analysis ...... 17 Water sampling and analyses ...... 17 Metals analysis ...... 18 Statistical analysis ...... 18 Results and Discussion ...... 18 Wastewater and duckweed biomass characterization ...... 18 Duckweed protein trends ...... 23 Duckweed protein economic value ...... 26 Conclusion ...... 27 Acknowledgements ...... 27 References ...... 28 Chapter 3 – Pathogen removal in the Penn State Eco-MachineTM and identification of links between microbiomes and microbiological water safety ...... 30 Introduction ...... 30 Materials and methods ...... 32 The Penn State Eco-MachineTM ...... 32

Water sampling and analyses ...... 32 Sample process prior to pathogen enrichment ...... 33 Enrichment and isolation of presumptive Salmonella spp...... 34 Enrichment and isolation of presumptive Listeria spp...... 34 DNA extraction for PCR confirmation ...... 35 PCR confirmation of Salmonella spp...... 35 PCR confirmation of the general Listeria monocytogenes and Listeria spp...... 35 Enumeration of Escherichia coli and coliforms...... 36 Results and discussion ...... 36 Water quality ...... 36 Pathogen removal in the Penn State Eco-MachineTM ...... 40 Duckweed coliform and E.coli inactivation by drying ...... 41 Indicators for common pathogens ...... 42 Conclusions and future work ...... 43 Acknowledgements ...... 43 References ...... 44 Chapter 4 – Coupling ecological wastewater treatment with production of livestock feed and irrigation water provides net benefits to human health and the environment: a life cycle assessment ...... 47 Abstract ...... 47 Introduction ...... 48 Methods ...... 50 Description of the Penn State Eco-MachineTM and inventory analysis ...... 50 Duckweed characteristics ...... 53 Life cycle impact assessment (LCIA) ...... 54 Sensitivity analysis ...... 54 Results and discussion ...... 54 Midpoint characterization ...... 54 Damage categories ...... 58 Sensitivity analysis of heating/greenhouse requirements ...... 64 Sensitivity analysis on vertical farming options to increase duckweed growth area ...... 65 Conclusions ...... 68 Acknowledgements ...... 69 References ...... 70

Chapter 5 – Duckweed protein supports the growth and organ development of mice: a feeding study comparison to conventional casein...... 74 Abstract ...... 74 Introduction ...... 75 Materials and methods ...... 78 Preparation of duckweed and diet formulation ...... 78 Mouse feeding experiment ...... 79 Statistical analysis ...... 79 Results and discussion ...... 79 Duckweed protein quality ...... 79 Body weight gain and relative organ weights ...... 81 Body weight gain and relative organ weights ...... 82 Food intake and feeding efficiency ...... 86 Conclusion ...... 86 Acknowledgements ...... 87 References ...... 88 Chapter 6 – Conclusions, significance, and future work ...... 93 References ...... 96 Appendix A – Duckweed grown on dairy cow manure - theoretical protein production and nitrogen and phosphorous removal in the Chesapeake Bay Watershed ...... 97 N and P removal from the Chesapeake Bay Watershed ...... 98 Protein requirement met, and N and P removal based on number of animals and growth area ...... 99 Techno-economic analysis of a theoretical 1000 dairy cow herd with 0-20 acres of plant growth area...... 100 Conclusions and future work ...... 101 Supplemental information ...... 104 Appendix B – Chapter 2 additional files ...... 107 Appendix C – Chapter 3 additional files ...... 116 Appendix D – Chapter 4 additional files ...... 117 Appendix E – Chapter 5 additional files ...... 119

vii

LIST OF FIGURES Figure 2.1. (A) Plan-view schematic of the Eco-MachineTM. Red square around the tanks indicates the four tanks from which water was diverted for this experiment (arrows indicate the direction of flow). Green outline on Aerobic 3 tank indicates where duckweed used to inoculate each growth tray was harvested from; (B) Plan-view photograph of the four, hydraulically separated duckweed growth trays. . 17

Figure 2.2. Crude protein content versus ammonium (A), total nitrogen (B), and dissolved oxygen (C) from each duckweed growth tray (n=8 per tray)...... 24

Figure 2.3. Photographs of typical duckweed roots in each growth tray (photo credit: Michael Shreve). 25

Figure 2.4. Comparison of growth rate, crude protein content, and protein yield of duckweed and common land-grown forage crops. (*Kennely et al., 1995; +Masuda & Goldsmith, 2009; Othis study). ... 27

+ - Figure 3.1. Water quality conditions (NH4 , NO3 , TN, and COD) throughout the Penn State Eco- MachineTM (n=18 collected from June 2019 – March 2020) ...... 38

Figure 3.2. Concentration of coliform and E.coli through the Penn State Eco-MachineTM (n=19 for Truck, Clarifier, and Pond; n=5 for UV)...... 40

Figure 3.3. E. coli counts by absence (0) or presence (1) of Salmonella, Listeria spp., and Listeria monocytogenes in three stages of the Penn State Eco-MachineTM (n=19 per sampling location). All samples taken after UV disinfection had no presence of Salmonella or Listeria (not shown)...... 42

Figure 4.1. Schematic of the Penn State Eco-MachineTM and scope of the LCA analysis. Solid arrows indicate the flow of wastewater through the treatment system. Dashed arrows indicate products. The dashed line is the system boundary...... 52

Figure 4.2. Midpoint category impacts of the Penn State Eco-MachineTM. Net impact value per million liter (ML) treated for each category is listed above (positive values represent detrimental impacts and negative values represent beneficial impacts)...... 55

Figure 4.3. Damage assessment of three phases of the Penn State Eco-MachineTM: a) construction; b) operation; c) products; and d) total (see Table 4.1 for inventory; positive values represent detrimental impacts and negative values represent beneficial impacts)...... 63

Figure 4.4. Impacts of the Eco-MachineTM with propane heating, natural gas heating, no heating, and no heating and no greenhouse per million liters (ML) wastewater treated at the four damage categories: a) human health; b) ecosystem quality; c) climate change; and d) resources. Net impact is shown adjacent to the bar in each plot (positive values represent detrimental impacts and negative values represent beneficial impacts).Sensitivity analysis on vertical farming options to increase duckweed growth area ...... 65

Figure 4.5. Vertical farming sensitivity analysis on the impact of the number of vertical trays for duckweed growth in each of the four damage categories: a) human health; b) ecosystem quality; c) climate change; and d) resources. The net impact for each damage category are shown adjacent to the bars (positive values represent detrimental impacts and negative values represent beneficial impacts)...... 67

viii

Figure 4.6. Vertical farming weighted impacts of damage categories based on the number of vertical trays used for duckweed growth. Positive values represent detrimental impacts and negative values represent beneficial impacts...... 68

Figure 5.1. (A) Body mass of mice fed three diets over four weeks: control (casein protein); 10% DWP (casein + 10% duckweed protein); and 25% DWP (casein + 25% duckweed protein). (B) Average growth rate of mice fed three different diets. (C) Organ percentage of body mass from the mice fed three diets after 4 weeks (n=10 per diet; 95% CI)...... Error! Bookmark not defined.

Figure 5.2. A) Food intake rates and B) feeding efficiency of the three diets (n=16 per diet for food intake; n=10 per diet for feeding efficiency; 95% CI)...... Error! Bookmark not defined.

Figure A.1. Theoretical % N and P removed from entering the Chesapeake Bay from manure based on duckweed or soybean growth area (assumes a 7-month growing season for duckweed and one harvest per year for soybean)...... 98

Figure A.2. Comparison of duckweed (DW) and soybean (SB) for fraction of protein requirements met for a 1000 dairy cow herd, with N- and P- uptake from manure fertilizer by both crops based on number of AU per duckweed/soybean growth area (assumes a 7-month growing season for duckweed and one harvest per year for soybean)...... 100

Figure A.3. Duckweed vs. soybean production on equivalent land areas: A) techno-economic analysis; B) fraction of protein requirements met for a 1000 dairy cow herd, with N- and P- uptake from manure fertilizer by the two crops. Analysis assumes a 7-month growing season in NE USA, and a crop value of $360/MT...... 101

Figure B.1. Duckweed protein content results from Cumberland Valley Analytical Services...... 113

LIST OF TABLES

Table 1.1. Amino acid composition (g/100g protein) of various duckweed species from each of the five genera and of common plant-based meals used for livestock and fish...... 6

Table 1.2. In-depth characterization of the nutritional quality of Wolffia microscopia (adapted from Appenroth et al., 2017; FA = fatty acids)...... 6

Table 3.1. Additional water quality parameters for the Penn State Eco-MachineTM from June 2019-March 2020 (n=18 per location ± the standard deviation)...... 39

Table 4.1. Inventory used in the life cycle assessment of the Penn State Eco-MachineTM (positive values represent material consumed and negative values represent material produced)...... 53

Table 4.3. Duckweed vertical growth tray inventory data used in the sensitivity analysis...... 66

Table 5.1. Composition of diets used in this study...... 78

Table 5.2. Proximate analysis and amino acid composition of the duckweed protein isolate used in this study...... 80

Table A.1. Characteristics of dairy cow manure, duckweed and soybean used for this analysis...... 99

Table B.1. Harvested duckweed masses, densities, and growth rates over the course of the experiment...... 108

Table B.2. Duckweed biomass metal concentrations from ICP-MS performed by Penn State Laboratory for Isotopes and Metals in the Environment...... 114

Table C.1. Plate counts of total coliforms, fecal coliforms, and E.coli from duckweed grown in the Penn State Eco-MachineTM at different drying temperatures...... 116

Table D.1. SimaPro Characterization of the Penn State Eco-MachineTM operated for 30 years using IMPACT 2002+ V2.15...... 117

Table E.1. Body weight of mice over the course of the study...... 119

Table E.2. Food intake rate of each group over the course of the study (DWP = duckweed protein) ..... 120

Table E.3. Organ weights of each mouse...... 121

ACKNOWLEDGEMENTS

The financial support of the Office of the Physical Plant, Department of Civil and Environmental

Engineering, and Institutes of Energy and Environment Seed Grant Program at Penn State are gratefully acknowledged.

Mostly, I would like to thank my advisor, Rachel Brennan, for her intellectual and moral support throughout this process. I can confidently say that I would not be here if you hadn’t inspired me to become involved with research at the Eco-MachineTM during the final semester of my Bachelor’s degree.

Thank you.

I would also like to thank my committee members for their patient advice and guidance throughout my graduate research, as well as the faculty, staff, and fellow graduate students of the Civil and Environmental Engineering Department for their support.

xii

Chapter 1 - Introduction Humans and agriculture

It is generally believed that modern humans evolved around 200,000 years ago, and that the population remained low (< 1 million) until the advent of farming around 10,000 BCE. The invention of agriculture was the first catalyst for rapid human population growth, leading to a global population of 170 million by the year 1 CE For the next ~1,800 years, population steadily grew to around 1 billion people. Over the last two centuries, the industrial revolution, improvements in public health, and intensive agriculture (the

‘green revolution’) enabled the human species to sky rocket its population to 7.8 billion people today, and it is projected the global population will reach 11.2 billion by the year 2100 (UN, 2019). Although necessary for supporting the global population, agriculture is arguably the most powerful force that humanity has unleashed onto the planet – roughly one-half of the Earth’s arable land has been transformed for agriculture, and over 75% of that area is used for grazing and growing feed grains for livestock (UN FAO, 2019). Agriculture has many negative effects on the environment, including producing greenhouse gas (GHG) emissions, releasing nutrients into natural water bodies and causing eutrophication, and depleting water and soil nutrients. In fact, food production is responsible for 25% of

GHG emissions globally, and is tied with electricity and heat production as the largest GHG emitter by any one sector (Poore & Nemecek, 2018).

Global food crisis

As population increases and developing regions become more affluent, the demand for all food products will inevitably rise. It is projected that by the year 2027, the demand for food will outweigh the supply, leading to a global shortage of 214 trillion Calories (kcal) per year, which equates to the same energy as

379 billion Big Macs per year – more than McDonald’s has sold in its entire 65-year existence (Gro

Intelligence, 2017). When distributed over the entire global population, this leads to a shortage of about

75 Calories per day per person (~3% of daily recommended Calorie intake). However, the predicted

Calorie shortage will not be spread evenly across the globe. Rapid population growth over the past decades in China, India, and many African countries has caused these nations to have large ‘Calorie gaps’

– the number of Calories these nations demand is much higher than what they produce, forcing them to import food from mainly North America, Europe, and South America. However, the agricultural output of these ‘Calorie rich’ nations is becoming stagnant and will not be able to meet the growing Calorie demand of the developing world. Strategies for combatting this looming food crisis include: changing consumption patterns (eating less meat); reducing food waste; and increasing agricultural yields.

The production of meat, especially beef, is extremely taxing on the environment: if all cattle on

Earth were their own nation, they would be the 3rd largest GHG emitter, only behind China and the United

States (Ranganathan et al., 2016). The obvious solution to reduce the impacts of meat production is to consume less meat. This is starting to occur in developed regions around the world. For example, a study done in the UK found that one-in-eight of their citizens are vegetarian or vegan, and over 60% of those have changed their diet within the last five years (Waitrose and Partners, 2019). Food waste is another issue leading to a global food crisis. In the United States, 30-40% of the food supply is wasted at the retail and consumer levels, corresponding to roughly 130 billion pounds (59 billion kg) and $161 billion worth of food wasted (US FDA, 2020). In fact, if just 15% of the food waste in the United States were recovered, it could feed 25 million people. In addition, food waste is the single largest category of material placed in landfills in the United States, comprising 22% of municipal solid waste (US EPA,

2020), and exacerbates environmental concerns related with landfills, such as GHG emissions, groundwater pollution, and dangers to human health (Lisk, 1991).

While these strategies for combating food shortage seem simple, they require individuals from

Calorie surplus regions to change their behavior on behalf of Calorie deficit regions, which is unlikely to occur and is therefore not practical for overcoming the food crisis. Increasing agricultural yields by commercializing small-scale farms in the developing world (essentially leapfrogging the green revolution reforms in the 1950-1960s), especially in India and Africa, is currently the most realistic solution for overcoming a global food shortage. In addition, developing new technologies/strategies to increase the

2 efficiency of agriculture and livestock production in the developed world will assist in providing food to a growing population.

Combining wastewater treatment with food production

One strategy for increasing food production efficiency is linking wastewater treatment with agriculture by utilizing photosynthetic plants to recover waste nutrients and convert those nutrients into high-protein biomass. This biomass can then be harvested and used for animal feed or crop fertilizer. However, most wastewater treatment systems are not designed to incorporate the production of beneficial by-products.

Conventional wastewater treatment systems use a series of aerated and anoxic tanks to manipulate bacteria to remove organics and nutrients from the water (Holmes et al., 2019). Nitrogen (N) is a major nutrient of concern in wastewater treatment, and entire systems are designed with the goal of removing nitrogen from the wastewater. Urea (CH4N2O) is the main source of nitrogen in wastewater, and it is

+ easily broken down into ammonium (NH4 ) in the early stages of wastewater treatment facilities. The

+ - NH4 is then oxidized to nitrate (NO3 ) in a metabolic process called nitrification that occurs under aerobic conditions. Although there are many different configurations, typically the wastewater is then put into

- anaerobic conditions where the NO3 is converted to nitrogen gas (N2) in a process called denitrification.

The N2 is allowed to escape into the atmosphere, essentially wasting a valuable source of nutrients.

Ecological wastewater treatment systems (e.g., constructed wetlands, Eco-MachinesTM) use the same principles as conventional wastewater treatment, but utilize eukaryotic, multicellular organisms (plants, snails, fish, etc.) in addition to microbes to imitate a natural ecosystem. These systems are capable of treating wastewater to the same extent as conventional systems, but with fewer energy and chemical inputs.

Safety concerns

A major concern surrounding the use of wastewater nutrients for food production is the risk of contamination by wastewater-borne pathogens and metals. Treated wastewater has been effectively used to irrigate crops with no significant health concerns regarding the presence of microbial contamination or

3 metals in the harvested plant tissue (Almuktar et al., 2015; Forslund et al., 2010; Kiziloglu et al., 2008; and Saffari & Saffari, 2013), but little work has been done to examine the risks of using plants grown hydroponically in a wastewater treatment system as a food source. However, pathogens associated with plant biomass have been shown to rapidly inactivate during drying ≥ 60oC (Lee & Kaletunc, 2002), which is already typically done in order to preserve the quality of the biomass. Thus, it is likely that wastewater- grown plants that are dried at ≥ 60oC would have a low risk for microbial contamination. Metals that are present in wastewater (commonly Al, Cd, Cr, Cu, Pb, Hg, Ni, Ag, and Zn) have been shown to concentrate up to 100 or 1000 times into plants, posing a potential health risk in utilizing these plants for food. However, the extent of metal uptake largely depends on the plant species and the growth medium

(Tangahu et al., 2011). Thus, it is necessary to examine specific plants on specific waste streams to accurately characterize potential health risks. After confirming the safety of using a wastewater-grown plant for food, the environmental benefits/impacts of these wastewater-to-food systems should be examined.

Life cycle assessment

Life cycle assessment (LCA) is a technique used to evaluate the potential environmental impacts in each step of a product or process by considering the material inputs, processing, disposal, and emissions/waste that a product or process uses during its life. LCA studies have been conducted on wastewater treatment systems (Dixon et al., 2003; Hospido et al., 2004; Li et al., 2013; and Tabesh et al.,

2019), ecological wastewater treatment systems (Hendrickson et al., 2015), and sustainable protein alternatives (Oonincx and de Boer, 2012; Halloran et al., 2016; Gnansounou and Raman, 2016; and van

Oirschot et al., 2017). The economic and environmental benefits of ecological wastewater treatment systems compared to conventional treatment systems are well documented – Eco-MachinesTM are cost competitive with conventional systems up to 2.3 million L/day (600,000 gal/day) when built in a greenhouse, and up to 3.8 million L/day (1 million gal/day) without a greenhouse, largely due to removing the need for heating (U.S. EPA, 2001). Additionally, an LCA conducted on a Living Machine

(similar to an Eco-Machine) for treating wastewater from an office building found that these ecological systems can reduce GHG emissions and energy consumption by 90% and 10%, respectively

(Hendrickson et al., 2015). To date, no study has been conducted to determine the additional economic and environmental benefits of coupling ecological wastewater treatment with sustainable protein production.

Duckweed as a protein source

One ideal plant for bridging the divide between ‘waste’ nutrients and valuable biomass is duckweed.

Duckweed is a floating aquatic plant that is the anatomically simplest and smallest flowering plant in the world. Duckweeds grow on every continent except Antarctica, and there are 37 species belonging to five genera: Landoltia, Lemna, Spirodela, Wolfia, and Wolffiella (Xu et al., 2014; Cui & Cheng, 2015). All of the species have flat, oval fronds from 1mm-1cm in diameter that float on the surface of the water, and some have root-like structures that assist in stabilizing the plant and obtaining nutrients (Leng, 1999).

They are capable of growing on water depths ranging from only a few millimeters to three meters and can tolerate a large range of water quality conditions. When grown under ideal conditions (namely water temperature, pH, nutrient concentrations, and light intensity), duckweeds can double their biomass in 16-

48 hours, and have a protein content up to 45% by dry matter (DM), making them a model candidate for sustainable protein production.

Duckweed has been studied in depth for its potential to sequester nutrients from waste streams, and its growth rate, protein content, and N and P uptake rates are well documented, establishing that duckweed is capable of producing a vast quantity of protein (Leng, 1999; Cheng et al., 2002; Xu & Shen,

2011; Ge et al., 2012; Xu et al., 2012; Sims et al., 2013; and Yu et al., 2014). The quality of duckweed protein has also been investigated - the amino acid composition of the five duckweed genera are similar, and are comparable to common plant-based meals used for livestock fodder (Table 1.1). Thus, based solely on amino acid composition, duckweed protein appears to be of sufficient quality to be used as a protein supplement for animal feed, and has been characterized by the World Health Organization as a

‘high quality’ protein source (WHO, 2007). In addition to protein, duckweed is a source of fatty acids, macro and micro elements, and phytosterols (Table 1.2).

Table 1.1. Amino acid composition (g/100g protein) of various duckweed species from each of the five genera and of common plant-based meals used for livestock and fish. Duckweed speciesa Common mealsb Amino acid Spriodela Landoltia Lemna Wolffiella Wolffia Soybean Peanut Rice Corn polyrhiza punctata minor hyalina microscopica Cysteine 0.8 1.1 0.9 1.0 1.2 1.9 1.6 1.3 1.7 Methionine 1.6 1.6 1.6 2.0 1.6 1.7 1.0 3.0 2.4 Asparagine 7.8 8.1 8.2 7.3 10.4 ND ND ND ND Threonine 4.3 4.1 4.0 4.2 4.7 3.9 1.6 3.8 3.0 Serine 4.1 4.0 4.1 4.3 4.7 4.2 ND ND ND Glutamine 9.6 9.5 9.8 10.5 10.9 18.4 17.7 ND ND Glycine 4.3 4.5 4.6 5.0 4.7 ND 5.0 ND ND Alanine 5.4 5.3 5.1 6.0 7.8 ND ND ND ND Valine 4.4 4.6 4.6 4.8 4.9 5.3 4.4 6.2 5.1 Isoleucine 3.3 3.5 3.7 3.9 3.7 6.0 4.6 5.2 4.9 Leucine 6.8 7.3 7.3 8.0 7.7 8.0 6.7 8.2 15.3 Tyrosine 3.1 3.1 3.7 3.8 3.3 4.0 4.4 5.7 2.3 Phenylalanine 4.0 4.5 4.4 5.1 4.2 5.3 5.1 5.0 5.6 Lysine 4.2 4.1 5.0 5.8 5.7 6.8 3.0 3.2 1.9 Histidine 1.6 1.6 1.5 1.7 1.7 2.9 2.1 1.7 2.1 Arginine 4.7 4.7 4.8 4.7 5.2 7.3 11.3 7.2 3.3 Proline 3.5 4.1 3.8 3.7 3.6 5.0 ND ND ND Tryptophan ND ND ND ND ND 1.4 1.0 1.3 0.5 aAppenroth et al., 2017, bLi et al., 2011, ND = not determined.

Table 1.2. In-depth characterization of the nutritional quality of Wolffia microscopia (adapted from Appenroth et al., 2017; FA = fatty acids). Fatty acid distribution (%) Macro elements Micro elements Heavy metals Phytosterols (g/kg DW) (mg/kg DW) (mg/kg DW) (50 mg/g lipid extract) Saturated FA 25.1 ± 0.1 Ca 6.0 Fe 240 Cd 0.40 Campesterol 18% Monounsaturated FA 3.8 ± 0.1 K 83 Mn 755 Pb 0.24 Stigamasterol 15% Polyunsaturated FA 71.1 ± 0.2 Mg 3.1 Cu 3.52 Hg 0.04 Sitosterol 53% Sum n3 44.2 ± 0.2 Na 0.3 Zn 30.8 As 0.05 ∆5-Avenasterol 11% Sum n6 26.8 ± 0.0 P 7.0 I 0.75 ∆5,24(25)-Stigmastadienol 1% n6/n3 ratio 0.61 ± 0.0 Se 0.06 Cycloartenol 2%

Many in vivo duckweed feeding studies have been conducted on various animals (dairy cattle, pigs, sheep, goats, poultry, and fish) to determine how duckweed protein compares to conventional feed sources for animal growth and survival (Haustein et al., 1994; Kabir et al., 2005; Ngamsaeng et al., 2004;

Khanum et al., 2005; Moss, 1999; Van et al., 1997; Damry et al., 2001; Reid, 2004; El-Shafai et al., 2004;

Tavares et al., 2008; and Effiong et al., 2009). However, the results from these feeding studies are inconsistent, making it difficult to establish duckweed as a reliable protein source. The confounding results found in these studies are largely due to the wide range of nutritional characteristics that duckweeds can display depending on the nutrient content of their growth medium, with protein contents ranging from 15-45% and fiber content ranging from 5-30% (Leng et al., 1995). In addition, many of these studies had poorly characterized animal diets and duckweed growth conditions, making it impossible to determine the reasons for the inconsistencies observed.

This dissertation addresses the key knowledge gaps preventing the establishment of integrated wastewater-food production systems utilizing duckweed by critically evaluating duckweed grown in a pilot-scale ecological wastewater treatment system (the Penn State Eco-MachineTM) by quantifying protein yields, food safety, economic and environmental impacts, and protein quality. The following four chapters are the foundation of this work, and are briefly described below:

Chapter 2: A beneficial by-product of ecological wastewater treatment: an evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture.

This chapter establishes the quantity of protein that duckweed can produce when grown on wastewater from a pilot-scale ecological wastewater treatment system. In addition, this chapter quantifies metal uptake into wastewater-grown duckweed, and provides a basis for the ideal location to grow duckweed in these types of systems.

This manuscript has been published in Ecological Engineering:

Roman, B., and Brennan, R.A. 2019. A beneficial by-product of ecological wastewater

treatment: an evaluation of wastewater-grown duckweed as a protein supplement for 7

sustainable agriculture. Ecological Engineering: X, 1:100004.

https://doi.org/10.1016/j.ecoena.2019.100004

Chapter 3: Pathogen removal in the Penn State Eco-MachineTM and identification of links between microbiomes and microbiological water safety.

Chapter 3 focuses on pathogen removal throughout the Penn State Eco-MachineTM as well as pathogen removal on wastewater-grown duckweed biomass from drying. This research is a portion of a larger manuscript in progress that involves linking the microbiome of various stages of the Penn State Eco-

MachineTM to the presence/absence of common pathogens found in wastewater.

This manuscript is still in preparation with our collaborators:

Yan, R., Roman, B., Brennan, R.A., and Kovac, J. (2020) Pathogen removal in the Penn

State Eco-MachineTM and identification of links between microbiomes and

microbiological water safety. (In preparation).

Chapter 4: Coupling ecological wastewater treatment with production of livestock feed and irrigation water provides net benefits to human health and the environment: a life cycle assessment

Chapter 4 is a life cycle assessment of the Penn State Eco-MachineTM treating municipal wastewater while producing animal feed derived from duckweed and irrigation water derived from UV disinfected

Eco-MachineTM effluent. The results show that when these systems are operated in a region that does not require heating, the produce beneficial impacts to human health, ecosystem quality, and climate change.

This manuscript has been submitted:

Roman, B., and Brennan, R.A. 2020. Life cycle assessment of an Eco-MachineTM

indicates beneficial environmental impacts for upcycling wastewater nutrients into

protein-rich duckweed. Journal of Environmental Management. (In review).

Chapter 5: Duckweed protein supports the growth and organ development of mice: a feeding study comparison to conventional casein.

The fifth chapter is a feeding study comparing duckweed protein to casein to determine if duckweed protein supports the body growth and organ development of laboratory mice. The results show that when casein was replaced with duckweed protein up to 25%, there were no adverse effects on mice body growth and organ development.

This manuscript has been submitted:

Roman, B., Brennan, R.A., and Lambert, J.D. 2020. Duckweed protein supports the

growth and organ development of mice: a feeding study comparison to conventional

casein. Journal of Food Science. (In review).

Chapter 6: Conclusions, significance, and future work.

The final chapter concludes the dissertation and lays out potential future work, namely that studies should be conducted within agricultural operations to fully realize wastewater-grown duckweed as a protein supplement for livestock.

References

Almuktar, S.A.A.A.N.; and Scholz, M. (2015) “Microbial contamination of Capsicum annuum irrigated with recycled domestic wastewater treated by vertical-flow wetlands.” Ecological Engineering, 82:404- 414. Appenroth, K.J., Sree, K.S., Bohm, V., Hammann, S., Vetter, W., Leiterer, M., and Jahreis, G. 2017. Nutritional value of duckweeds (Lemnaceae) as human food. Food Chemistry, 217:266-273. https://doi.org/10.1016/j.foodchem.2016.08.116 Ashbey, E., and Wangermann, E. 1949. Senescence and rejuvenation in Lemna minor. Nature, 164:187. Cheng, J., Landesman, L., Bergmann, B. a, Classen, J.J., Howard, J.W., Yamamoto, Y.T., 2002. Nutrient Removal from swine lagoon liquid by Lemna minor 8627. Trans. ASAE 45, 1003–1010. Cheng, J.J., Stomp, A.M., 2009. Growing duckweed to recover nutrients from wastewaters and for production of fuel ethanol and animal feed. Clean Soil Air Water, 37:17-26. https://doi- org.ezaccess.libraries.psu.edu/10.1002/clen.200800210 Cui, W., Cheng, J.J., 2015. Growing duckweed for biofuel production: A review. Plant Biol. 17, 16–23. doi:10.1111/plb.12216 Dixon, A.; Simon, M.; and Burkitt, T. (2003) “Assessing the environmental impact of two options for small-scale wastewater treatment: comparing a reedbed and an aerated biological filter using a life cycle approach.” Ecological Engineering, 20:297-308. https://doi.org/10.1016/S0925-8574(03)00007-7 Ellis, E.C., Klein Goldewijk, K., Siebert, S., Lightman, D., & Ramankutty, N. 2010. Anthropogenic transformation of the biomes, 1700 to 2000. Global Ecology and Biogeography, 19(5):589-606. Forslund, A.; Ensink, J.H.J.; Battilani, A.; Kljujev, I.; Gola, S.; Raicevic, V.; Jovanovic, Z.; Stikic, R.; Sandei, L.; Fletcher, T.; and Dalsgaard, A. (2010) “Faecal contamination and hygiene aspect associated with the use of treated wastewater and canal water for irrigation of potatoes (Solanum tuberosum).” Agricultural Water Management, 98:440-450. Ge, X., Zhang, N., Phillips, G.C., Xu, J., 2012. Growing Lemna minor in agricultural wastewater and converting the duckweed biomass to ethanol. Bioresour. Technol. 124, 485–488. Gnansounou, E.; and Raman, J.K. (2016) “Life cycle assessment of algae biodiesel and its co-products.” Applied Energy, 161:300-308. https://doi.org/10.1016/j.apenergy.2015.10.043 Gro Intelligence, 2017. How can we avoid a food crisis that’s less than a decade away. Published donline at gro-intelligence.com. Retrieved from: https://gro-intelligence.com/insights/articles/214-trillion-calories Halloran, A.; Roos, N.; Eilenberg, J.; Cerutti, A.; and Bruun, S. (2016) “Life cycle assessment of edible insects for food protein: a review.” Agronomy for Sustainable Development, 36:57. https://doi.org/10.1007/s13593-016-0392-8 Holmes, D.E.; Dang, Y.; and Smith, J.A. (2019) “Chapter four – nitrogen cycling during wastewater treatment.” Advances in Applied Microbiology, 106:113-192. https://doi.org/10.1016/bs.aambs.2018.10.003 Hendrickson, T.P.; Nguyen, M.T.; Sukardi, M.; Miot, A.; Horvath, A.; and Nelson, K.L. (2015) “Life- cycle energy use and greenhouse gas emissions of a building-scale wastewater treatment and nonpotable

10 reuse system.” Environmental Science and Technology, 49”10303-10311. https://doi.org/10.1021/acs.est.5b01677 Hospido, A.; Moreira, M.T.; Fernandez-Couto, M.; and Feijoo, G. (2004) “Environmental performance of a municipal wastewater treatment plant.” International Journal of Life Cycle Assessment, 9(4):261-271. https://doi.org/10.1007/BF02978602 Kiziloglu, F.M.; Turan, M.; Sahin, U.; Kuslu, Y.; and Dursun, A. (2008) “Effects of untreated and treated wastewater irrigation on come chemical properties of cauliflower (Brassica olerecea L. var. botrytis) and red cabbage (Brassica olerecea L. var. rubra) grown on calcareous soil in Turkey.” Agricultural Water Management, 95:716-724. Lee, J.; and Kaletunç, G. (2002) “Evaluation of the heat inactivation of Escherichia coli and Lactobacillus plantarum by differential scanning calorimetry.” Applied and Environmental Microbiology, 68:5379- 5386. Leng, R.A., 1999. Duckweed: a tiny aquatic plant with enormous potential for agriculture and environment. Food & Agriculture Organization, Animal Production and Health Division. Rome, Italy. Leng, R.A., Stombolie, J.H., and Bell, R. (1995). Duckweed – a potential high-protein feed resource for domestic animals and fish. Livestock Research for Rural Development, 7(1). Li, X.; Rezaei, R; Li, P.; and Wu, G. (2011) “Composition of amino acids in feed ingredients for animal diets.” Amino Acids, 40:1159-1168. Li, Y.; Luo, X.; Huang, X.; Wang, D.; and Zhang, W. (2013) “Life cycle assessment of a municipal wastewater treatment plant: a case study in Suzhou, China.” Journal of Cleaner Production, 57:221-227. https://doi.org/10.1016/j.jclepro.2013.05.035 Lisk, D.J. (1991) “Environmental effects of landfills.” Science of the Total Environment, 100:415-468. Oonincx, D.G.A.B.; and de Boer, I.M.J. (2012) “Environmental impact of the production of mealworms as a protein srource for humans – a life cycle assessment.” PLoS ONE (7)12: e51145. https://doi.org/10.1371/journal.pone.0051145 Ranganathan, J., Vennard, D., Waite, R., Lipinski, B., Searchinger, T., and GlOBAGRI-WRR model authors. 2016. Shifting diets for a sustainable food future. World Resources Institute. Installment 11 of “Creating a Sustainable Food Future. Roman, B., and Brennan, R.A. 2019. A beneficial by-product of ecological wastewater treatment: an evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture. Ecological Engineering: X, 1:100004. https://doi.org/10.1016/j.ecoena.2019.100004 Roser, M. 2013. Future population growth. Published online at OurWorldInData.org. Retrieved from: https://ourworldindata.org/future-population-growth Saffari, V.R.; and Saffari, M. (2013) “Effect of treated municipal wastewater on bean growth, soil chemical properties, and chemical fractions of zinc and copper.” Arabian Journal of Geosciences, 6:4475- 4485. Sims, A., Gajaraj, S., Hu, Z., 2013. Nutrient removal and greenhouse gas emissions in duckweed treatment ponds. Water Res. 47, 1390–1398. doi:10.1016/j.watres.2012.12.009

Tabesh, M.; Masooleh, M.F.; Roghani, B.; and Motevallian, S.S. (2019) “Life-cycle assessment (LCA) of wastewater treatment plants: a case study of Tehran, Iran.” International Journal of Civil Engineering, 17:1155-1169. https://doi.org/10.1007/s40999-018-0375-z Tangahu, B.V., Abdullah, S.R.S.; Basri, H., Idris, M., Anuar, N., and Mukhlisin, M. 2011. A review on heavy metals (As, Pb, and Hg) uptake by plants through phytoremediation. International Journal of Chemical Engineering, 939161. United States Environmental Protection Agency. 2001. The Living Machine® wastewater treatment technology. An evaluation of performance and system cost. Office of Water (4204). EPA-832-R-01-004. United States Environmental Protection Agency. 2020. Sustainable management of food. Published online at epz.gov. Retrieved from: https://www.epa.gov/sustainable-management-food United States Food and Drug Administration. 2020. Food loss and waste. Published online at fda.gov. Retrieved from https://www.fda.gov/food/consumers/food-loss-and-waste van Oirschot, R.; Thomas, J.B.E.; Grondahl, F.; Fortuin, K.P.J.; Brandenburg, W.; and Potting, J. (2017) “Explorative environmental life cycle assessment for system design of seaweed cultivation and drying.” Algal Research, 27:43-54. https://doi.org/10.1016/j.algal.2017.07.025 Waitrose and Partners, 2019. Food and Drink Report 2018-19. Published online at globalagriculture.org. WHO (2007) “Protein and amino acid requirements in human nutrition.” Geneva: WHO. Report of a joint FAO/WHO/UNU Expert Consultation. World Health Organisation Technical Reports Series no. 935. Xu, J., Cheng, J.J., Stomp, A.M., 2012. Growing Spirodela polyrrhiza in Swine Wastewater for the Production of Animal Feed and Fuel Ethanol: A Pilot Study. Clean - Soil, Air, Water 40, 760–765. doi:10.1002/clen.201100108 Xu, J., Shen, G., 2011. Growing duckweed in swine wastewater for nutrient recovery and biomass production. Bioresour. Technol. 102, 848–853. doi:10.1016/j.biortech.2010.09.003 Xu, Y., Ma, S., Huang, M., Peng, M., Bog, M., Sree, K.S., Appenroth, K.J., Zhang, J., 2014. Species distribution, genetic diversity and barcoding in the duckweed family (Lemnaceae). Hydrobiologia 743, 75–87. doi:10.1007/s10750-014-2014-2 Yu, C., Sun, C., Yu, L., Zhu, M., Xu, H., Zhao, J., Ma, Y., Zhou, G., 2014. Comparative analysis of duckweed cultivation with sewage water and SH media for production of fuel ethanol.

Chapter 2 – A beneficial by-product of ecological wastewater treatment: an evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture

Roman, B., and Brennan, R.A. 2019. A beneficial by-product of ecological wastewater treatment: an evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture. Ecological

Engineering: X, 1:100004. https://doi.org/10.1016/j.ecoena.2019.100004

Abstract

Ecological wastewater treatment systems that incorporate aquatic plants (like duckweed) have the potential to recover nutrients and produce high-quality protein, simultaneously alleviating two global issues: hunger and lack of sanitation. Although protein production by duckweed in simple lagoon systems and laboratory trials has been reported, its physiology throughout more complex wastewater treatment systems containing a range of different environmental conditions has not been examined. In this study, a duckweed co-culture (Lemna japonica/minor and Wolffia columbiana) was grown on wastewater from four different stages of a pilot-scale ecological treatment system. Contrary to the literature, the protein content of duckweed did not consistently increase with increasing aqueous nitrogen concentrations, but rather appeared to also be dependent on chemical and microbial interactions. This study indicates that with proper management, duckweed grown in ecological wastewater systems can sustainably produce protein at rates exceeding those of common land-grown forage crops (10.1 tonne ha-1 yr-1).

Introduction

In the developing world, up to 90% of domestic wastewater is discharged without any treatment, releasing valuable nutrients into aquatic systems, leading to poor water quality, eutrophication, and dead zones

(FAO UN, 2015). Often in similar locations, roughly 780 million people suffer from protein-energy undernourishment, due to scarcity of high-quality food (Swaminathan et al., 2012). As human population grows and developing countries become more affluent, the demand for animal-derived proteins is escalating: the Food and Agriculture Organization (FAO) estimates that dairy and meat consumption will increase by 82% and 102%, respectively, between 2000 and 2050 (Boland et al., 2013). In addition, protein-rich fodder to support livestock growth is often a limiting factor for meat production (Van Huis et al., 2013), and many regions with minimal arable land and water scarcity are being forced to import fodder, further decreasing the sustainability of local foods (USDA, 2013). There are also significant environmental concerns associated with the production of animal proteins, including land-use, greenhouse gas emissions, and degraded water quality, which all require careful consideration and proper management (McAllister et al., 2011).

Modern wastewater treatment technologies and intensive farming methods may at first appear to provide a solution to our growing sanitation and protein needs, but these are typically restricted to developed areas, require large amounts of energy, and often release excess nutrients and untreated contaminants into aquatic systems. Ecological wastewater treatment systems (ex., constructed wetlands;

Eco-MachinesTM) utilize diverse life forms, typically housed in a series of ponds or tanks, to clean wastewater to the same, or better, effluent quality as conventional wastewater treatment systems (ex., activated sludge plants). Compared to conventional technologies, ecological wastewater treatment systems have a smaller energy and chemical footprint, making them well suited for use in small or developing communities (US EPA, 2002). Moreover, certain aquatic plants grown in these systems (ex., duckweed) can sequester nutrients from wastewater, producing high-protein biomass that can be harvested and reused in sustainable agriculture.

The aquatic plants of the subfamily Lemnoideae (common name: duckweeds) require only sunlight to treat contaminated water, and simultaneously produce a concentrated source of protein and nutrients that can easily be harvested and reused for food production. Lemnoideae have been shown to grow rapidly on the surface of municipal, dairy, swine, industrial, and aquaculture wastewaters, reducing chemical oxygen demand (COD) and removing substantial amounts of nitrogen (N) and phosphorous (P) (Frederic et al., 2006; Adhikari et al., 2015; Cheng et al, 2002; Chaiprapat et al., 2003). Nitrogen uptake by duckweed is believed to be the most

+ critical factor affecting its protein content, with a preference of ammonium (NH4 ) over nitrate

- (NO3 ) for synthesizing amino acids (Landesman et al., 2005).

The general composition of different duckweed species are quite similar to each other, consisting of macro-nutrients (ex., Ca, Na, K, Fe, Mg) and low fiber and lignin contents (Culley

& Epps, 1973). However, the protein content of duckweed species can range from 15 to 45% by weight, depending on the quality of the water in which they are grown (Chantiratikul et al., 2010).

When grown in nutrient-rich environments like wastewater, the resulting high protein content of

Lemnoideae biomass can yield revenue as fodder for livestock and fish (Ansal et al., 2010;

Mohedano et al., 2012; Fang, 2013; Zhao et al., 2014). Predicting the protein content of duckweed grown in different environments is necessary for effective implementation in various locations with different water quality characteristics. Previous lab and full-scale lagoon studies have examined the effect of nutrient concentrations on duckweed growth and protein content

(Chaiprapat et al., 2003; Cheng & Stomp, 2009; Mohedano et al., 2012); however, an analysis of how duckweed is affected when grown on various stages of a more complex wastewater treatment system has yet to be investigated. The purpose of this study was to examine how N concentrations and N speciation affect the growth and protein content of duckweed grown in four different stages of an ecological wastewater treatment system.

Materials and Methods

Experimental setup

Experiments were conducted in a pilot-scale ecological wastewater treatment system (Eco-MachineTM) located in a 50 m2 greenhouse at The Pennsylvania State University (University Park, PA). This system operates year-round on primary influent municipal wastewater (following rag and grit removal), which is delivered several times per week from the local treatment plant to an underground outdoor holding tank

(11.35 m3, 4 d hydraulic residence time (HRT)). The wastewater is pumped from the holding tank (with a typical dissolved oxygen (DO) < 0.5 mg/L) into the greenhouse intermittently to achieve a flow rate of

2.65 m3/day, and then flows by gravity through components of the system in the following order (Figure

2.1A): closed Anaerobic tank (1.7 m3; 15 hr HRT); closed Anoxic tank (1.7 m3; 15 hr HRT); three open

Aerobic tanks (each 3.79 m3; 34 hr HRT); a clarifier (1.4 m3; 13 hr HRT); a horizontal subsurface flow wetland (44.6 m3; 17 day HRT); and a final display pond (1.8 m3; 16 hr HRT). The clarifier, horizontal subsurface wetland, and display pond were not part of this study. The Anaerobic and Anoxic tanks are identical and closed to the atmosphere. Wastewater from the Aerobic 3 tank, which is high in nitrate, is recycled to the Anoxic tank at 50% of the influent flow rate to achieve denitrification. Aerobic 1 and 2 are aerated periodically and have floating rafts vegetated with mature purple taro plants (Colocasia esculenta). Aerobic 3 is not aerated, and the water surface is covered with duckweed previously identified as a co-culture of Lemna japonica/minor and Wolffia columbiana (Calicioglu and Brennan, 2018).

Peristaltic pumps (Masterflex L/S 07528-10, Gelsenkirchen, Germany) were used to pump wastewater out of the sequential Anaerobic, Anoxic, Aerobic 1, and Aerobic 2 tanks at 27 mL/min (24 hr

HRT) and into the head of four separate shallow growth trays (each 115 x 33 x 10 cm; Figure 2.1B). The same pumps were used to pump wastewater out of the opposite end of each tray – the trays were hydraulically separated and not connected together. The trays were named for the tanks from which their water was supplied. The experiment was operated from January to May 2017, with daylight hours over this period ranging from 10 to 14 hours.

A Recycle Influent from Effluent Holding Anaerobic Anoxic Aerobic 1 Aerobic 2 Aerobic 3 Clarifier to tank Wetland

Figure 2.1. (A) Plan-view schematic of the Eco-MachineTM. Red square around the tanks indicates the four tanks from which water was diverted for this experiment (arrows indicate the direction of flow). Green outline on Aerobic 3 tank indicates where duckweed used to inoculate each growth tray was harvested; (B) Plan-view photograph of the four, hydraulically separated duckweed growth trays.

Duckweed harvesting, growth rate, and protein analysis

To start the experiment, each tray was filled with water from its respective tank, and then one square foot of duckweed harvested from the Aerobic 3 tank was added. A two-week acclimation period (with continuous flow) was allowed before sampling was initiated. During the experiment, half of the duckweed

(based on surface area) was harvested from each tray once per week using a net, rinsed with tap water, and dried at 45oC in an Econotherm Lab Gravity Convection Oven (Precision Scientific, Winchester, IL) until a constant weight was achieved (2 – 3 days). The dried duckweed was stored at room temperature in an air tight desiccator for a maximum of six months, until it was ground into a powder and analyzed for crude protein content (Cumberland Valley Analytical Services, Hagerstown, MD).

Water sampling and analyses

Temperature, conductivity, dissolved oxygen (DO), pH, and oxidation-reduction potential (ORP) within each tray were determined weekly on site using a YSI 556 multi-probe system (YSI Inc., Yellow Springs,

OH). At the same sampling events, water samples were taken from the influent and effluent tubing of each duckweed tray, collected into 50 mL centrifuge tubes, placed on ice in a cooler, and transported to

+ the lab. NH4 was measured using an Orion Star Series portable meter and electrode (Thermo Scientific,

Waltham, MA). Chemical oxygen demand (COD) was measured according to Standard Methods (5220

D; Clesceri et al., 1998). Total nitrogen (TN) was determined using a Shimadzu TOC-VCSH/CSN

- 3- analyzer (Shimadzu, Columbia, MD). After filtering (0.45 μm), NO3 , phosphate (PO4 ), and sulfate

2- (SO4 ) were quantified using a Dionex ICS-1100 ion chromatograph equipped with an AS-18 column and

30 mM potassium hydroxide eluent (Dionex, Sunnyvale, CA). All analyses were performed within two hours of collection.

Metals analysis

Duckweed dried at 45oC for 48 hours was ground using a mortar and pestle and digested in 70%

OmniTrace nitric acid in a MARS 6 microwave digestion system for 25 min (CEM Corporation,

Matthews, NC). Digested samples were diluted to 3% nitric acid and analyzed using inductively coupled plasma mass spectrometry (ICP-MS) at the Penn State Laboratory for Isotopes and Metals in the

Environment (LIME). Metals targeted in the analysis were: Aluminum (Al); Arsenic (As); Cadmium

(Cd); Chromium (Cr); Copper (Cu); Iron (Fe); Lead (Pb); Nickel (Ni); Silver (Ag); and Zinc (Zn).

Statistical analysis

SAS 9.4 was used to conduct a repeated measure analysis of covariance (ANCOVA) with an autoregressive lag 1 covariance structure to determine if a significant difference existed between the duckweed protein content, growth rate, and nutrient removal observed in each tray. Linear regressions were also conducted for duckweed protein content versus ammonia, total nitrogen, and dissolved oxygen in each tray to determine if a significant trend was observed.

Results and Discussion

Wastewater and duckweed biomass characterization

In general, the wastewater characteristics entering the Anaerobic tray (Table 2.1) were similar to that of low-strength domestic wastewater, with the exception of COD. The lower than average COD values entering the Anaerobic tray can be attributed to removal in the holding tank of the Eco-MachineTM: an

18 average of 36.7% COD removal was observed from the wastewater delivery truck to the holding tank, which is comparable to COD removal in septic tanks (US EPA, 2002). Temperature, conductivity, and pH stayed relatively constant through each tray, while DO and ORP were higher in the Aerobic trays due to aeration of their respective tanks, consistent with trends observed in conventional wastewater treatment.

Table 2.1. Average water quality characteristics from the influent, effluent, and within each duckweed growth tray, and duckweed biomass characteristics throughout the experiment (n=18; ± one standard deviation). Trays are named for the Eco-MachineTM tanks from which their flow is derived. Duckweed Growth Trays Parameter Anaerobic Anoxic Aerobic 1 Aerobic 2 Influent Effluent Influent Effluent Influent Effluent Influent Effluent COD (mg L-1) 114 ± 21 72.7 ± 20 70 ± 18 61.7 ± 22 39 ± 9.4 33.7 ± 12 38 ± 14 35.4 ± 15 TN (mg L-1) 32.9 ± 8.1 25.3 ± 7.9 25.7 ± 7.8 20.4 ± 7.2 25.6 ± 7.2 20.6 ± 7.5 22.9 ± 7.1 17.7 ± 7.8 + -1 NH4 -N (mg L ) 24.2 ± 9.1 17.6 ± 7.0 18.3 ± 7.0 12.9 ± 6.7 17.8 ± 6.8 11.0 ± 6.2 9.1 ± 5.4 5.1 ± 4.3 - -1 NO3 -N (mg L ) 0.0 ± 0.0 1.0 ± 1.5 0.0 ± 0.1 2.3 ± 2.0 1.6 ± 1.3 4.7 ± 4.4 8.8 ± 3.9 8.8 ± 4.4 3- -1 PO4 -P (mg L ) 3.2 ± 1.2 3.0 ± 1.2 2.9 ± 1.1 3.1 ± 0.9 2.9 ± 1.0 3.3 ± 0.9 3.3 ± 1.0 3.2 ± 0.7 2- -1 SO4 (mg L ) 18.3 ± 6.9 25.5 ± 5.7 31.7 ± 10.5 29.6 ± 4.8 31.7 ± 4.3 30.3 ± 3.5 30.6 ± 3.7 30.2 ± 3.6

Temp. (oC) 19.6 ± 3.4 19.5 ± 3.1 20.0 ± 3.3 20.0 ± 2.6 Cond. (mS/cm) 2.3 ± 0.9 2.3 ± 0.7 2.2 ± 0.5 2.1 ± 0.5 pH 7.3 ± 0.2 7.4 ± 0.2 7.4 ± 0.2 7.5 ± 0.2 DO (mg L-1) 0.5 ± 0.1 1.0 ± 0.6 1.7 ± 0.9 2.9 ± 1.4 ORP (mV) -197 ± 79 -15 ± 113 110 ± 70 151 ± 37

Duckweed growth rate 8.0 ± 4.0 7.0 ± 3.6 8.2 ± 3.9 6.9 ± 3.5 (g m-2 d-1) Duckweed protein content 37.0 ± 1.9 37.4 ± 2.3 37.0 ± 1.6 36.0 ± 2.2 (% DM) Protein yield 2.9 ± 1.4 2.6 ± 1.2 3.0 ± 1.5 2.5 ± 1.3 (g m-2 d-1)

Fe (mg kg-1) 408 ± 100 397 ± 82 298 ± 66 199 ± 30 Al (mg kg-1) 84.4 ± 24 73.3 ± 32 52.5 ± 33 65.7 ± 64 Zn (mg kg-1) 44.1 ± 8.8 42.2 ± 8.2 49.8 ± 9.8 62.8 ± 23 Cu (mg kg-1) 19.5 ± 8.7 15.5 ± 5.3 12.2 ± 3.4 14.3 ± 3.5 Pb (mg kg-1) 2.7 ± 1.5 2.6 ± 1.5 2.4 ± 1.3 3.1 ± 1.6 Ni (mg kg-1) 2.5 ± 1.5 1.8 ± 0.4 0.99 ± 0.2 0.89 ± 0.3 Cr (mg kg-1) 0.37 ± 0.1 0.36 ± 0.1 0.25 ± 0.1 0.27 ± 0.2 As (mg kg-1) 0.23 ± 0.06 0.25 ± 0.06 0.18 ± 0.05 0.15 ± 0.07 Cd (mg kg-1) 0.03 ± 0.03 0.02 ± 0.01 0.02 ± 0.01 0.03 ± 0.01 Ag within the duckweed biomass was below detection limit for all samples.

Nitrogen removal in the duckweed trays was similar to that measured in other studies (Zhao

+ et al., 2014; Mohedano et al., 2012). As expected, TN and NH4 concentrations decreased from the

- Anaerobic to the Aerobic trays, but interestingly, NO3 production did not correspond directly with

+ NH4 removal (Table 1). Previous studies have shown simultaneous nitrification/denitrification occurring in duckweed growth ponds, and have speculated the occurrence of denitrification in the benthic layer (Zimmo et al., 2003). Although the duckweed trays used in this study did not have a sediment layer, wastewater solids did accumulate on the bottom of each tray over the course of the

- experiment, coincident with decreasing NO3 concentrations in the effluent. These observations suggest that as solids accumulated, an anoxic zone was formed at the bottom of the trays that provided a niche environment for denitrifying bacteria.

TN removal was similar for all trays (20-25%). In ecological wastewater treatment systems, a combination of N transformation mechanisms are responsible for N removal, including:

+ + mineralization of organic nitrogen (ON) to NH4 ; sedimentation of ON; regeneration of NH4 from

+ - sediment; decay of plant biomass to ON; volatilization of NH3; nitrification of NH4 to NO3 ;

- + - denitrification of NO3 to nitrogen gas (N2); plant uptake of NH4 and NO3 ; and microbial uptake of

+ - NH4 and NO3 (Kadlec & Wallace, 2008). It is likely that a combination of these processes was contributing to N removal in the duckweed trays, making it difficult to attribute N removal to any one mechanism; however, the % TN removal by duckweed uptake in each tray was estimated to be:

13.8% for Anaerobic; 16.6% for Anoxic; 16.9% for Aerobic 1; and 13.2% for Aerobic 2, based on the calculations described in Eq. 1 – 3:

푇푁 푟푒푚표푣푒푑 푏푦 푑푢푐푘푤푒푒푑 = ∗ ∗ Eq. 1 .

푇푁 푟푒푚표푣푒푑 푖푛 푡푟푎푦푠 = ∗ − ∗ ∗ Eq. 2

. % 푇푁 푟푒푚표푣푒푑 푏푦 푑푢푐푘푤푒푒푑 = ∗100% = ∗ 100% Eq. 3 . The average TN removal from the influent of the Anaerobic tank to the effluent of the Aerobic 2 tank in the pilot-scale Eco-MachineTM system was 45%, similar to the removal typically observed in free water 21

surface wetlands (US EPA, 2000), suggesting that duckweed can play a significant role in N recovery from these systems.

Phosphate concentrations throughout the duckweed trays showed very little change (Table 2.1).

Duckweed is known to concentrate P up to 1.5% of its dry weight (Leng et al., 1994). However, duckweed is capable of drawing P from its biomass, and has been known to grow on waters devoid of P once it has been accumulated (Leng, 1999). Since the duckweed in this experiment was originally grown on wastewater within the Eco-MachineTM, it is likely that the biomass had already accumulated a sufficient amount of P to continue to reproduce, thus not needing to utilize the P present in the wastewater of the growth trays. Also, it has been shown that P in the biomass of duckweed is highly soluble and is rapidly released into the medium upon death of the plant (Stambolie & Leng, 1994). It is possible that due to the harvesting frequency used within this experiment (7 days), some of the duckweed may have been dying and releasing P back into the wastewater. Further studies examining P removal with varying harvesting frequencies of duckweed grown on domestic wastewater should be conducted.

Sulfate concentrations showed little change from the influent to the effluent of each growth tray, with exception of the Anaerobic tray, which had a significant increase (p=0.002; Table 2.1). It is likely

2- that the increase in SO4 observed within the Anaerobic tray can be attributed to the changing environmental conditions between the Anaerobic Eco-MachineTM tank and the Anaerobic duckweed growth tray. The shallow depth of the tray likely enabled some dissolution of oxygen from the atmosphere into the water, leading to higher DO and ORP values than in the Anaerobic tank. However, the DO and ORP values found in the bulk liquid of the Anaerobic duckweed growth tray are below the

2- values required to oxidize reduced sulfur (S) to SO4 (Bouroushian, 2010). This suggests that duckweed could be providing a niche aerobic environment for sulfur-oxidizing bacteria in its rhizosphere via

2- 2- photosynthesis, enabling the oxidation of S to SO4 and thereby higher concentrations of SO4 in the effluent of the Anaerobic tray than the influent. Little research has been done on the S requirements of

+ duckweed; however, high levels of S-amino acids have been observed when growth rate is high and NH4 in non-limiting (Leng et al., 1994), suggesting that low S concentrations have the potential to limit growth

22 and protein content. However, since S is abundant in wastewaters, it is an unlikely candidate for deficiencies in practical applications.

Metals concentrations within the duckweed biomass from the four growth trays are provided in

Table 1. All metals were below the regulatory limits for fodder set by the European Parliament and US

FDA. Of the analyzed metals, Fe had the highest biomass concentration, ranging from 160 to 600 mg/kg duckweed biomass, which is higher than many other crops used for livestock feed, including barley, corn, soy, and wheat (Skinner & Peterson, 1928). The low availability of Fe in soils is a key limiting factor for crop production in arid and semiarid regions, leading to Fe-deficient crops (Khan et al., 2011); thus, wastewater-grown duckweed could provide a valuable source of Fe-rich fodder in these areas. These results indicate that wastewater grown duckweed meets the requirements for metals in fodder, and may be beneficial in areas with Fe deficiencies.

Duckweed protein trends

Contrary to previous work (Landesman et al., 2005; Leng et al., 1994), which showed that duckweed protein content increased with increasing aqueous N concentrations up to 60 mg N/L, duckweed crude protein content in this study did not consistently increase with increasing N concentrations in the aqueous phase (Figure 2.2A & 2.2B). Instead, differences in duckweed crude protein content and growth rates between trays were statistically insignificant. The average crude protein content and growth rate of duckweed in this study was 36.9 % dry matter (DM) and 27.5 tonne/ha-1 yr-1, respectively (n = 32 duckweed collection events; 8 from each tray). A statistically significant positive correlation between

+ NH4 and crude protein was observed in the Aerobic 1 and Aerobic 2 trays (Figure 2.2A; p ≤ 0.05), while a negative correlation was observed in the Anaerobic tray (Figure 2.2A; p ≤ 0.05). Interestingly, a strong positive correlation between the crude protein content of duckweed and DO in the Anaerobic tray was observed (Figure 2.2C; p ≤ 0.01). Since microbial communities in wastewater change with DO, it is likely that plant-microbe interactions had a substantial effect on duckweed protein content.

45 A Anaerobic Anoxic Aerobic 1 Aerobic 2 Slope = -0.13 Slope = -0.08 Slope = 0.22 Slope = 0.37 42 R2 = 0.31 R2 = 0.72 R2 = 0.80 R2 = 0.77 p = 0.05 p = 0.05 p = 0.01 p = 0.05 39

Crude Protein (% DM) (% Protein Crude 33

30 0 10 20 30 40 50 Ammonium (mg L-1) 45 B Anaerobic Anoxic Aerobic 1 Aerobic 2 Slope = -0.04 Slope = 0.23 Slope = 0.18 Slope = 0.32 2 42 R2 = 0.01 R2 = 0.25 R2 = 0.25 R = 0.48 p = 0.87 p = 0.45 p = 0.17 p = 0.15 39

33 Crude Protein (% DM) (% Protein Crude

30 20 25 30 35 40 45 Total Nitrogen (mg L-1) 45 C Anaerobic Anoxic Aerobic 1 Aerobic 2 Slope = 11.4 Slope = 2.85 Slope = 2.25 Slope = 3.54 42 R2 = 0.67 R2 = 0.47 R2 = 0.47 R2 = 0.59 p = 0.01 p = 0.21 p = 0.09 p = 0.75 39

33 Crude Protein (% DM) (% Protein Crude

30 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Dissolved Oxygen (mg L-1)

Figure 2.2. Crude protein content versus ammonium (A), total nitrogen (B), and dissolved oxygen (C) from each duckweed growth tray (n=8 duckweed collection points per tray).

The rhizosphere of aquatic plants experiences different chemical conditions from the surrounding environment due to a range of processes induced by the plant roots and their associated microbial activity. During photosynthesis, water is converted to oxygen at the roots (the roots of duckweed are photosynthetic), providing a niche environment for ammonia-oxidizing bacteria (AOB), increasing the potential for N transformations, and potentially affecting nutrient uptake. Thick biofilms were visible on the roots of the duckweed harvested from the Anaerobic and Anoxic trays, whereas the duckweed roots from the Aerobic 1 and Aerobic 2 trays did not have a visible biofilm (Figure 2.3). It is speculated that the dense biofilms present on the duckweed roots in the Anaerobic and Anoxic trays may have contained a large population of AOB, which may have out-competed duckweed for the

+ NH4 in the wastewater, resulting in lower than anticipated protein yields. To confirm this hypothesis,

+ the expression of NH4 oxidizing genes from the duckweed rhizospheric microbial communities will be analyzed in future work.

Anaerobic Anoxic Aerobic 1 Aerobic 2

Figure 2.3. Photographs of typical duckweed roots in each growth tray (photo credit: Michael Shreve).

Duckweed protein economic value

Based on the average growth rate and protein content of duckweed measured over this 5-month study, an annual protein yield of 10.1 tonne protein ha-1 yr-1 was estimated as shown in Eq. 4:

, 푝푟표푡푒푖푛 푦푖푒푙푑 = ∗ ∗ ∗ ∗ Eq. 4 ∗ ∗

This protein yield was used to estimate the potential monetary value of duckweed at $4,800 USD ha-1 yr-1 using the current market value of soybean meal ($465/tonne) as a surrogate for duckweed, since it has similar protein quality and content (World Bank, 2018). This duckweed protein yield dwarfs that of common land-grown forage crops (Figure 2.4), while not occupying arable land. The yield of duckweed protein achieved in this experiment can be used as an estimation for full-scale, year-round operation in controlled indoor environments or subtropical or tropical regions (since the Eco-MachineTM is operated at a minimum air temperature of 20oC); however, lower protein yields should be anticipated in temperate regions in which indoor production, or year-round outdoor operation, is not feasible. Thus, incorporating wastewater grown duckweed into animal fodder can simultaneously: 1) improve nutrient removal from wastewater streams; 2) serve as a protein source for supporting livestock growth; and 3) reduce land dependence for forage crops.

Protein yield = Growth rate ∙ Protein content 40 12 Growth rate Protein content Protein yield

35 ) 10 -1 ) ) yr -1

30 -1 yr

-1 8 25

20 6

15 4

10 ha (ton yield Protein & Crude protein (% DM) (% Crude protein & Growth rate (ton ha (ton rate Growth 2 5

0 0 O Oats* Soybeans+ Barley* Maize* Sorghum* Duckweed

Figure 2.4. Comparison of growth rate, crude protein content, and protein yield of duckweed and common land-grown forage crops. (*Kennely et al., 1995; +Masuda & Goldsmith, 2009; Othis study). Conclusion

Recovering nutrients from wastewater to produce protein-rich plant biomass may improve water and food security while simultaneously decreasing nutrient pollution and lowering demands on prime agricultural land. Understanding wastewater characteristics that promote high protein duckweed is imperative for optimizing protein production. This study indicates that, contrary to previous work, N concentrations in actual wastewater cannot be solely used to predict protein content in duckweed biomass, but rather a suite of chemical and microbial interactions must also be understood. The potential uptake of trace organic contaminants into duckweed should also be considered when selecting growth locations, which will be evaluated in future work. Growing duckweed for protein on the end stages of a wastewater treatment system may be a viable approach for reducing the risk of contamination from pathogens and chemicals, while not sacrificing protein yield.

Acknowledgements

The support of the Office of the Physical Plant and the Department of Civil and Environmental Engineering at The Pennsylvania State University are gratefully acknowledged, as is the graduating classes of 2000 and

1950 for gifting the university with the Eco-MachineTM.

References

Adhikari, U., Harrigan, T., Reinhold, D.M., 2015. Use of duckweed-based constructed wetlands for nutrient recovery and pollutant reduction from dairy wastewater. Ecological Engineering 78, 6-14. Ansal, M.D., Dhawan, M., Kaur, V.I., 2010. Duckweed based bio-remediation of village ponds: an ecologically and economically viable integrated approach for rural development through aquaculture. Livestock Research for Rural Development 22, Article #129. Boland, M.J., Rae, A.N., Vereijken, J.M., Meuwissen, M.P.M., Fischer, A.R.H., van Boekel, M.A.J.S., Rutherfurd, S.M., Gurppen, H., Moughan, P.J., Hendrisk, W.H., 2013. The future supply of animal- derived protein for human consumption. Trends in Food Science & Technology 29, 62-73. Calicioglu, O., Brennan, R.A., 2018. Sequential ethanol fermentation and anaerobic digestion increases bioenergy yields from duckweed. Bioresource Technology 257, 344–348. Chaiprapat, S., Cheng, J., Classen, J.J., Ducoste, J.J., Liehr, S.K., 2003. Modeling nitrogen transport in duckweed pond from secondary treatment of swine wastewater. Journal of Environmental Engineering 129, 731-739. Chantiratikul, A., Chinrasri, O., Chantiratikul, P., Sangdee, A., Maneechote, U., Bunchasak, C., 2010. Effect of replacement of protein from soybean meal with protein from Wolffia meal (Wolffia globosa (L). Wimm.) on performance and egg production in laying hens. International Journal of Poultry Science 9, 283-287. Cheng, J., Landesman, L., Bergmann, B.A., Classen, J.J., Howard, J.W., Yamamoto, Y.T., 2002. Nutrient removal from swine lagoon liquid by Lemna minor 8627. American Society of Agricultural Engineers 45, 1003-1010. Cheng, J.J., Stomp, A.M., 2009. Growing duckweed to recover nutrients from wastewaters and for production of fuel ethanol and animal feed. Clean Soil Air Water 37, 17-26. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1998. Standard methods for the examination of qater and qastewater, 20th ed. American Public Health Association, American Water Works Association, Water Environment Federation, Baltimore, MD. Culley, D.D., Epps, E.A., 1973. Use of duckweed for waste treatment and animal feed. Journal of the Water Pollution Control Federation 45 (2), 337-347. Fang, Y., 2013. Wastewater treatment by duckweed and production of high starch biomass: the pilot-scale studies. The Second International Conference on Duckweed Research and Applications, Rutgers, NJ. Food and Agriculture Organization of the United Nations, 2015. The improved global governance for hunger reduction programme. Frederic, M., Samir, L., Louise, M., Abdelkrim, A., 2006. Compreshensive modeling of mat density effect on duckweed (Lemna minor) growth under controlled eutrophication. Water Research 40, 2901- 2910. Henze, M., Comeau, Y., 2008. Wastewater characteristics. Biological wastewater treatment: principles, modelling and design. IWA Publishing, London, UK. Kadlec, R.H., Wallace, S.D., 2008. Treatment wetlands. Taylor & Francis Group, Boca Raton, Florida.

Kennely, J., Okine, E., Khorasani, R., 1995. Barley as a grain and forage source for ruminants. The Western Canadian Dairy Seminar. Khan, Z., Ahmad, K., Kashaf, S., Ashraf, M., Al-Qurainy, F., Danish, M., Fardous, A., Gondal, S., Ejaz, A., Valeem, E., 2011. Evaluation of iron content in a potential fodder crop oat (Avena Sativa L.) grown on soil treated with sugarcane filter cake. Pakistan Journal of Botany, 43(3), 1547-1550. Landesman, L., Parker, N.C., Fedler, C.B., Konikoff, M., 2005. Modeling duckweed growth in wastewater treatment systems. Livestock Research for Rural Development 17 (6). Leng, R.A., Stambolie, J.H., Bell, R.E., 1994. Duckweed a potential high protein feed resource for domestic animals and fish. Improving animal production systems based on local feed resources. 7th AAAP Animal Science Congress, 100-117. Masuda, T., Goldsmith, P.D., 2009. World soybean production: area harvested, yield, and long-term projections. International Food and Agribusiness Management Review 12, 143-162. McAllister, T.A., Beauchemin, T.A., McGinn, K.A., Hao, X.Y., Robinson, P.H., 2011. Greenhouse gases in animal agriculture – finding a balance between food production and emissions. Animal Feed Science and Technology 166, 1-6. Mohedano, R.A., Costa, R.H.R., Tavares, F.A., Filho, P.B., 2012. High nutrient removal rate from swine wastes and protein biomass production by full-scale duckweed ponds. Bioresource Technology 112, 98- 104. Skinner, J., Peterson, W., 1928. The iron and manganese content of feeding stuffs. The Journal of Biological Chemistry 79, 679-687. Swaminathan, S., Vaz, M., Kurpad, A.V., 2012. Protein intakes in India. British Journal of Nutrition 108, 50-58. United States Department of Agriculture, 2013. USDA agricultural projections to 2022. Interagency Agricultural Projections Committee. United States Environmental Protection Agency, 2002. Wastewater technology fact sheet. The Living Machine®. Office of Water. EPA 832-F-02-025. Van Huis, A. 2013. Potential of insects as food and feed in assuring food security. Annual Review of Entomology 58, 563-583. Zhao, Y., Fang, Y., Jin, Y., Huang, J., Bao, S., Fu, T., He, T., Wang, F., Zhao, H., 2014. Potential of duckweed in the conversion of wastewater nutrients to valuable biomass: a pilot-scale comparison with water hyacinth. Bioresource Technology 163, 82-91. Zimmo, O.R., van der Steen, N.P., Gijzen, H.J., 2003. Comparison of ammonia volatilization rates in algae and duckweed-based waste stabilization ponds treating domestic wastewater. Water Research 37, 4587-4594.

Chapter 3 – Pathogen removal in the Penn State Eco-MachineTM and identification of links between microbiomes and microbiological water safety

Yan, R., Roman, B., Brennan, R.A., and Kovac, J. (In preparation).

The following chapter is a part of a collaborative project in which Runan Yan completed the molecular testing of water samples and Ben Roman completed the water quality analyses and duckweed bacterial counts.

Introduction

As global population and food demand increases, there will be a growing need for crops used for animal feed and for irrigation water. However, it is estimated that in some currently irrigated regions around the world (western United States; China; and West, South, and Central Asia), the reversion of 20-60 Mha of irrigated cropland to rainfed management could be necessary due to decreased fresh water availability from climate change (Elliot et al., 2013). Lack of water sources for irrigation is a major factor accounting for the looming global food crisis – one study found that by the year 2027, there could be a global food shortage of 214 trillion Calories (Gro Intelligence, 2017). Recycled water (ex., treated wastewater from sewage, industry, and storm water runoff) is a promising source to supply the growing demand for irrigation in agriculture. However, there are concerns associated with contamination by pathogens that need to be addressed to demonstrate safety for public health.

Ecological wastewater treatment systems (ex., constructed wetlands, Eco-MachinesTM, Living

Machines®) are capable of treating wastewater to the same extent as conventional wastewater treatment systems, but require less energy and chemical inputs (US EPA, 2001). In addition, these systems promote the growth of plants (e.g., duckweed), which can sequester nutrients from the wastewater, and be harvested and used for beneficial purposes, such as protein feed, slow-release fertilizer, or biofuel substrate. These systems can naturally remove 70-80% E.coli (Sheehan, 2012), and the low turbidity of the effluent enables it to be effectively UV disinfected (by solar or UV bulbs), which can reduce E.coli

30 and coliforms levels to below the regulatory limits to be safely used for irrigation. Thus, these systems can produce valuable plant biomass and irrigation water while simultaneously treating wastewater.

Duckweed is a small, floating aquatic plant that has the capability of producing large quantities of protein compared to land grown crops due its prolific growth rate and ability to hyperaccumulate nutrients. Duckweed has been successfully grown on various wastewaters, including those generated by municipalities, industry, and agriculture (Oron, 1994; Al-Nozaily et al., 2000; Monette et al., 2006;

Hammouda et al., 1995; Priya et al., 2012; Adhikari et al., 2015; Cheng et al., 2002; Porath & Pollock,

1982). Duckweed has also been used as a protein supplement for various animals, including cows, pigs, sheep, goat, poultry, and fish (Cheng & Stomp, 2009). However, there are concerns surrounding the safety of using wastewater-grown duckweed as a food source due to potential pathogen contamination.

E.coli and coliforms are currently the recommended microbial indicators of food and water safety hazards, but have been shown to poorly correlate with the presence of foodborne pathogens (McEgan et al., 2013; Draper et al., 2016), allowing for pathogen contamination to slip under the radar even when testing is in place. Pathogen outbreaks in food products not only cause illness and death, but can cost up to $15.6 billion annually in the United States alone (USDA, 2020). Hence, there is a knowledge gap linking easily identifiable safety indicators and complex microbiomes found in natural systems (i.e., agricultural fields, Eco-MachinesTM, etc.), which is of growing importance with the recent necessity of using recycled wastewater in agriculture.

The goal of this study is to contribute to the existing knowledge regarding the safety of using treated wastewater for irrigation by analyzing common wastewater bacteria throughout a pilot scale ecological wastewater treatment system, followed by UV disinfection. In addition, duckweed grown in the system was harvested and dried at different temperatures to observe bacterial inactivation.

Materials and methods

The Penn State Eco-MachineTM

The Penn State Eco-MachineTM is a pilot-scale ecological wastewater treatment system located in a 50 m2 greenhouse at The Pennsylvania State University (University Park, PA). The Eco-MachineTM has a treatment capacity of 3,800 L/day, but was operating at 1,500 L/day during the sampling period of this study (May 2019 – March 2020). The Eco-MachineTM can be separated into three phases of treatment: the holding tank (pseudo septic tank); secondary treatment (follows a modified Ludzack-Ettinger design); and the wetland (horizontal subsurface). Primary influent wastewater from the Penn State Wastewater

Treatment Plant is delivered to an underground holding tank outside of the greenhouse weekly. The wastewater is pumped from the holding tank into the greenhouse intermittently to achieve the desired daily flowrate, and then flows by gravity through the remainder of the system, which includes an anaerobic tank, an anoxic tank, three aerobic tanks, a clarifier, and a horizontal subsurface wetland (see

Chapter 2 for additional details on the Penn State Eco-MachineTM). During the sampling period, a filtration (5 µm polypropylene cartridge filter) and UV disinfection system (PurTest 12 Ultraviolet

Disinfection Unit) was installed to remove the remaining pathogens from the effluent of the Penn State

Eco-MachineTM. Samples were taken from the UV disinfection system every two weeks starting in

January 2020.

Water sampling and analyses

Samples were taken in 50 mL centrifuge tubes from the delivery truck and all stages of the Penn State

Eco-MachineTM, stored on ice, and analyzed within two hours of collection. Temperature, conductivity, dissolved oxygen (DO), pH, and oxidation reduction potential (ORP) was analyzed using a YSI 556 multi-probe system (YSI Inc., Yellow Springs, OH). Chemical oxygen demand (COD) was measured

+ according to Standard Methods (5220 D; Clesceri et al., 1998). Ammonium (NH4 ) was measured using an Orion Star Series portable meter and electrode (Thermo Scientific, Waltham, MA). Total nitrogen

(TN) was determined using a Shimadzu TOC-VCSH/CSN analyzer (Shimadzu, Columbia, MD). After

- 3- 2- filtering (0.45 μm), NO3 , phosphate (PO4 ), and sulfate (SO4 ) were quantified using a Dionex ICS-1100 ion chromatograph equipped with an AS-18 column and 30 mM potassium hydroxide eluent (Dionex,

Sunnyvale, CA).

Duckweed sampling and pathogen analysis

Half of the duckweed grown in the third aerobic tank was harvested once every two weeks, placed in a sterile plastic bag, and transported to the lab on ice. The harvested duckweed was separated into three groups and dried as follows: air dry (room temperature in the laboratory), 40oC, and 60oC. One gram subsamples of the duckweed from each drying temperature were taken periodically over the course of one week, ground using a mortar and pestle, placed into 9 mL of sterile saline, sonicated for 30 minutes, and

0.1 mL of the solution was added to spread plates for total coliform (m-Endo LES agar), fecal coliform

(m FC agar), and E.coli (MacConkey agar; Hardy Diagnostics, Springboro, OH). All plates were spread aseptically in a laminar flow hood. The plates were allowed to incubate at 35oC for 24 hr and the number of colonies recorded.

The following analyses were performed by Runan Yan

Sample process prior to pathogen enrichment

A total volume of 250 mL of water samples collected from the truck, anoxic tank, aerobic 3 tank, clarifier, and pond were used for enrichment of pathogens. All water samples were kept on ice until processed on the same day of collection. Water samples were centrifuged (Beckman Avanti J-26 XPI, Rotor JLA-

10.500) twice at 5,000 G for 20 min at 4°C to pellet bacterial cells. The supernatants were then filtered through 0.45 µm pore-size membrane filters (Thermo Scientific Nalgene, 1452045) under vacuum conditions, so that bacterial cells that were not pelleted could be collected on the filter membrane. To prevent over-drying of the filter membrane, which potentially has negative impacts on the survivability of pathogens, the filtration process was always terminated before the membrane reached visual dryness.

Upon completion of the filtration, the filter membranes were transferred to storage bottles, and enriched together with the cell pellets collected from the corresponding samples.

Enrichment and isolation of presumptive Salmonella spp.

The primary and secondary enrichment of Salmonella was conducted following the Food and Drug

Administration Bacteriological Analytical Manual (FDA BAM) protocol (FDA 2020) with minor modifications. A total volume of 225 mL of buffered peptone water supplemented with novobiocin

(working concentration of 20 mg/L) was added into the bottles where cell pellets and filter membranes were stored (Weller et al. 2020). Following incubation at 35°C for 24 hours, 1 mL and 0.1 mL of the primary enrichments were transferred into 9 mL of tetrathionate (BD, 249120) and 9.9 mL of Rappaport-

Vassiliadis broth (Thermo Scientific™, CM0866B) for secondary enrichment, respectively. Tetrathionate enrichment broth was incubated at both 35°C and 41.5°C, and RV was only incubated at 41.5°C. After 24 h incubation, 10 µL of each secondary enrichment broth was streaked onto both Xylose Lysine

Deoxycholate agar (Hardy Diagnostics CRITERION™, C7321) and Hekton Enteric agar (Hardy

Diagnostics CRITERION™, C5841), followed by incubation at 35°C for 24 h for isolation. Two presumptive Salmonella colonies from each plate were picked and streaked on non-selective Brain heart infusion agar (BD, 241830) and incubated at 35°C for 24 h for DNA extraction.

Enrichment and isolation of presumptive Listeria spp.

The enrichment and isolation of Listeria spp. was conducted following the FDA BAM protocol (FDA

2017) with minor modifications. Briefly, a total volume of 225 mL of buffered Listeria enrichment broth

(OxoidTM, CM0879B) supplemented with sodium pyruvate (working concentration 11.1g/L) were added into the bottles where cell pellets and filter membranes were stored. Following incubation at 30°C for 4 h, a cocktail of antimicrobial supplements containing nalidixic acid, cycloheximide, and acriflavine hydrochloride was added into the enrichments to inhibit microbial competitors, and the working concentration of each antibiotic was 40 mg/L, 50 mg/L and 10 mg/L, respectively. After incubation for an additional 44 h at 30°C, 10 µL of each enrichment was streaked onto both Agar Listeria Ottavani &

Agosti (Biorad, 3563695) and Rapid L’mono (Biorad, 3563694) agars for isolation. After incubation at

37°C for up to 48 h, two colonies of both presumptive L. monocytogenes and other Listeria spp. were

34 streaked onto BHI agar plates incubated at 37°C for 24-48 h. Well isolated colonies from BHI agar were used for PCR confirmation.

DNA extraction for PCR confirmation

One well-isolated colony was selected for PCR confirmation. Briefly, one well-isolated colony was resuspended in 100 µL of nuclease free water, and heated at 95°C for 10 min to lyse cells. The heated suspensions were cooled and then centrifuged at 10,000 G for 5 min to remove cell debris, and 50 µL supernatant was collected and stored at -20°C for PCR amplification.

PCR confirmation of Salmonella spp.

Salmonella spp was confirmed based on the presence of invA gene (Park and Ricke 2015). Briefly, 2.5 µL of DNA template was used in each 25 µL PCR reaction with 8.1 pmol of primers Salm3 (AGC GTA CTG

GAA AGG GAA AG) and Salm4 (ATA CCG CCA ATA AAG TTC ACA AAG) targeting the

Salmonella specific invA gene. A known strain of S. enterica Typhimurium (FSL S5-0536) (Vangay et al.

2013) was used as a positive control and the nuclease free water as negative control. The initial denaturation step at 95°C for 5 min was followed by 35 cycles of amplification (denaturation at 95°C for

90 s, annealing at 60°C for 60 s, and extension at 72°C for 90 s), ending with a final extension at 72°C for

7 min. Successful PCR amplification was confirmed by gel electrophoresis using 1.5% agarose gel, and a band at 389 bp was expected for Salmonella spp. If negative PCR results were obtained, additional presumptive colonies were analyzed until Salmonella spp. were successfully confirmed or all presumptive colonies were identified as non-Salmonella.

PCR confirmation of the general Listeria monocytogenes and Listeria spp.

Confirmation of the general Listeria genus was based on the presence iap gene, and L. monocytogenes was confirmed based on the presence of an additional lmo2234 gene (Chen and Knabel 2007). Briefly, 2.5

µL of DNA template was used in each singleplex 25 µL PCR reaction. Primers iapF (ATG AAT ATG

AAA AAA GCA AC) and iapR (TTA TAC GCG ACC GAA GCC AAC) were used to confirm the general Listeria spp. (8.75 pmol/reaction), and additional primers lmo2234F (TGT CCA GTT CCA TTT

TTA ACT) and lmo2234R (TTG TTG TTC TGC TGT ACG A) were used to confirm L. monocytogenes

(7.5 pmol/reaction). A known strain of L. monocytogenes (PS00010) was used as a positive control and nuclease free water was used as a negative control.

The following touchdown PCR amplification procedure was used: initial denaturation at 95 °C for 15 minutes, 15 cycles of denaturation at 94°C for 1 minute, annealing at 55°C to 51°C for 1 minute with a touch down of 3 cycles per temperature, extension at 72°C for 1 minute. The following 15 cycles started with denaturation at 94°C for 1 minute, annealing at 50°C for 1 minute, extension at 72°C for 1 minute and the final extension at 72°C for 8 minutes. Successful PCR amplification was confirmed by gel electrophoresis using 1.5% agarose gel. A band between 1450 bp to 1600 bp (for iap) was expected for

Listeria spp., and an additional band at 420 bp (for lmo2234) was expected for L. monocytogenes. If negative PCR results were obtained, additional presumptive colonies were analyzed until Listeria spp. were successfully confirmed or all presumptive colonies were identified as non-Listeria.

Enumeration of Escherichia coli and coliforms

The enumeration of E. coli and coliforms was conducted using the 3M Petrifilm E. coli/Coliform Count

Plate (3M Company, 6414) following the manufacture’s instruction. Briefly, water samples were first serially diluted using phosphate buffer saline (containing 137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4 and 2 mM KH2PO4, pH 7.4) and 1 mL of the diluted samples was then plated in duplicate. Only plates of those dilutions yielding 15-150 colonies were enumerated after both 24 and 48 h of incubation at 35°C.

Data were analyzed following the Guidance for Data Quality Assessment (EPA 2000) if below the limit of detection (1 colony forming unit per mL).

Results and discussion

Water quality

+ - TM The water quality performance for COD, TN, NH4 , and NO3 in the Penn State Eco-Machine follows a

+ similar trend to conventional treatment, where the COD is removed throughout the system, NH4 is

- - oxidized to NO3 in the aerobic tanks, and NO3 is reduced to N2 in the anoxic tank and the wetland

(Figure 3.1). The holding tank is not designed to treat wastewater, but naturally the COD is degraded and

+ organic forms of N are degraded into NH4 . The secondary treatment phase (modified Ludzack-Ettinger design) consists of six treatment tanks, and is responsible for the majority of the treatment in the system.

+ TM The TN and NH4 concentrations exiting the clarifier of the Penn State Eco-Machine (40 mg TN/L and

+ 30 mg NH4 /L) are typically higher than what is seen in the effluent of conventional wastewater treatment

+ systems (3 mg TN/L and 1.0 mg NH4 /L; Rohrbacher et al., 2013). This is managed by directing clarifier effluent through a subsurface wetland, where plants and biological growth are intended to assimilate the remaining N (see removal between clarifier and pond in Figure 3.1). This presents a unique opportunity to recover N into duckweed biomass prior to the wetland, and thereby produce a prolific source of protein that can be upcycled into sustainable agricultural operations. During this study, the focus was on pathogen and water quality measurements, not duckweed management, so there was still some residual N in the

+ system effluent. In order to remove the remaining NH4 from the system, duckweed could be harvested more frequently, and/or the aeration in the aerobic tanks could be increased to completely oxidize the

+ TM NH4 to NO3. Additional water quality parameters through the Penn State Eco-Machine are shown in

Table 3.1.

90 600 TN NH4 75 NO3 500 COD

60 400

45 300 COD (mg/L)COD 30 200 Nitrogen species (mg/L) species Nitrogen

15 100

0 0 Truck Anaerobic Anoxic Aerobic 1 Aerobic 2 Aerobic 3 Clarifier Pond

Secondary treatment Holding tank Wetland (Modified Ludzack Ettinger)

+ - Figure 3.1. Water quality conditions (NH4 , NO3 , TN, and COD) throughout the Penn State Eco- MachineTM (n=18 collected from June 2019 – March 2020)

Table 3.1. Additional water quality parameters for the Penn State Eco-MachineTM from June 2019-March 2020 (n=18 per location ± the standard deviation). Parameter Truck Anaerobic Anoxic Aerobic 1 Aerobic 2 Aerobic 3 Clarifier Pond Cl- 360 210 210 220 220 210 300 300 (mg/L) ±490 ±100 ±99 ±93 ±72 ±56 ±190 ±170 2- SO4 30 5.3 4.7 19 23 23 24 38 (mg/L) ±12 ±4.9 ±3.6 ±7.3 ±7.8 ±7.8 ±11 ±28 3- PO4 4.0 5.6 5.6 5.9 5.1 5.1 4.8 3.0 (mg/L) ±1.3 ±1.7 ±1.6 ±1.8 ±1.8 ±1.6 ±2.0 ±1.9 DO 2.3 0.8 0.9 2.3 5.0 1.4 0.9 1.2 (mg/L) ±1.2 ±0.3 ±0.3 ±1.4 ±1.4 ±0.5 ±0.3 ±0.3 Water temp. 20 16 16 16 16 16 17 16 (oC) ±2.1 ±3.3 ±3.3 ±2.7 ±2.7 ±2.8 ±3.7 ±3.9 ORP -66 -270 -240 37 76 17 -30 -75 (mV) ±88 ±54 ±46 ±87 ±24 ±49 ±87 ±99 pH 7.3 6.7 6.9 7.1 7.0 6.9 6.9 6.8 ±0.5 ±0.2 ±0.3 ±0.2 ±0.2 ±0.3 ±0.2 ±0.3

Pathogen removal in the Penn State Eco-MachineTM

Pathogens are removed in ecological wastewater treatment systems through a variety of mechanisms, including sedimentation, natural die-off, temperature, oxidation, predation, unfavorable water chemistry, adhesion to biofilm, mechanical filtration, exposure to biocides, and UV radiation from the sun (Weber &

Legge, 2014). The following mechanisms are believed to be the major forms of pathogen removal in the

Penn State Eco-MachineTM: sedimentation – bacteria will ‘settle’ out of the water and into the sediment at the bottom of the tanks (Karim et al., 2004); oxidation – enteric bacteria are either facultative or obligate anaerobes, and the presence of oxygen creates unfavorable conditions (Vymazal, 2005); and predation – grazing by protozoan ciliates and flagellates (Kadlec & Knight, 1996); and filtration by adsorption of pathogens to the wetland media and biofilms attached to the media (Stott & Tanner, 2005; Williams et al.,

1995; and Wand et al., 2007). This results of this study indicate that the Penn State Eco-MachineTM is capable of reducing Coliform and E.coli by 3-4 log through the secondary treatment and wetland phases

(Figure 3.2). After filtration and UV disinfection, <0.3 log CFU coliform/mL were present, and no E.coli remained.

6 Coliform 5 E.coli 4

log CFU/mLlog 1

-1 Truck Clarifier Pond UV

Wetland Secondary treatment Filtration & (Modified Ludzack Ettinger) UV disinfection

Figure 3.2. Concentration of coliform and E.coli through the Penn State Eco-MachineTM (n=19 for Truck, Clarifier, and Pond; n=5 for UV).

Duckweed coliform and E.coli inactivation by drying

The inactivation rate of total coliform, fecal coliform, and E.coli tested on the duckweed grown in the

Aerobic 3 tank were similar, with air dried samples still having activity after one week, 40oC samples showing no activity after two days, and the 60oC samples showing no activity after only four hours

(Figure 3.3). This result is consistent with previous studies performed on E.coli inactivation by drying of plant biomass, where various plants dried at 60oC showed rapid E.coli inactivation (Lee & Kaletunc,

2002). However, drying at high temperatures induces a risk of deteriorating the nutritional quality of the plant biomass, as well as requiring more energy than a passive drying system. Thus, in order to determine the ideal temperature for drying duckweed, further studies should be conducted with smaller increment temperatures (2-5oC) and to consider the economic and environmental impacts of drying at each temperature.

110

100

90 Coliform E.coli 80 40 C 70 60 C (CFU/mL) 60

E.coli E.coli 50

30 Coliform and Coliform 20

0 0 2 4 6 8 12 24 48 Drying time (hr)

Figure 3.3. Coliform and E.coli inactivation on duckweed biomass over time by drying at 40 and 60 C.

Indicators for common pathogens

The presence of Salmonella and Listeria decreased through the Penn State Eco-MachineTM (Figure 3.4), and after filtration and UV disinfection, no pathogens were observed (not shown in the figure).

Salmonella was present in 100% of truck, 63% of clarifier, 21% of pond, and 0% of UV disinfected samples. Listeria spp. was present in 79% of truck, 58% of clarifier, 37% of pond, and 0% of UV disinfected samples. Listeria monocytogenes was present in 74% of truck, 21% of clarifier, and 0% of pond and UV disinfected samples. All of the samples collected from the UV disinfection unit were below the regulatory limit of 126 CFU E.coli /100mL, further establishing the safety of using UV disinfected effluent from ecological wastewater treatment systems as irrigation water. E.coli concentrations appear to have a relationship with the presence/absence of Salmonella and Listeria (Figure 3.4), especially in the clarifier, where the concentrations of E.coli are generally higher when there is a presence of Salmonella or

Listeria. The only statistical difference between E.coli concentrations in the clarifier was for Listeria spp. for this set of data, but more samples will be collected in the future.

Salmonella Listeria spp. Listeria mono.

0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 Truck Clarifier Pond Truck Clarifier Pond Truck Clarifier Pond

Figure 3.4. E. coli counts by absence (0) or presence (1) of Salmonella, Listeria spp., and Listeria monocytogenes in three stages of the Penn State Eco-MachineTM (n=19 per sampling location). All samples taken after UV disinfection had no presence of Salmonella or Listeria (not shown).

Conclusions and future work

This study evaluated the safety of using UV disinfected Eco-MachineTM effluent for irrigation water and wastewater-grown duckweed as animal feed. The results show that after UV disinfection, E.coli,

Salmonella, and Listeria were below detection in the water, indicating that this effluent is suitable for irrigation water. In addition, wastewater-grown duckweed dried at 40oC and 60oC showed no presence of

E.coli after 48 hours and 4 hours, respectively. Future work evaluating the energy and economic benefits of using more sustainable methods of drying and disinfection (i.e., utilizing solar drying and disinfection) should be investigated. Additionally, Metagenomic DNA is currently being extracted from Penn State

Eco-MachineTM water samples, with the hope of identifying microbiome fingerprints that can serve as early indicators of ecological changes in waters that are associated with increased risk of the presence of microbiological food safety hazards. The metagenomic data will be linked with pathogen data, and an attempt will be made to predict pathogen presence based on microbiome composition.

Acknowledgements

The support of the Institutes of Energy and the Environment Seed Grant Program at The Pennsylvania

State University is gratefully acknowledged.

References

Adhikari, U.; Harrigan, T.; and Reinhold, D.M. (2015) “Use of duckweed-based constructed wetlands for nutrient recovery and pollutant reduction from dairy wastewater.” Ecological Engineering, 78:6-14. Al-Nozaily, F., Alaerts, G., and Veenstra, S. (2000) "Performance of duckweed-covered sewage lagoons - II. Nitrogen and Phosphorous balance and plant productivity." Water Resources, 34, 2734-2741. Chen, Yi, and Stephen J. Knabel. 2007. “Multiplex PCR for Simultaneous Detection of Bacteria of the Genus Listeria, Listeria Monocytogenes, and Major Serotypes and Epidemic Clones of L. Monocytogenes.” Applied and Environmental Microbiology 73 (19): 6299–6304. https://doi.org/10.1128/AEM.00961-07. Cheng, J.; Landesman, L.; Bergmann, B.A.; Classen, J.J.; Howard, J.W.; and Yamamoto, Y.T. (2002) “Nutrient removal from swine lagoon liquid by Lemna minor 8627.” American Society of Agricultural Engineers, 45:1003-1010. Cheng, J.J., Stomp, A.M., 2009. Growing duckweed to recover nutrients from wastewaters and for production of fuel ethanol and animal feed. Clean Soil Air Water 37, 17-26. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1998. Standard methods for the examination of water and wastewater, 20th ed. American Public Health Association, American Water Works Association, Water Environment Federation, Baltimore, MD. Dieter CA, Maupin MA, Caldwell RR, Harris MA, Ivahnenko TI, Lovelace JK, Barber Draper AD, Doores S, Gourama H, LaBorde LF. Microbial Survey of Pennsylvania Elliott, J., Deryng, D., Müller, C., Frieler, K., Konzmann, M., Gerten, D., Glotter, M., Flörke, M., Wada, Y., Best, N., Eisner, S., Fekete, B., Folberth, C., Foster, I., Gosling, S., Haddeland, I., Khabarov, N., Ludwig, F., Masaki, Y., and Wisser, D. 2014. Constraints and potentials of future irrigation water availability on agricultural production under climate change. Proceedings of the National Academy of Sciences of the United States of America, 111:3239-3244. EPA. 2000. “Practical Methods for Data Analysis QA00 UPDATE.” FDA. 2017. “BAM Chapter 10: Detection of Listeria Monocytogenes in Foods and Environmental Samples, and Enumeration of Listeria Monocytogenes in Foods.” Accessed July 13, 2020. https://www.fda.gov/food/laboratory-methods-food/bam-chapter-10-detection-listeria-monocytogenes- foods-and-environmental-samples-and-enumeration. FDA. 2020. “BAM Chapter 5: Salmonella | FDA.” Accessed July 13, 2020. https://www.fda.gov/food/laboratory-methods-food/bam-chapter-5-salmonella. Geological Survey; 2018 p. 76. Report No.: 1441. Available from: http://pubs.er.usgs.gov/publication/cir1441 Gro Intelligence, 2017. How can we avoid a food crisis that’s less than a decade away. Published donline at gro-intelligence.com. Retrieved from: https://gro-intelligence.com/insights/articles/214-trillion-calories Hammouda, O.; Gaber, A.; and Abdel-Hameed, M.S. (1995) “Assessment of effectiveness of treatment of wastewater-contaminated aquatic systems with Lemna gibba.” Enzyme and Microbial Technology, 17:317-323.

Kadlec, R. H., & Knight, R. L. (1996). Treatment wetlands. Boca Raton: Lewis Publishers. Karim, M.R., Manshadi, F.D., Karpiscak, M.M., and Gerba, C.P. 2004. The persistence and removal of enteric pathogens in constructed wetlands. Water Research, 38(7):1831-1837. https://doi.org/10.1016/j.watres.2003.12.029 Lee, J.; and Kaletunç, G. (2002) “Evaluation of the heat inactivation of Escherichia coli and Lactobacillus plantarum by differential scanning calorimetry.” Applied and Environmental Microbiology, 68:5379- 5386. McEgan R, Mootian G, Goodridge LD, Schaffner DW, Danyluk MD. Predicting Salmonella populations from biological, chemical, and physical indicators in Florida surface waters. Appl. Environ. Microbiol. 2013 Jul;79(13):4094–105. Monette, F., Lasfar, S., Millette, L., and Abdekrim, A. (2006). “Comprehensive modeling of mat density effect on duckweed (Lemna minor) growth under controlled eutrophication.” Water Research, 40:2901- 2910. NL, Linsey KS. Estimated use of water in the United States in 2015 [Internet]. Reston, VA: U.S. Oron, G. (1994) "Duckweed culture for wastewater renovation and biomass production." Agricultural Water Management, 26: 27-40. Park, S.H., and S.C. Ricke. 2015. “Development of Multiplex PCR Assay for Simultaneous Detection of Salmonella Genus, Salmonella Subspecies I, Salm . Enteritidis, Salm . Heidelberg and Salm. Typhimurium.” Journal of Applied Microbiology 118 (1): 152–60. https://doi.org/10.1111/jam.12678. Porath, D.; and Pollock, J. (1982) “Ammonia stripping by duckweed and its feasibility in circulating aquaculture.” Aquatic Botany, 13:125-131. Priya, A.; Avishek, K; and Pathak, G. (2012) “Assessing the potentials of Lemna minor in the treatment of domestic wastewater at pilot scale.” Environmental Monitoring and Assessment, 184:4301-4307. Rohrbacher, J.; Bilyk, K.; Pitt, P; Latimer, R.J.; and Matthews, R. (2013) “Successfully reducing effluent total nitrogen using conventional nutrient removal strategies.” Hazen and Sawyer Environmental Consultants. Fairfax County, VA. Sheehan, W. 2012. Optimizing the solar disinfection method to produce potable water from ecologically- treated wastewater using recycled polyethylene terephthalate bottles. Undergraduate honors thesis submitted the Penn State Civil Engineering department. Stott, R., and Tanner, C.C. 2005. Influence of biofilm on removal of surrogate faecal microbes in a constructed wetland and maturation pond. Water Science and Technology, 51(9):315-322. Surface Water Used for Irrigating Produce Crops. J. Food Prot. 2016;79(6):902–12. United States Department of Agriculture. 2020. Cost estimates of foodborne illnesses. Economic Research Service. United States Environmental Protection Agency. 2001. The Living Machine® wastewater treatment technology. An evaluation of performance and system cost. Office of Water (4204). EPA-832-R-01-004.

Vangay, Pajau, Eric B. Fugett, Qi Sun, and Martin Wiedmann. 2013. “Food Microbe Tracker: A Web- Based Tool for Storage and Comparison of Food-Associated Microbes.” Journal of Food Protection 76 (2): 283–94. https://doi.org/10.4315/0362-028X.JFP-12-276. Vymazal, J. 2005. Removal of enteric bacteria in constructed treatment wetlands with emergent macrophytes: A review. Journal of Environmental Science and Health, 40(6):1355-1367. https://doi.org/10.1081/ese-200055851 Wand, H., Vacca, G., Kuschk, P., Kruger, M., and Kastner, M. 2007. Removal of bacteria by filtration in planted and non-planted sand columns. Water Research, 41(1):159-167. https://doi.org/10.1016/j.watres.2006.08.024 Weber, K.P., and Legge, R.L. 2013. Comparison of the catabolic activity and catabolic profiles of rhizospheric, gravel-associated and interstitial microbial communities in treatment wetlands. Water Science and Technology, 67(4):886-893. https://doi.org/10.2166/wst.2012.637 Weller, Daniel, Natalie Brassill, Channah Rock, Renata Ivanek, Erika Mudrak, Sherry Roof, Erika Ganda, and Martin Wiedmann. 2020. “Complex Interactions Between Weather, and Microbial and Physicochemical Water Quality Impact the Likelihood of Detecting Foodborne Pathogens in Agricultural Water.” Frontiers in Microbiology 11 (February). https://doi.org/10.3389/fmicb.2020.00134. Williams, J., Bahgat, M., May, E., Ford, M., and Butler, J. 1995. Mineralisation and pathogen removal in Gravel Bed Hydroponic constructed wetlands for wastewater treatment. Water Science and Technology, 32(3): 49-58. https://doi.org/10.1016/0273-1223(95)00604-4

Chapter 4 – Coupling ecological wastewater treatment with production of livestock feed and irrigation water provides net benefits to human health and the environment: a life cycle assessment

Roman, B., and Brennan, R.A. 2020. (In review). “Coupling ecological wastewater treatment with production of livestock feed and irrigation water provides net benefits to human health and the environment: a life cycle assessment.” Journal of Environmental Management.

Abstract

Conventional wastewater treatment systems have remained largely unchanged for over a century, and require large amounts of energy to remove organic material and nutrients from wastewater. Ecologically- designed wastewater treatment systems (ex., Eco-MachinesTM) are a lower-energy alternative that can treat wastewater to the same extent using an ecosystem of plants and microorganisms, while also providing a carbon sink through photosynthesis. The environmental benefit of Eco-MachinesTM can be theoretically maximized by incorporating hyperaccumulating aquatic plants (ex., duckweed) to facilitate nutrient recovery and conversion into protein-rich biomass, which can then be harvested for a range of agricultural and bioenergy applications. Although it has been established that ecological wastewater treatment systems are more cost- and energy-efficient than conventional wastewater treatment systems, a systematic life cycle assessment (LCA) of an Eco-MachineTM coupled with its beneficial by-products has not been conducted. In this study, a series of LCAs were performed on different operational scenarios for a 1000 gallon per day, pilot-scale Eco-MachineTM that, in addition to producing irrigation-quality water, also produces duckweed biomass for aquaculture. The analysis revealed that Eco-MachinesTM located in warm climates, which do not require a greenhouse or supplemental heating, use approximately a third of the energy and produce half of the greenhouse gas emissions compared to conventional wastewater treatment systems in similar locations, while also providing benefits to human health, ecosystem quality, climate change, and resources. In addition, increasing the growth area for duckweed using vertical

47 farming techniques improves the overall impact of the system. This study suggests that with proper management, ecological wastewater treatment systems that upcycle nutrients and water into beneficial products can provide a net benefit to human health and the environment.

Introduction

Biological wastewater treatment is the most common form of sewage treatment in the developed world, and has gone mostly unchanged since the early 1900s. In a conventional wastewater treatment system, organic material and nutrients are removed from the wastewater by bacteria that grow in a large aerated/mixed tank (activated sludge tank). Although there are many possible configurations, the wastewater typically then proceeds to an anoxic tank, where denitrifying bacteria reduce nitrate to nitrogen gas. Although effective, conventional wastewater treatment systems require large amounts of energy to operate, and essentially waste a valuable source of nitrogen by letting it escape into the atmosphere. In the U.S. alone, wastewater treatment was responsible for 1.8% (69.4 billion kWh) of the total electricity use in 2011 (Copeland and Carter, 2017), and released approximately 600,000 tons of aqueous total nitrogen (TN) and 300,000 tons of nitrous oxide (N2O, a greenhouse gas that is 300 times more potent than CO2) into the environment (Maupin & Ivahnenko, 2011; US EPA, 2012).

As a sustainable, decentralized alternative to conventional wastewater treatment, ecological wastewater treatment systems (ex., Eco-MachinesTM) utilize a series of tanks containing a variety of microorganisms, macroinvertebrates, and plants to treat wastewater and recover nutrients, all with no chemical input. The U.S. EPA estimates that Eco-MachinesTM are operationally cost competitive with conventional systems up to flow rates of 1 million gallons (3,785 m3) per day in warm climates, and

600,000 gallons (2,270 m3) per day in cooler climates that require a greenhouse and supplemental heating

(U.S. EPA, 2002). In addition to direct cost benefits, Eco-MachinesTM can be configured to provide a range of environmental benefits. Although natural predation throughout the system significantly reduces pathogen levels (Sheehan, 2012), additional disinfection (ex., with UV light) can make the water exiting an Eco-MachineTM suitable for irrigation, reducing the need for local freshwater withdrawals. In addition,

48 certain plants grown within the Eco-MachineTM (ex., duckweed) can sequester nutrients from the wastewater and produce protein-rich biomass that can be harvested and used for a range of applications, including feed for livestock or fish (Roman and Brennan, 2019). Since the human demand for meat is projected to rise 74% by 2050 (FAO, 2012), the amount of plant protein required to feed these animals is anticipated to similarly increase. To reduce the negative global impacts from intensive agriculture/aquaculture, additional sources of sustainable plant-based proteins are crucial. For example, nearly 90% of small, ocean-caught fish are used to feed farmed fish in the aquaculture industry (Tacon and Metian, 2009), often requiring more than six pounds of ocean-caught fish to produce one pound of farmed fish, which is both environmentally and economically unsustainable (Marina Aquaculture Task

Force, 2007).

Duckweeds, also known as water lentils, are floating aquatic plants from the subfamily Lemnoideae which require only a few millimeters of water depth to grow and can tolerate a large variety of water quality conditions. When grown in nutrient rich environments like wastewater, duckweed grows rapidly and can obtain dry weight protein concentrations of up to 45% (Leng, 1999). Protein rich duckweed has been effectively used as a feed supplement for dairy cows, pigs, sheep, goats, ducks, chickens, and fish

(Cheng and Stomp, 2009; Ansal et al., 2010; Mohedano et al., 2012; Fang, 2013; and Zhao et al., 2014).

Additionally, duckweed is capable of yielding 5 – 10 times more protein than common land-grown crops used for feed (Roman and Brennan, 2019). Due to its ability to create a vast amount of protein, duckweed has the potential to help fill growing protein demands while simultaneously cleaning the water on which it grows. Despite these benefits, the environmental impacts of producing duckweed on wastewater have not been quantified.

Life cycle assessment (LCA) is a technique to evaluate the environmental impacts of a product or process, considering energy and material inputs, processing or manufacturing, use, disposal, and the emissions and waste that a process or product produces over its lifespan. LCA studies have been conducted on wastewater treatment systems (Dixon et al., 2003; Hospido et al., 2004; Li et al., 2013; and

Tabesh et al., 2019) and sustainable protein alternatives, such as insects, seaweed, and algae (Oonincx and

49 de Boer, 2012; Halloran et al., 2016; Gnansounou and Raman, 2016; and van Oirschot et al., 2017). One study analyzed a Living Machine® (similar to an Eco-MachineTM) for treating domestic wastewater from an office building, and found that these decentralized systems can reduce total greenhouse gas (GHG) emissions by almost 90% and total energy consumption by 10%, when compared to conventional wastewater treatment systems (Hendrickson et al., 2015). To date, no LCA studies have been conducted on coupling an ecological wastewater treatment system with protein production.

The aim of the LCA conducted in this study was to identify and quantify the environmental impacts of operating a pilot-scale Eco-MachineTM treating municipal wastewater while concomitantly producing animal feed (derived from duckweed) and irrigation water (derived by UV disinfection of the treated water). Although this study evaluates one pilot-scale system with a discrete set of products, this analysis is a necessary first step toward understanding where environmental impacts in these systems originate, so that future systems can be designed and operated to maximize environmental sustainability. Additional studies would need to be conducted to validate these results for full-scale systems.

Methods

Description of the Penn State Eco-MachineTM and inventory analysis

The model system utilized for this study is a pilot-scale Eco-MachineTM with a capacity of 3,785 L day-1 that is located at The Pennsylvania State University (University Park, PA, USA) campus. Since this location is within a temperate forest biome that experiences freezing temperatures during part of the year, and the vegetation within the system is mostly tropical, the Penn State Eco-MachineTM is sheltered within a greenhouse which receives supplemental heat via a propane-powered furnace when temperatures drop below 18 oC. A solar power array located outside of the greenhouse, consisting of ten 175 W panels on a tracker system that follows the sun’s position throughout the day (Sharp NT-175UC1 panels; Zomeworks

UTRF 168 solar tracker), provides nearly 80% of the electricity used within the system. Routine operation of this facility includes weekly delivery of municipal wastewater from the Penn State Wastewater

Treatment Plant following rag and grit removal (i.e., primary influent). Wastewater is delivered to an

50 outdoor underground holding tank, from which wastewater is pumped into the greenhouse, where it is treated through a series of six tanks and a subsurface wetland, before passing through a UV disinfection unit. The system is configured as a Modified Ludzack-Ettinger (MLE) process, where the anoxic tanks are located upstream of the aerobic tanks (Figure 1), and a portion of the nitrate that was converted from ammonia in the aerobic tanks is recycled back to the anoxic tanks for denitrification. A subsurface wetland inside the greenhouse removes the remaining BOD and nutrients. Although natural predation in the Eco-MachineTM has been shown to remove E.coli by 70% (Sheehan, 2012), a UV disinfection unit was added to remove remaining bacteria from the system effluent, to ensure they are below the regulatory limits of 126 CFU/100mL for irrigation water (FDA, 2016).

Duckweed harvested from the Penn State Eco-MachineTM has been tested as a source of protein for animal fodder (Roman and Brennan, 2019), a substrate for bioethanol and biomethane production

(Calicioglu et al., 2019), and as a slow-release sustainable soil amendment/fertilizer (Kreider et al., 2019).

Of these options, using duckweed as a protein supplement has been calculated to have the highest monetary value on today’s markets (Calicioglu, 2019). In this study, the environmental benefits of using the duckweed grown in the Eco-MachineTM as a source of tilapia feed were analyzed, since the need for sustainable aquaculture feed is well established, and duckweed has been previously shown to be an effective supplement for tilapia in various studies (Gaigher et al., 1984; Moreau et al., 1986; and Fasakin et al., 1999). Data collected from the Penn State Eco-MachineTM from June 2016 to May 2020 was used for the analysis.

The LCA of the Penn State Eco-MachineTM inventory was broken into three phases: construction; operation; and products (Table 1). The construction phase consists of six HDPE tanks used for treatment; a concrete underground holding tank; a greenhouse, which includes a block foundation, glass windows, and steel beams; gravel and HDPE liner for the wetland; piping and valves; solar panels; and a UV disinfection unit. The operation phase consists of: electricity used for aeration, pumping, air humidification, and UV disinfection; and propane used to heat the greenhouse in the winter months.

Products include irrigation water (UV disinfected treated wastewater); duckweed used for tilapia feed; and treated wastewater.

Table 4.1. Inventory used in the life cycle assessment of the Penn State Eco-MachineTM (positive values represent material consumed and negative values represent material produced). Phase IMPACT 2002+ database item Amount Construction Treatment tanks 6 HDPE tanks Polyethylene, high density, granulate, recycled {US} 322.9 kg 1 concrete tank Concrete, 20MPa {GLO} 4.5 m3 Greenhouse Block foundation Concrete block {GLO} 19,320 kg Glass windows Flat glass, uncoated {GLO} 1,600 kg Steel beams Steel, unalloyed {GLO} 1,570 kg Wetland Gravel Gravel, round {GLO} 45,740 kg Liner Polyethylene, high density, granulate {GLO} 2,596 kg Piping and valves Polyvinylchloride, suspension polymerized {GLO} 168.3 kg Solar panels Solar glass, low-iron {GLO} 172.4 kg UV disinfection Ultraviolet lamp {GLO} 30 p Operation Aeration Electricity, low voltage {NPCC, US only} 2,450 kWh/yr Pumping Electricity, low voltage {NPCC, US only} 115.7 kWh/yr Climate control Furnace Propane {GLO} 2,010 kg/yr Humidifier Electricity, low voltage {NPCC, US only} 40 kWh/yr Wastewater delivery Diesel {GLO} 235 kg/yr Solar panel Electricity, low voltage {NPCC, US only} -2,200 kWh/yr UV disinfection Electricity, low voltage {NPCC, US only} 36.5 kWh/yr Products Irrigation water Irrigation {US} 1,380 m3/yr Duckweed Tilapia feed, 24-28% protein {GLO} 38.3 kg/yr Wastewater treatment Wastewater, unpolluted, from residence {GLO} -1,380 m3/yr GLO = global US = United States NPCC = Northeast Power Coordinating Council

Duckweed characteristics

Duckweed grown in the Penn State Eco-MachineTM was previously reported to have an average growth rate of 7 g m-2 day-1 and a protein content of 38% (Roman and Brennan, 2019). For the LCA, a growth area of 44.5 m2 (the size of the wetland in the Penn State Eco-MachineTM) was assumed available to grow duckweed year-round. Tilapia feed (24-28% protein content) was used as a proxy for the duckweed in the

LCA since duckweed has been shown to be an effective supplement to tilapia diets, where replacing 30% of conventional fishmeal with duckweed has been shown to have no effect on tilapia growth (Fasakin et al., 1999). Thus, the duckweed produced in the Eco-MachineTM was assumed to provide 30% of tilapia

53 feed (i.e., one-tonne dried duckweed  0.3 tonne tilapia feed) as the output for this assessment. It was assumed that aquaculture operations would be co-located with the Eco-Machine, therefore, downstream duckweed processing such as packaging, transportation, etc. was not included in this study.

Life cycle impact assessment (LCIA)

SimaPro 9.0 was used for the life cycle inventory analysis and impact assessment calculations, coupled with the EcoInvent 3.0 database. Impact 2002+ was used as the methodology for the impact assessment, quantifying four damage categories: human health; ecosystem quality; climate change; and resources.

Within these damage categories, 15 midpoint categories were quantified: carcinogens; non-carcinogens; respiratory inorganics; ionizing radiation; ozone layer depletion; respiratory organics; aquatic ecotoxicity; terrestrial ecotoxicity; terrestrial acidification/nutrification; land occupation; aquatic acidification; aquatic eutrophication; global warming; non-renewable energy; and mineral extraction (Humbert et al., 2012).

Sensitivity analysis

A Monte-Carlo analysis was performed within SimaPro to determine the uncertainty in the inventory data for the damage categories (95% CI). A sensitivity analysis was performed to determine how a greenhouse and heating affect the overall impacts of the system. Additionally, using vertical farming techniques to increase the duckweed growth area and yield was investigated.

Results and discussion

Midpoint characterization

The midpoint impact characterization for the individual components of the Penn State Eco-MachineTM are shown in Figure 4.2. The detrimental impacts (positive values) outweighed the beneficial impacts

(negative values) in the majority of midpoint categories, except carcinogens, non-carcinogens, land occupation, and mineral extraction.

Detrimental impacts

As the pilot system in this study is currently operated, including a greenhouse and supplemental heating, the major detrimental impacts are from climate control, aeration, and construction of the greenhouse/wetland. Some of these impacts, however, could be mitigated with changes in design or operation.

Climate control is responsible for the largest detrimental impact for 10 midpoint categories: respiratory inorganics (56.5%); ozone layer depletion (77.8%); respiratory organics (64.7%); aquatic ecotoxicity (59.0%); terrestrial ecotoxicity (58.7%); terrestrial acidification/nutrification (55.6%); aquatic acidification (64.8%); aquatic eutrophication (68.7%); global warming (54.6%); and non-renewable energy (67.5%). Climate control in the Penn State Eco-MachineTM is provided by a propane furnace and a humidifier; there is no air conditioning. Propane used to heat the greenhouse is responsible for nearly all of the detrimental impacts from climate control. The Eco-MachineTM had an average propane usage of

1,077 gallons per year (based on data collected from 2018-2020). Operating Eco-MachinesTM (and other indoor ecological wastewater treatment systems) in climates with moderate temperatures year-round would remove the need for heating, and greatly reduce the detrimental impacts.

Aeration is responsible for the largest detrimental impact for three midpoint categories: non- carcinogens (14.1%); ionizing radiation (62.5%); and mineral extraction (25.5%). Aeration is provided to the aerobic tanks by two air compressors that each operate for six hours per day, consuming nearly 2,500 kWh of electricity per year. This result is consistent with conventional wastewater treatment systems, where aeration typically accounts for over 75% of the total energy used (Cantwell et al., 2017). Strategies for reducing the energy cost for aeration in Eco-MachinesTM include: installing fine-bubble diffusers and reducing fouling; installing mechanical mixers in the aerobic tanks to increase oxygen use efficiency; and adding a trickling system that allows the wastewater to pass over shallow steps to naturally aerate between tanks.

The construction of the greenhouse and wetland is responsible for the largest detrimental impact for the final midpoint category: carcinogens (38.5%). This study assumed that raw materials were used to build the greenhouse/wetland; however, recycled/reclaimed materials could be at least partially used to construct such systems (ex., recycled concrete blocks for the foundation, recycled gravel for the wetland, etc.). In addition, a greenhouse would not be necessary if the system were located in a tropical/equatorial region, which would greatly reduce construction impacts.

Although not a leading detrimental impact in any category, wastewater delivery is responsible for

>5% of the total impact in most midpoint categories. Currently, a 2000 gallon (7.6 m3) tanker truck is used to transport primary influent wastewater one mile (1.6 km) from the wastewater treatment plant to the Eco-MachineTM. Ideally, wastewater would be piped directly to the Eco-MachineTM, but shallow bedrock on the site prohibited the construction of underground sewage pipes. In practice, an Eco-

MachineTM should be connected directly to the sewage pipes to avoid having to transport wastewater to the system by truck.

The remaining detrimental impacts stem from the treatment tanks, piping and valves, pumps, and UV disinfection, which collectively account for less than 5% of the total impacts for all but two midpoint categories: aquatic ecotoxicity (5.8%); and terrestrial ecotoxicity (8.6%).

Beneficial impacts

The beneficial impacts from the Eco-MachineTM in this study result from the renewable electricity produced by the solar panels, wastewater treatment, irrigation water (treated wastewater), and tilapia feed

(duckweed) produced within the system.

The integrated solar array is responsible for the largest beneficial impact for seven midpoint categories: ionizing radiation (-56.1%, by offsetting other energy producing processes that emit radiation like nuclear and fossil fuels); ozone layer depletion (-9.0%); aquatic ecotoxicity (-19.4%); terrestrial acid/nutri (-21.7%); aquatic acidification (-14.7%); global warming (-22.1%); and non-renewable energy

(-14.6%). The solar array provides roughly 80% of the total electricity used by the Eco-MachineTM annually, however, during the summer months, the solar panels provide more electricity than the Eco-

MachineTM consumes, and the additional electricity is sent to the university grid. Since this relatively small solar panel system provides the majority of the electricity needed throughout the year, it suggests that Eco-MachinesTM could be easily designed to operate with only electricity produced by a larger solar array, greatly reducing their long-term detrimental impacts.

Wastewater treatment is responsible for the largest beneficial impact for four midpoint categories: carcinogens (-65.8%); non-carcinogens (-76.7%); terrestrial ecotoxicity (-33.3%); and mineral extraction

(-51.5%). This study assumes that if not treated by the Eco-MachineTM, the wastewater would otherwise remain untreated to fairly compare this system to other wastewater treatment systems. From a global perspective, this is realistic, as approximately 80% of the world’s wastewater is discharged without treatment into the environment (NRDC, 2018). Ecological wastewater treatment systems are ideal for addressing this problem, especially in developing regions, since they require little energy and infrastructure to operate.

Irrigation water produced by the Eco-MachineTM is responsible for the largest beneficial impact for three midpoint categories: respiratory inorganics (-32.2%); respiratory organics (-15.4%); aquatic eutrophication (-17.9%).

Tilapia feed is responsible for the largest beneficial impact for one midpoint category: land occupation (-55.1%).

Damage categories

The damage category impacts (human health, ecosystem quality, climate change, and resources) caused by the Penn State Eco-MachineTM were grouped into three phases to facilitate the analysis and discussion that follows: construction (Figure 4.3a), operation (Figure 4.3b), and products (Figure 4.3c).

Human health

The human health damage category is the sum of the following midpoint categories: carcinogens; non- carcinogens; respiratory inorganics; ionizing radiation; ozone layer depletion; and respiratory organics.

Human health impacts are expressed in “Disability-Adjusted Life Years” (DALY), which characterizes

58 disease severity, accounting for both mortality and morbidity. Human health is dominated by respiratory effects caused by inorganic substance emitted into air (Humbert et al., 2012).

The net impact to human health from the Penn State Eco-MachineTM is 1.69E-4 ± 3.86E-4

DALY/million liters (ML) treated wastewater (Table 4.2). The near-zero value indicates that the system has little to no detrimental impacts on human health. Although minimal, the largest of the detrimental impacts to human health are caused by climate control (53.5%), followed by aeration (24.1%), and greenhouse construction (8.4%). The largest beneficial impacts to human health are from wastewater treatment (-36.2%), irrigation (-31.5%), and the solar panels (-21.4%).

For the Eco-MachineTM, the propane consumed, and thus the detrimental impacts from natural gas processing and oil refining, is the largest contributor to detrimental human health effects. Electricity consumption, and thus the emissions from the power plants producing the electricity, is the other largest contributor to the detrimental human health effects. However, since the solar array produces about 80% of the total electricity consumed by the system, the detrimental human health impacts from producing electricity in a power plant are largely offset. Additionally, the energy required to treat wastewater in a conventional plant is not needed if wastewater is treated in the Eco-MachineTM, creating a beneficial impact to human health. Lastly, utilizing the irrigation water produced by the Eco-MachineTM on nearby farmland removes the detrimental impacts of the infrastructure and energy needed to pump water from a surface/ground water source for irrigation.

Ecosystem quality

The ecosystem quality damage category is the sum of the midpoint categories: aquatic ecotoxicity; terrestrial ecotoxicity; terrestrial acidification/nutrification; land occupation; aquatic acidification; and aquatic eutrophication. Ecosystem quality impacts are expressed in “Potentially Disappeared Fraction of species over a certain amount of m2 during a certain amount of year” (PDF-m2-yr). Ecosystem quality is dominated by terrestrial ecotoxicity and land occupation (Humbert et al., 2012).

The net impact to ecosystem quality from the Penn State Eco-MachineTM is 170 ± 480 PDF-m2-yr/ML

(Table 4.2). The largest detrimental impacts to ecosystem quality are from climate control (58.7%),

59 aeration (18.4%), and treatment tanks (6.6%). Similar to human health, the largest beneficial impacts to ecosystem quality are from wastewater treatment (-31.8%), irrigation water (-27.6%), and the solar panels

(-16.4%).

Climate change

The climate change damage category includes one midpoint category: global warming. Climate change impacts are expressed in kg CO2-eq and are dominated by greenhouse gas emissions (Humbert et al.,

2012).

TM The net impact to climate change from the Penn State Eco-Machine is 958 ± 114 kg CO2-eq/ML

(Table 4.2). The largest detrimental impacts to climate change are from climate control (54.6%), aeration

(24.9 %), and greenhouse construction (7.0%). The largest beneficial impacts to climate change are from solar panels (-22.1%), irrigation (-15.9%), and wastewater treatment (-14.9%). This analysis likely underestimates the beneficial impacts to climate change from the Eco-MachineTM, as photosynthesis by plants in the system was neglected. For example, Taro (Colocasia esculenta) grows in the aerobic tanks, and is known to sequester carbon at a rate of 10.4 ± 5.2 g C/m2-day (Saunders et al., 2012). Given an estimated surface area of 2.33 m2 for Taro in the Eco-MachineTM, it could theoretically sequester ~24 g

C/day, or 8.8 kg C/year, which would offset approximately 0.7% of the total CO2 emissions from operating the Eco-MachineTM for one year (Belles, 2020). If C sequestration from all the plants in the

TM Eco-Machine were included, it would likely reduce the CO2 emissions from the system by less than 5%

(Belles, 2020). Frequent harvesting of duckweed in a vertical farming system would likely increase C uptake, however, so beneficial impacts to climate change in that scenario would require further investigation.

Consistent with human health and ecosystem quality, over half of the detrimental impacts on climate change are from the oil refining and processing required to create the propane which is burned to heat the greenhouse in the winter. Replacing propane with natural gas would reduce the detrimental impacts in the climate change category by nearly 70%.

Although operating the Eco-MachineTM in a temperate zone with a propane furnace results in net detrimental impacts to climate change, it is still only about a third of the climate change impacts reportedly produced by the operation of a conventional wastewater treatment plant (3,000 kg CO2-eq/ML;

Hendrickson et al., 2015). However, the conventional wastewater treatment study includes emissions of

GHG (NOx, CO2, and CH4) that are biologically produced during the treatment process, which was not accounted for in the Eco-MachineTM analysis conducted here. It has been estimated that GHG emissions from microbial mediated reactions account for 2-5% of the total GHG emissions from wastewater treatment (Campos et al., 2016). Thus, it is expected that if the GHG production by microbial degradation of N and COD were accounted for in the Eco-MachineTM, the results would only increase slightly.

Resources

The resources damage category is the sum of non-renewable energy and mineral extraction midpoint categories. Resources impacts are expressed in Megajoules (MJ) and are largely dominated by non- renewable energy consumption (Humbert et al., 2012).

The net impact to resources from the Penn State Eco-MachineTM is 85,700 ±13,900 MJ/ML (23,800 kWh/ML; Table 4.2). Climate control (67.5%) is responsible for over half of the detrimental impact, followed by aeration (16.3%), and wastewater delivery (8.6%). The solar array (-14.6%) provides the largest beneficial impact to resources, followed by irrigation water (-5.4%), and wastewater treatment (-

3.5%).

As currently operated, the detrimental impact to resources from the Penn State Eco-MachineTM

(23,800 kWh/ML) is 4-5 times higher than the impact to resources from conventional wastewater treatment (~5,000 kWh/ML; Hendrickson et al., 2015). This is for two reasons: 1) the furnace used to heat the greenhouse in the winter currently consumes propane; and 2) biogas produced during conventional treatment is typically recovered and used to heat various treatment processes throughout the treatment plant (Noyola et al., 2005). In its current configuration, the Penn State Eco-MachineTM does not produce a significant quantity of biogas as it does not contain an anaerobic digester. Methane is below detection in the tanks, so it is likely that any methane production occurring deep within the tanks is consumed by

61 methanotrophic organisms in the upper water layers that are exposed to oxygen through diffusion. For methane capture to be possible, either an anaerobic digester would need to be added to the system, or the anaerobic tanks would need to be redesigned to avoid aerobic zones in the upper layers.

When the need for heating the greenhouse with propane is removed, the impact to resources from the

Eco-MachineTM is reduced to 2,750 kWh/ML, or about half of the impacts from conventional wastewater treatment.

Table 4.2. Damage assessment results of the three phases of the Penn State Eco-MachineTM per million liters (ML) treated. The Monte-Carlo analysis standard deviation is shown for each phase at each damage category (95% CI; positive values represent detrimental impacts and negative values represent beneficial impacts). Damage category Construction Operation Products Total Human health 3.1E-4 ±2.4E-5 1.2E-3 ±2.2E-4 -1.4-3 ±2.9E-4 1.7E-4 ±3.9E-4 (DALY/ML) Ecosystem quality 110 ±20 500 ±370 -470 ±64 170 ±480 (PDF-m2-yr/ML) Climate Change 320 ±14 1,300 ±90 -690 ±70 960 ±110 (kg CO2 eq./ML) Resources 7,300 ±220 87,000 ±14,000 -10,000 ±1,000 86,000 ±14,000 (MJ/ML)

Sensitivity analysis of heating/greenhouse requirements

Based on the results of the damage assessment of the Penn State Eco-MachineTM, it is obvious that removing the need for housing the system within a heated greenhouse would greatly reduce its detrimental impacts. A comparison of the impacts from four scenarios are shown in Figure 4.4: propane heated greenhouse, natural gas heated greenhouse, non-heated greenhouse, and no heat and no greenhouse. The largest reduction in detrimental impacts comes from the need to heat the greenhouse, which transforms the Eco-MachineTM from producing detrimental impacts at all four damage categories to producing a beneficial impact (negative value) in the human health, ecosystem quality, and climate change damage categories (Figure 4.4). Changing from propane to natural gas heating reduces the detrimental impacts for ecosystem quality and climate change, but increases the detrimental impacts for human health and resources. IMPACT 2002+ offers a ‘weighted’ single score impact, that accounts for the overall importance of each damage category and weights them accordingly, allowing for the four damage categories to be compared with the same units. When the Eco-MachineTM heated with propane versus natural gas are compared with weighting, propane heating produces 1.4% more total detrimental impacts than natural gas, indicating that natural gas heating would slightly improve the overall sustainability of the Penn State Eco-MachineTM. This result indicates that Eco-MachinesTM are capable of producing net benefits to all damage categories except resources, when located in areas that do not require heating.

Sensitivity analysis on vertical farming options to increase duckweed growth area

One of the benefits of duckweed is that it only requires a few millimeters of water depth to grow, making it well suited for vertical farming (Leng, 1999). Therefore, a sensitivity analysis was performed to examine the potential beneficial impacts of utilizing multiple stacks (1-4) of duckweed growth platforms

65 covering the area of the wetland (44.5 m2) to increase the duckweed growth area in the Penn State Eco-

MachineTM. A steel frame was used to support the vertical duckweed growth trays, which were assumed to be made of plywood sheets lined with an HDPE liner, which is a common configuration for hydroponic systems. LED light strips installed underneath the trays provide light to the duckweed growing in the tray beneath. No LED lights were used above the top tray or in the single stack scenario, since sunlight would not be impeded by another tray above. The inventory used for the vertical growth sensitivity analysis is presented in Table 3.

Table 4.3. Duckweed vertical growth tray inventory data used in the sensitivity analysis. Amount IMPACT 2002+ database item 1 stack 2 stacks 3 stacks 4 stacks Steel, unalloyed {GLO} 170 kg 240 kg 310 kg 380 kg Polyethylene, high density, granulate, recycled {US} 0.863 kg 1.726 kg 2.589 kg 3.452 kg Plywood, for indoor use {RoW} 0.565 m3 1.13 kg 1.695 kg 2.26 kg Electricity, low voltage {NPCC, US only} - 1,117 kWh 2,234 kWh 3,351 kWh Tilapia feed, 24-28% protein {GLO} 1150 kg 2,300 kg 3,450 kg 4,600 kg GLO = global US = United States NPCC = Northeast Power Coordinating Council

Increasing the number of vertical stacks used to grow duckweed provides additional beneficial impacts for human health (Figure 4.5a), ecosystem quality (Figure 4.5b), and climate change (Figure 4.5c), but increases detrimental impacts for resources (Figure 4.5d). The majority of the detrimental impact in the resource category was due to the electricity used to power the LED grow lights. In each damage category aside from resources, the plywood used as the base of the growth platform provided the largest detrimental impact, suggesting that the use of recycled wood or another recycled container system could eliminate a substantial portion of the detrimental impacts of the vertical farming system. Nonetheless, the results suggest that the additional beneficial impacts from producing more duckweed outweigh the detrimental impacts from constructing vertical farming system and the energy required to power the LED lights (Figure 4.6).

Conclusions

A life cycle assessment conducted on a pilot-scale Eco-MachineTM showed that the detrimental impacts from these systems are dominated by the fuel source used to heat the greenhouse and aeration. It was determined that if the greenhouse were located in a region that does not require heating, the system would transform detrimental impacts (positive values) for human health, ecosystem quality, and climate change into beneficial impacts (negative values). Fine bubble diffusers, mechanical mixers, or a passive trickling system incorporated into the Eco-MachineTM could reduce the detrimental impacts from aeration. The integrated solar array provides the largest beneficial impact to the system by producing nearly 80% of the electricity that the Eco-MachineTM consumes. Currently, the solar array only consists of ten solar panels, and could be easily expanded to produce all of the energy needs of the Eco-MachineTM. As currently

TM operated, the Penn State Eco-Machine produces about a third of the CO2 emissions as conventional wastewater treatment systems. The impact of the Penn State Eco-MachineTM to resources is about 4-5 times higher than conventional wastewater treatment; however, if the need for heating were removed, it would produce about one-half of the impacts to resources as conventional wastewater treatment. Finally,

68 providing additional area for duckweed growth in the form of a vertical farming system increases the overall sustainability of the system by producing more protein-rich plant biomass. Future research should investigate how these types of systems scale with treatment capacity, and validate the safety of duckweed biomass generated in different wastewater sources for use as animal feed. In addition, LCAs of ecological treatment systems coupled with duckweed growth for use in other products (ex., fertilizer, bioenergy) should be conducted to quantify the impact of using these systems for meeting various community needs.

Acknowledgements

The support of the Office of the Physical Plant and the Department of Civil and Environmental

Engineering at The Pennsylvania State University are gratefully acknowledged, as is the graduating classes of 2000 and 1950 for gifting the university with the Eco-MachineTM.

References

Almuktar, S.A.A.A.N.; and Scholz, M. (2015) “Microbial contamination of Capsicum annuum irrigated with recycled domestic wastewater treated by vertical-flow wetlands.” Ecological Engineering, 82:404- 414. Ansal, M.D.; Dhawan, M.; and Kaur, V.I.; (2010) “Duckweed based bio-remediation of village ponds: an ecologically and economically viable integrated approach for rural development through aquaculture.” Livestock Research for Rural Development, 22 Article #129. Belles, E. (2020) “A systematic carbon and energy balance of the Penn State Eco-MachineTM and the implementation of a vertical farming system to increase duckweed production.” A undergraduate honors thesis submitted to Penn State Department of Civil Engineering. University Park, PA. Campos, J.L.; Valenzuela-Heredia, D.; Pedrouso, A.; Van del Rio, A.; Belmonte, M.; and Mosquera- Corral, A. (2016) “Greenhouse gases emissions from wastewater treatment plants: minimization, treatment, and prevention.” Journal of Chemistry, 3796352. https://doi.org/10.1155/2016/3796352 Calicioglu, A.O. (2019) “Feasibility assessment of large-scale duckweed biorefineries: technical, environmental, and economic evaluation of integrated wastewater treatment and bioenergy production using duckweed.” Ph.D. dissertation. The Pennsylvania State University, Department of Civil and Environmental Engineering. University Park, PA. Calicioglu, A.O.; Richard, T.L.; and Brennan, R.A. (2019) “Anaerobic bioprocessing of wastewater- derived duckweed: maximizing product yields in a biorefinery value cascade.” Bioresource Technology, 289:121716. https://doi.org/10.1016/j.biortech.2019.121716 Cantwell, J.; Dasek, A,; Erickson, B.; Fillmore, L.; Gheewala, S.; Kottwitz, J.; Levy, M.; Liner, B.; Lung, B.; Nimbalkar, S.; Rao, P.; and Rickert, M. (2017) “Energy data management manual for the wastewater treatment sector.” U.S. Department of Energy and Environmental Protection Agency. Cheng, J.J.; and Stomp, A.M. (2009) “Growing duckweed to recover nutrients from wastewaters and for production of fuel ethanol and animal feed.” Clean, 37(1):17-26. https://doi.org/10.1002/clen.200800210 Copeland, C.; and Carter, N.T. (2017) “Energy-water nexus: the water sector’s energy use.” Congressional Research Service Report, 7-5700, R43200. Dixon, A.; Simon, M.; and Burkitt, T. (2003) “Assessing the environmental impact of two options for small-scale wastewater treatment: comparing a reedbed and an aerated biological filter using a life cycle approach.” Ecological Engineering, 20:297-308. https://doi.org/10.1016/S0925-8574(03)00007-7 Fang, Y. (2013) “Wastewater treatment by duckweed and production of high starch biomass: the pilot- scale studies.” The Second International Conference on Duckweed Research and Applications, Rutgers, NJ. Fasakin, E.A.; Balogun, A.M.; and Fasuru, B.E. (1999) “Use of duckweed, Spirodela polyrrihiza L. Schleiden, as a protein feedstuff in practical diets for tilapia, Oreochromis niloticus L.” Aquaculture Research, 30;313-318. https://doi.org/10.1046/j.1365-2109.1999.00318.x Food and Agriculture Organization. (2012) “Livestock and landscapes.” Gridded Livestock of the World database.

Food and Drug Administration. (2016) “Food safety modernization act.” Published online at fda.gov. Retrieved from: https://www.fda.gov/food/food-safety-modernization-act-fsma/fsma-final-rule-produce- safety Forslund, A.; Ensink, J.H.J.; Battilani, A.; Kljujev, I.; Gola, S.; Raicevic, V.; Jovanovic, Z.; Stikic, R.; Sandei, L.; Fletcher, T.; and Dalsgaard, A. (2010) “Faecal contamination and hygiene aspect associated with the use of treated wastewater and canal water for irrigation of potatoes (Solanum tuberosum).” Agricultural Water Management, 98:440-450. Gaigher, I. G.; Porath, D.; and Granoth, G. (1984) “Evaluation of duckweed (Lemna gibba) as feed for tilapia (Oreochromis niloticus X O. aureus) in a recirculating unit.” Aquaculture, 41:235-244. Gnansounou, E.; and Raman, J.K. (2016) “Life cycle assessment of algae biodiesel and its co-products.” Applied Energy, 161:300-308. https://doi.org/10.1016/j.apenergy.2015.10.043 Halloran, A.; Roos, N.; Eilenberg, J.; Cerutti, A.; and Bruun, S. (2016) “Life cycle assessment of edible insects for food protein: a review.” Agronomy for Sustainable Development, 36:57. https://doi.org/10.1007/s13593-016-0392-8 Hendrickson, T.P.; Nguyen, M.T.; Sukardi, M.; Miot, A.; Horvath, A.; and Nelson, K.L. (2015) “Life- cycle energy use and greenhouse gas emissions of a building-scale wastewater treatment and nonpotable reuse system.” Environmental Science and Technology, 49”10303-10311. https://doi.org/10.1021/acs.est.5b01677 Hospido, A.; Moreira, M.T.; Fernandez-Couto, M.; and Feijoo, G. (2004) “Environmental performance of a municipal wastewater treatment plant.” International Journal of Life Cycle Assessment, 9(4):261-271. https://doi.org/10.1007/BF02978602 Humbert, S.; De Schryver, A.; Bengoa, X.; Margni, M.; and Jolliet, O. (2012) “IMPACT 2002+: user guide.” Swiss Federal Institute of Technology Lausanne, Switzerland. Kiziloglu, F.M.; Turan, M.; Sahin, U.; Kuslu, Y.; and Dursun, A. (2008) “Effects of untreated and treated wastewater irrigation on come chemical properties of cauliflower (Brassica olerecea L. var. botrytis) and red cabbage (Brassica olerecea L. var. rubra) grown on calcareous soil in Turkey.” Agricultural Water Management, 95:716-724. Kreider, A.N.; Fernandez Pulido, C.; Bruns, M.A.; and Brennan, R.A. (2019) “Duckweed as an agricultural amendment: nitrogen mineralization, leaching, and sorghum uptake.” Journal of Environmental Quality, 48(2):469-475. https://doi.org/10.2134/jeq2018.05.0207 Leng, R.A. (1999) “Duckweed: a tiny aquatic plant with enormous potential for agriculture and environment.” Food & Agriculture Organization, Animal Production and Health Division. Rome, Italy. Li, Y.; Luo, X.; Huang, X.; Wang, D.; and Zhang, W. (2013) “Life cycle assessment of a municipal wastewater treatment plant: a case study in Suzhou, China.” Journal of Cleaner Production, 57:221-227. https://doi.org/10.1016/j.jclepro.2013.05.035 Marine Aquaculture Task Force (2007) “Sustainable marine aquaculture: fulfilling the promise; managing the risks.” Report of the Marine Aquaculture Task Force, p. 93. Maupin, M.A.; and Ivahnenko, T. (2011) “Nutrient loading to streams of the Continental United States from municipal and industrial effluent.” Journal of the American Water Resources Assosciation, 47(5):950-964. https://dx.doi.org/10.1111%2Fj.1752-1688.2011.00576.x

Mohedano, R.A.; Costa, R.H.R.; Tavares, F.A.; and Filho, P.B. (2012) “High nutrient removal rate from swine wastes and protein biomass production by full-scale duckweed ponds.” Bioresource Technology, 112:98-104. https://doi.org/10.1016/j.biortech.2012.02.083 Moreau, J.; Orachungwong, C.; Segura, G.; and Tanthipwon, P. (1986) “Alimentation du jeune tilapia, application au developpement de son elevage intensif. (Feeding of young tilapia, and its application to intensive rearing.)” In: Aquaculture research in the Africa region. Proceedings of the African seminar on aquaculture organisedorganized by the International Foundation for Science (IFS), Stockholm, Sweden, held in Kisumu, Kenya, 7-11 October 1985. Ed. E.A. Huisman, Wageningen, Netherlands; Pudoc. pp. 60- 96. National Resources Defense Council (2018) “Water pollution: everything you need to know.” Noyola, A.; Morgan-Sagastume, J.M.; and Lopez-Hernandez, J.E. (2005) “Treatment of biogas produced in anaerobic reactors for domestic wastewater: odor control and energy/resource recovery.” Reviews in Environmental Science and Bio/Technology, 5:93-114. https://doi.org/10.1007/s11157-005-2754-6 Oonincx, D.G.A.B.; and de Boer, I.M.J. (2012) “Environmental impact of the production of mealworms as a protein srource for humans – a life cycle assessment.” PLoS ONE (7)12: e51145. https://doi.org/10.1371/journal.pone.0051145 Roman, B.; and Brennan, R.A. (2019) “A beneficial by-product of ecological wastewater treatment: an evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture.” Ecological Engineering X, 1:100004. https://doi.org/10.1016/j.ecoena.2019.100004 Saffari, V.R.; and Saffari, M. (2013) “Effect of treated municipal wastewater on bean growth, soil chemical properties, and chemical fractions of zinc and copper.” Arabian Journal of Geosciences, 6:4475- 4485. Saunders, M. J.; Kansiime, F.; Jones, M. B. (2012) “Agricultural encroachment: implications for carbon sequestration in tropical African wetlands.” Global Change Biology, 18:1312-1321. https://doi.org/10.1111/j.1365-2486.2011.02633.x Sheehan, M.W. (2012) “Optimizing the solar disinfection method to produce potable water from ecologically treated wastewater using recycled polyethylene terephthalate bottles.” Undergraduate honors thesis. The Pennsylvania State University, Department of Civil Engineering. University Park, PA. Tabesh, M.; Masooleh, M.F.; Roghani, B.; and Motevallian, S.S. (2019) “Life-cycle assessment (LCA) of wastewater treatment plants: a case study of Tehran, Iran.” International Journal of Civil Engineering, 17:1155-1169. https://doi.org/10.1007/s40999-018-0375-z Tacon, A.G.J.; and Metian, M. (2009) “Fishing for feed or fishing for food: increasing global competition for small pelagic forage fish.” Ambio, 38:294-293. United State Geological Survey (2018) “Estimated use of water in the United States in 2015.” Water Availability and Use Science Program. Open-file report 2017-1131. United States Environmental Protection Agency (2000) “Wastewater technology fact sheet: free water surface wetlands.” EPA 832-F-00-024.

United States Environmental Protection Agency (2012) “Global anthropogenic non-CO2 greenhouse gas emissions: 1990-2030. Office of Atmospheric Programs. Washington, DC. EPA 430-R-12-006.

72 van Oirschot, R.; Thomas, J.B.E.; Grondahl, F.; Fortuin, K.P.J.; Brandenburg, W.; and Potting, J. (2017) “Explorative environmental life cycle assessment for system design of seaweed cultivation and drying.” Algal Research, 27:43-54. https://doi.org/10.1016/j.algal.2017.07.025 Wu, L.; Chen, W.; French, C.; and Chang, A. (2009) “Safe application of reclaimed water reuse in the southwestern United States.” University of California. Division of Agriculture and Natural Resources. Publication 8357. Zhao, Y.; Fang, Y.; Jin, Y.; Huang, J.; Bao, S.; Fu, T.; He, T.; Wang, F.; and Zhao, H. (2014) “Potential of duckweed in the conversion of wastewater nutrients to valuable biomass: a pilot-scale comparison with water hyacinth.” Bioresource Technology, 163:82-91. https://doi.org/10.1016/j.biortech.2014.04.018

Chapter 5 – Duckweed protein supports the growth and organ development of mice: a feeding study comparison to conventional casein.

Roman, B., Brennan, R.A., and Lambert, J.D. 2020. Duckweed protein supports the growth and organ development of mice: a feeding study comparison to conventional casein. Journal of Food Science. (In review).

Abstract

As global population growth and meat consumption increases, sustainable alternatives to conventional protein-rich fodder crops for livestock are needed to reduce negative environmental impacts. Duckweed, a small floating aquatic plant, can generate 5-10 times higher protein yields than conventional land-grown crops. Although some in vivo feeding trials with duckweed have been conducted, those measuring animal weight are limited, and those examining organ development are nonexistent. To secure broad acceptance of new protein sources, such controlled studies are critical. This study measured the growth, food intake, and final organ and adipose tissue mass of male CF-1 mice fed a semi-purified diet containing casein or diets in which 10% or 25% of the casein was replaced with duckweed protein (DWP). Proximate analysis showed that the DWP preparation used contained 39.9% protein (w/w), and contained all of the essential amino acids with Met as the limiting amino acid. The average growth rates were not significantly different among the treatment groups: 0.21 g/day; 0.24 g/day; and 0.25 g/day for the control; 10%; and

25% DWP protein diets, respectively. The daily food intake of both DWP diets was 6.5 – 8.0% higher than the control diet, but feeding efficiency did not differ among diets. The relative weight of the liver, spleen, kidneys, heart, and epidydimal fat, and colon length were not significantly different between treatment groups. The results from this study show that replacement of up to 25% dietary casein with

DWP has no adverse effects on the growth rate and final organ and adipose tissue weights of laboratory mice.

Introduction

Due to global population growth and increasing meat consumption in developing regions, the demand for animal-derived proteins is expected to double by 2050 (FAO, 2011). This increase in animal production will have compounding effects that are expected to negatively impact the environment: increased livestock production will require the conversion of forests, wetlands, and grasslands into grazing lands, essentially converting CO2 sinks into greenhouse gas producing areas. Furthermore, feeding livestock takes significant quantities of plant protein, with 85 million tons of plant protein required to produce 64 million tons of livestock protein annually (Steinfeld et al., 2006), concomitantly increasing land use, nutrient loads in water bodies, and reducing water availability, soil fertility, and biodiversity (Roser &

Ritchie, 2017). Potential solutions to alleviate the environmental impacts of animal agriculture include decreasing animal protein (i.e., meat) consumption, and switching to more sustainable sources of plant protein to meet the demands for food and fodder. It is estimated that alternatives to animal protein could claim up to one-third of the market by 2054 (Lux Research, 2014). Indeed, young consumers are increasingly looking to replace animal-derived proteins with plant-based proteins in their diets, with 46% agreeing that meat alternatives are healthier than meat (Scott-Thomas, 2015; McCarthy & Dekoster,

2020).

Many attempts have been made to find abundant and inexpensive protein-rich substitutes for conventionally grown feedstocks. For example, insect meal has been used in poultry, rabbit, pig, and fish diets (Adeniji, 2007; Duwa, Saleh, & Iqwebuike, 2014; Adeniji, 2008; and Adewolu, Ikenweiwe, &

Mulero, 2010), but the protein content is low (< 20% dry matter (DM)) (Lundy & Parrella, 2015). Soy, wheat, corn, sorghum, peas, and lupin have been used as supplements in fish meal (Allan et al., 2000;

Burel et al., 2000; and Kaushik et al., 1995), but due to their lower nutritional value, only a small portion of these plant-based proteins are typically blended with conventional meat-based feed. Another emerging protein source is algae: 24 million tons are currently farmed globally, with the majority being used in health foods, cosmetics, and animal feed. Although some microalgae contain nearly 50% protein DM,

75 algae protein development has been hampered by high production costs and technical challenges

(Henchion, Hayes, Mullen, Fenelon, & Tiwari, 2017). Clearly, additional research and development of sustainable protein sources is required to support growing demand.

Duckweed, a small floating aquatic plant, has shown a high potential to produce large amounts of protein by recovering nutrients from polluted waters. When grown under ideal conditions, duckweed can double its mass every 16-48 h and accumulate a crude protein content of up to 45% of the DM (Leng,

1999). In fact, it has been determined that duckweed grown on partially treated municipal sewage can generate protein yields that are 5-10 times higher than conventional land grown crops, such as soybeans, maize, and oats (Roman & Brennan, 2019). In addition to its prolific growth rates, another benefit of duckweed is the ease with which it can be skimmed from water and used for beneficial purposes.

Although algae and seaweed have long been used as components of plant-based nutrition (De Pinto &

Verhoff, 1977), harvesting algae from water can be difficult and energy-intensive, and seaweed is only sustainably available in coastal regions. Duckweed, however, can grow anywhere there is water: various genera of duckweed (most common: Lemna, Spirodela, and Wolffia) are found worldwide, and the dominant genera in the region depends on the water quality, topography, and climate (Leng, 1999).

Although the quantity of protein producible by duckweed is well documented, less work has been done on the quality of duckweed protein. Several studies have examined the amino acid composition of various duckweeds and concluded that based on its amino acid profile, the quality of duckweed protein is suitable for consumption by fish, poultry, cattle, and humans (Appenroth et al., 2018; Hanczakowski,

Szymczyk, & Wawrzynski, 1995; Sharma et al., 2019). Duckweed can also accumulate trace minerals (K,

Ca, Mg, Na, Fe) which are often deficient in the feed available to small livestock farmers (Leng, 1999), as well as micronutrients (iodine and vitamin A) which are commonly deficient in the diets of malnourished people (Valdimirova & Georgiyants, 2014). Although duckweed can provide a source of minerals and vitamins, it is generally only regarded as a protein supplement due to the high protein yields that can be achieved. Duckweed feeding trials have been conducted with pigs, ruminants, poultry, and fish, with

76 varying results. In some studies, duckweed has been shown to be able to supply all protein requirements for certain animals with no adverse health effects (Cheng & Stomp, 2009), while in other studies, small additions of duckweed into animal diets has been shown to decrease their weight gain and food intake

(Sonata, Rekiel, & Batorska, 2019). The confounding results in prior duckweed feeding studies are likely due to several factors, with arguably the most important being the broad range of nutritional quality of the duckweed used, which is highly dependent on its growth medium. In addition, no studies to date have examined organ development in animals fed a duckweed supplemented diet, which is critical for securing broad acceptance of protein supplements.

In this study, duckweed protein (DWP) was extracted, analyzed, and formulated into two diets in which 10% and 25% of the protein content was derived from duckweed protein, and the balance was casein. These duckweed-supplemented diets were compared to a conventional casein diet and fed to 30

CF-1 male mice for four weeks. Body weight gain, food intake rate, and final organ weight were analyzed.

Materials and methods

Preparation of duckweed and diet formulation

The duckweed used in the study was a polyculture of Lemna minor and Wolffia spp. grown on a nutrient media with a protein content of 40% DM (Real Source Foods, Melbourne, FL). In the lab, the duckweed was defatted and polyphenols removed by extracting sequentially with hexane and ethyl acetate, respectively. In brief, 50 g of dried duckweed was combined with 200 mL solvent for 10 min, draining the solvent through a Buchner funnel, and repeating. Extraction with each solvent was performed a total of three times. The diets: AIN93G (casein); AIN93G + 10% duckweed protein (DWP) replacement (4.3% duckweed mass addition); and AIN93G + 25% DWP replacement (10.7% duckweed mass addition) were prepared by Research Diets Inc (New Brunswick, NJ, USA). Diets were matched for energy, macronutrient, and micronutrient composition (Table 1). Proximate analysis and amino acid analysis were performed on the defatted/polyphenol-depleted duckweed was also analyzed (The University of Missouri

Agriculture Experiment Station Chemical Laboratories, Columbia, MO).

Table 5.1. Composition of diets used in this study. AIN93G (control) AIN93G + 10% AING93G + 25% DWP (4.3% mass) DWP (10.7% mass) mass Energy mass Energy mass Energy Ingredient (g) (kcal) (g) (kcal) (g) (kcal) Casein 200 800 180 720 150 600 L-Cystine 3.0 12 3.0 12 3.0 12 Corn starch 397 1590 397 1590 397 1590 Maltodextrin 10 132 528 132 528 132 528 Sucrose 107 428 107 428 107 428 Cellulose 50 0 32 0 5.0 0 Soybean oil 70 630 70 630 70 630 T-Butylhydroquinone 0.01 0 0.01 0 0.01 0 Mineral mix 3.5 0 3.5 0 3.5 0 Calcium carbonate 12.5 0 10.9 0 8.4 0 Potassium phosphate 6.9 0 6.4 0 5.8 0 Potassium citrate 2.5 0 2.3 0 2.1 0 Sodium chloride 2.6 0 2.6 0 2.6 0 Vitamin mix 10 40 10 40 10 40 Choline Bitartrate 2.5 0 2.5 0 2.5 0 Duckweed 0 0 43 79 108 199 Total 1000 4028 1003 4027 1007 4027

Mouse feeding experiment

The feeding study was approved by the Pennsylvania State University Institutional Animal Care and Use

Committee (Protocol No. 45380). Male CF-1 mice (5 wks old, 33.3g ± 2.4g at Day 1) were purchased from Charles River Laboratory (Wilmington, MA, USA). Mice were housed with 12 h light/dark cycles held at 18-23oC and 50-60% relative humidity, and allowed acclimate to the room for one week prior to starting the study. Mice were randomized by body weight (n = 10 per diet) to each diet. The body mass of each mouse was recorded twice a week for four weeks. Food consumption was measured twice weekly.

Fresh water was provided ad libitum. At the end of the feeding study, the mice were anesthetized and euthanized by exsanguination. The blood samples were centrifuged at 3200 g for 15 min to separate the plasma which was then stored at -80oC for future analysis. The liver, spleen, lungs, kidneys, heart, and epididymal (epi) fat were removed, rinsed with ice-cold 0.9% NaCl, trimmed of connective tissue and fat, blotted dry, and weighed. Organ and adipose tissue masses were normalized to final body weight. The length of the colon was also recorded. Feeding efficiency was calculated as the ratio of rate of body weight gain (g/d) to food intake (g/d), representing the change in body weight as a function of the mass of food consumed.

Statistical analysis

The growth rate, food intake rate, and organ masses were compared separately using a one-way ANOVA and Tukey post-test in Minitab to determine differences for each variable depending on the diet (p<0.05).

Results and discussion

Duckweed protein quality

Proximate analysis showed that the defatted/polyphenol-depleted duckweed preparation used in our experiments contained 39.9% protein, 14.2% fiber, and 1.1% fat (Table 2). Amino acid analysis showed that the material contained all the essential amino acids (EAA) in appreciable amounts, and all but two non-essential (Asp and Sec) and one conditionally essential (Gln) amino acids (Table 2). The EAA composition of duckweed protein was compared to other common plant-protein sources (Table 3), and the

79 mass required to meet the WHO recommended daily EAA intake for a 70 kg adult human was calculated for each protein source (Table 4). The limiting EAA for duckweed is Met, and assuming that duckweed is

100% digestible, it would require 92.2 g of defatted/polyphenol-depleted duckweed powder/day (Table

4). Compared to the mass required of the limiting EAA for the other plant-protein sources, defatted/polyphenol-depleted duckweed out-competes all but brown rice and potato protein isolates

(Table 4), suggesting that duckweed is an ideal candidate as a protein supplement for a human diet.

Table 5.2. Proximate analysis and amino acid composition of the duckweed protein isolate used in this study. Amino acid composition (mg/g) Proximate analysis Essential Nonessential Conditionally (%) essential Crude protein 39.9 His 8.9 Ala 24.1 Arg 24.9 Crude Fat 1.1 Ile 20.3 Asp 35.3 Cys 4.7 Crude Fiber 14.2 Leu 36.1 Glu 43.1 Pro 17.9 Ash 5.4 Lys 26.4 Gly 21.1 Moisture 14.4 Met 7.9 Ser 15.3 Phe 23.0 Tyr 14.0 Thr 17.4 Trp 6.8 Val 24.9

Table 5.3. Essential amino acid (EAA) composition of protein isolate from duckweed used in this study and other common plant-protein sources. Values are represented in g per 100 g raw material. EAA content of protein isolate (g/100 g) EAA Duckweed Oat Lupin Wheat Soy Brown rice Pea Corn Potato His 0.9 0.9 1.2 1.4 1.5 1.6 1.6 1.1 1.4 Ile 2.0 1.3 1.5 2 1.9 2.3 2.3 1.7 3.1 Leu 3.6 3.8 3.2 5 5 5.7 5.7 8.8 6.7 Lys 2.6 1.3 2.1 1.1 3.4 4.7 4.7 1 4.8 Met 0.8 0.1 0.2 0.7 0.3 2 0.3 1.1 1.3 Phe 2.3 2.7 1.8 3.7 3.2 3.7 3.7 3.4 4.2 Thr 1.7 1.5 1.6 1.8 2.3 2.3 2.5 1.8 4.1 Trp 0.7 N/A N/A N/A N/A N/A N/A N/A N/A Val 2.5 2 1.4 2.3 2.2 2.7 2.7 2.1 3.7 EAA values for other common plant-protein sources from Gorissen et al., 2018; N/A = not analyzed.

Table 5.4. Mass of protein isolate required to meet the daily recommended essential amino acid (EAA) intake for an adult human by the World Health Organization. Values are represented in g per 70 kg body weight per day. Bolded and underlined values indicate the limiting essential amino acid. Mass to meet the daily human EAA requirement (g/70 kg bw/d) EAA Duckweed Oat Lupin Wheat Soy Brown rice Pea Corn Potato His 78.7 77.8 58.3 50.0 46.7 43.8 43.8 63.6 50.0 Ile 69.0 107.7 93.3 70.0 73.7 60.9 60.9 82.4 45.2 Leu 75.6 71.8 85.3 54.6 54.6 47.9 47.9 31.0 40.7 Lys 79.5 161.5 100.0 190.9 61.8 44.7 44.7 210.0 43.8 Met 92.2 728.0 364.0 104.0 242.7 36.4 242.7 66.2 56.0 Phe 76.1 64.8 97.2 47.3 54.7 47.3 47.3 51.5 41.7 Thr 60.3 70.0 65.6 58.3 45.7 45.7 42.0 58.3 25.6 Trp 41.2 N/A N/A N/A N/A N/A N/A N/A N/A Val 73.1 91.0 130.0 79.1 82.7 67.4 67.4 86.7 49.2 Daily recommended EAA intake for humans from WHO, 2007; N/A = not analyzed.

Body weight gain and relative organ weights

There was no significant difference in body weight or rate of body weight gain for the mice fed different diets (Fig. 1A and B, p=0.559). The relative masses of the liver, spleen, kidneys, heart, and epididymal fat were not different among the mice treated with the three different diets (Fig. 1C). The lungs of the mice fed the 25% DWP diet were significantly larger (about 20%) than those of the mice on the control diet

(p=0.028). This difference could be due to chance, since it has been reported that there is no difference in the size of lungs from obese mice (>50g) and lean mice (<30 g) (Guivarch et al., 2018), which is a much larger gap in body mass than that observed between the control- and DWP-fed mice in this study. The average colon lengths of the mice fed the three diets were not significantly different (Fig. 1C). There were no gross differences in organ morphology between treatment groups. These results indicate that DWP can support normal growth and development of major organs. To our knowledge, this is the first study to exam the effect of DWP on organ weights.

Body weight gain and relative organ weights

A 45 B 0.40 Control 10% DWP 25% DWP 0.35 42 0.30

39 0.25 0.20 36 0.15 Body weight (g) weightBody

Growth rate (g/day) rate Growth 0.10 33 0.05 A A A 30 0.00 0 5 10 15 20 25 30 Control 10% 25% Day of dietary treatment DWP DWP C 7% 20% Control 10% DWP 25% DWP 18% 6% 16% 5% 14%

4% 12% 10% 3% 8% % of body mass body % of

2% mass length/body Colon 6% 4% 1% 2% A A A A A A A A B B A A A A A A A A A 0% 0% A A A Liver Spleen Lungs Kidneys Heart Epi Fat Colon Length

A number of duckweed feeding studies were conducted in the 1970s and 1980s (reviewed in

Culley, Rejmankova, Kvet, & Frye, 1981). Table 5 summarizes the results and limitations of studies conducted since 1990. Studies in non-ruminant mammals have focused on pigs. Consistent with our present results, one study with piglets found that replacing up to 60% of soybean meal mass in their diets with duckweed mass resulted in similar or greater growth (Moss, 1999). While the studies in piglets are

82 generally positive, with duckweed substitution yielding equivalent or superior increases in body weight, most are confounded by the use of different base diets between treatment groups, poorly characterized nutrient content for the overall diet, and/or incomplete characterization of the duckweed used in the studies.

Similar confounding issues exist in studies with other agricultural species. For example, two studies examined the impact of dried duckweed on body weight gain in broiler chickens, but found different results. One study found that the chickens fed corn-based diets supplemented with 10% and 15% duckweed showed comparable bodyweight gain to the chickens fed a solely corn-based diet (Haustein,

Gilman, & Skillicorn, 1994), while another study found that chickens fed an undefined commercial diet supplemented with 4%, 8%, and 12% duckweed had decreasing body weight as the % of duckweed increased (Kabir, Islam, Ahammad, & Howlider, 2005). Without clearly defined basal diets, it is impossible to determine the reasons for the differences observed between these studies.

It is well-established that duckweeds can exhibit a large range of nutritional characteristics depending on the nutrient content of their growth medium, with protein content ranging from 15-45% and fiber content ranging from 5-30% (Leng, Stombolie, & Bell, 1995). A strength of the current study is the use of a well-defined duckweed supplement and a consistent basal diet, allowing a more robust comparison between duckweed-derived protein and other protein sources. Thus, future duckweed feeding studies in agricultural species should formulate duckweed into diets based on a % of protein substitution, rather than a % of mass substitution to accurately represent the effect of DWP diets on animal growth.

Table 5.5. Review of duckweed feeding studies published since 1990 (adapted from Sonta et al., 2019). Bolded Positive/Negative/Neutral in the result column indicates the effect on including duckweed in the diet compared to the conventional diet. Animal Diet Result Citation Titan broilers fed commercial diet supplemented Negative: Chickens fed diets over 150 g/kg DW showed decreased Haustein et al. (1992) with 0, 100, 150, 200, 250, or 300 g/kg DW body weight gain and feed intake. Chickens fed the no DW diet reached much higher body weight gains and feed intake. Titan and Arbor Acres broilers fed maize diets Positive (≤15% DW supplement): Body weight gains and feed Haustein, Gilman, supplements with 0%, 10%, 15%, or 25% dried intake were comparable up to 15% DW, but deteriorated at 25% and Skillicorn (1994) Broiler DW DW. chickens Vencobb broilers sesame oil cake diet Positive (≤6% DW supplement): Final body weight and feed Ahammad, Swapon, supplemented with 0%, 3%, 6%, or 9% dried DW conversion were higher in the 3% and 6% DW diets. Poorest results Yeasmin, Rahman, were observed by the 9% DW supplement diet. and Ali (2003) Vencobb broilers fed commercial diets Negative: Increasing % of DW in diet decreased body growth, feed Kabir, Islam, supplemented with 0%, 4%, 8%, or 12% dried conversion ratio, and protein and energy intake Ahammad, and DW Howlider (2005) Star Cross Brown laying hens fed rice polish and Negative: No differences were observed between all groups for Akter, Chowdhury, fish meal diet replaced with 0, 50, 70, 110, 130, body weight, egg weight, and livability. Feed intake, egg production, Yeasmin, and Khatun and 150 g/kg dried DW. and feed conversion ratio all decreased with increasing DW in diet. (2011) Laying Hy-Line laying hens fed soybean meal diet Neutral: Egg production and nutrient composition of eggs were Anderson, Lowman, hens replaced with 0% or 12.5% DW. similar between the two diets, except omega-3 acids were higher in Stomp, and Chang the DW diet than in the soybean diet. (2011) Lohmann Brown layers fed a commercial diet Positive: Egg quality parameters were improved in the DW diet and Witkowska et al. supplemented with inorganic salt (C) or DW (E) birds had lower Cd in yolk and Pb in blood when fed the DW diet. (2012) Muscovy ducks fed rice diet supplemented with: Positive: Daily weight gain was lowest in E1 and highest in E2. Ngamsaeng, Thy, E1) 80% water spinach; E2) 80% DW; and E3) Feed conversion ratio was higher in E2 compared to E1 and E3. The and Preston (2004) 35% water spinach and 45% DW use of DW in duck diets was advised. Xingding ducks fed commercial diet Negative: All E3 ducks died within the first 3 weeks, and was Khanum, Chwalibog, supplemented with fresh (wet) DW: (C) concluded that DW cannot be used as the sole feed for ducks. Final and Huque (2005) Ducks commercial compound feed; (E1) 50% C + fresh body weight and daily weight gain were significantly lower in the E collected DW ad libitum (E1); 50% C + DW groups compared to C. DW inclusion in the diet caused a decrease forage (E2); and (E3) DW lagoon forage. on the growth of the ducks. Jinding ducks fed mustard oil cake diet replaced Negative: Body weight gain, egg weight, and feed conversion ratio Khandaker, Khan, with 0%, 5%, 10%, or 15% dried DW. were similar between groups, but egg production decreased with Shahjalaal, and increasing % DW in diet. Rahman (2007) Piglets fed soybean diets supplemented with 0%, Positive: Daily body weight gains were significantly higher in pigs Moss (1999) Pigs 20%, 40%, 60% dried DW fed 40 & 60% DW diets compared to the 0 (control soybean diet) & 20% DW diets.

Pigs fed sorghum and soybean meal (C) replaced Neutral: Weight gain, feed conversion ratio, and final body weight Gutierrez, Sangines, with 10% dried DW (E). were similar between groups. Perez, and Martinez (2001). Pigs fed sweet potato vines (C) or fresh DW (E) Positive: Body weight, live weight gain, and feed conversion ratio Du (1998) were higher for the DW diet. Pigs fed control diet compared to DW diet: Neutral: Pigs fed the C diet had higher final body weight, but the Van, Men, Son, and C) 60% broken rice; 33% rice bran; % fish meal; feed conversion ratio was similar. The carcasses of E had thinner Preston (1997) and 2% soybean meal and E) 69% cassava root; backfat than C. 8.6% DW; and 22.4% supplements Merino ewes fed oaten chaff diets supplemented Neutral: Hair coat parameters (wool yield, rate of wool elongation, Damry, Nolan, Bell, with DW: fiber diameter) did not differ among the groups. It was concluded and Thomson (2001) C) 700 g/day oaten chaff that DW is a valuable source of protein for the ruminants. E1) 630 g/day oaten chaff + 50 g/day dried DW E2) 540 g/day oaten chaff + 100 g/day dried DW Ruminants E3) 630 g/day oaten chaff + 1 kg/day fresh DW Boer goats fed soybean meal supplemented with Neutral: Nitrogen intake, excretion, and serum urea level, Reid (2004) 0%, 33%, and 66% dried DW phosphorous intake and excretion, and rumen pH, ammonium, and volatile fatty acids were similar for all diets. It was concluded that DW is nutritionally comparable to soybean meal. Nile tilapia fed fish meal (C) supplemented with Neutral: Weight gains were significantly higher in the C and E3 El-Shafai, El-Gohary, 20% (E1) or 40% (E2) dried DW and 20% (E3) diets. The remaining diets showed similar weight gains. Feed Verreth, Schrama, or 40% (E4) fresh DW conversion ratio and protein efficiency were similar for all groups. and Gijzen (2004) Fish fed E diets had greater phosphorous and protein content and lower lipid content than fish fed the C diet. Fish fed the C diet had higher dry matter content and lower ash content than fish fed the E Fish diets. Tilapia fed commercial feed supplemented with Neutral (≥50% DW supplement): Final body weight of fish fed the Tavares, Rodrigues, 0% (C), 50% (E1) and 100% (E2) dried DW C and E1 diets were similar, but fish fed the E2 diet had lower final Fracalossi, Esquivel, body weight. and Roubach (2008) Vundu fed fish meal supplemented with 0%, Negative: Growth rate, body weight, and feed conversion ratio were Effiong, Sanni, and 10%, 20%, and 30% dried DW similar for 0 and 10% DW fed fish, but were lower for 20 and 30 % Fakunle (2009) DW fed fish. Pacific white shrimp fed fish meal replaced with Positive: Survival rates were similar in all groups. The best growth Flores-Miranda et al. Shrimp 0%, 5%, 15%, 25%, or 35% fermented DW. efficiency was observed in shrimp fed the 35% DW diet. (2015) CF-1 male mice fed casein diet supplement with Neutral: Mice body mass growth was similar for each diet and food This study This study 0%, 10%, and 25% DW protein (DWP) intake rate was higher for both DWP diets.

Food intake and feeding efficiency

Food intake of the DWP diets were significantly greater than the control casein diet (p=0.006, Fig. 2A).

The feeding efficiency of the diets were not significantly different (p=0.678, Fig. 2B), however, indicating that there was no difference in utilization of protein and nutrients between the diets.

Furthermore, there was no trend present (R2=0.43) for the rate of body weight gain against the mean daily food intake rate for each group, establishing that the DWP diets provided no additional nutritional benefit.

25 1.8% A B 1.6% 23 1.4% 1.2% 21 1.0% 0.8% 19 0.6% Feeding Feeding efficiency Food intake (g/day) 17 0.4% 0.2% A B B A A A 15 0.0% Control 10% 25% Control 10% 25% DWP DWP DWP DWP

Figure 5.2. A) Food intake rates and B) feeding efficiency of the three diets (n=16 per diet for food intake; n=10 per diet for feeding efficiency; 95% CI). Conclusion

Compared to conventional land grown crops, duckweed can produce larger quantities of protein for animal feed, which makes it a prime candidate to meet growing global demands for animal-derived proteins. This study showed that replacement of up to 25% of casein in the basal diet with duckweed protein had no adverse effects on the growth or organ development in mice. Future studies investigating the effect of duckweed diets on the development of agricultural animals should use a consistent basal diet and substitute duckweed for other protein-rich feed based on protein content rather than total dry mass.

Finally, given the growing demand for plant-based proteins for human consumption, studies that examine the palatability and ingredient functionality of duckweed protein are needed.

Acknowledgements

The support of the Institutes of Energy and the Environment Seed Grant Program at The Pennsylvania

State University is gratefully acknowledged.

References

Adeniji, A.A. (2007). Effect of replacing groundnut cake with maggot meal in the diet of broilers. International Journal of Poultry Science, 6, 822-825. http://dx.doi.org/10.3923/ijps.2007.822.825 Adeniji, A.A. (2008). The feeding value of rumen content-maggot meal mixture in the diets of early weaned piglets. Asian Journal of Animal and Veterinary Advances, 3, 115-119. http://dx.doi.org/10.3923/ajava.2008.115.119 Adewolu M.A., Ikenweiwe N.B., and Mulero S.M. (2010). Evaluation of an animal protein mixture as a replacement for fishmeal in practical diets for fingerlings of Clarias gariepinus (Burchell, 1822). Israeli Journal of Aquaculture-Bamidgeh, 62(4), 237-244. Ahammad, M.U., Swapon, M.S.R., Yeasmin, T., Rahman, M.S., and Ali, M.S. (2003). Replacement of sesame oil cake by duckweed (Lemna minor) in broiler diet. Pakistan Journal of Biological Sciences, 6, 1450-1453. http://dx.doi.org/10.3923/pjbs.2003.1450.1453 Akter, M., Chowdhury, S.D., Yeasmin, A., and Khatun, M.A. (2011). Effect of duckweed (Lemna minor) meal in the diet of laying hen and their performance. Bangladesh Research Publications Journal, 5, 252-261. Allan, G.L., Parkinson, S., Booth, M.A., Stone, D.A.J., Rowland, S.J., Frances, J., Warner- Smith, R. (2000). Replacement of fish meal in diets for Australian silver perch, Bidyanus bidyanus: I. Digestibility of alternative ingredients. Aquaculture, 186, 293-310. https://doi.org/10.1016/S0044-8486(99)00380-4 Anderson, K.E., Lowman, Z., Stomp, A.M., and Chang, J. (2011). Duckweed as a feed ingredient in laying hen diets and its effect on egg production and composition. International Journal of Poultry Science, 10(1), 4-7. http://dx.doi.org/10.3923/ijps.2011.4.7 Appenroth, K.J., Sree, K.W., Bog, M., Ecker, J., Seeliger, C., Bohm, V., Lorkowski, S., Sommer, K., Vetter, W., Tolzin-Banasch, K., Kirmse, R., Leiterer, M., Dawczynski, C., Liebisch, G., and Jahreis, G. (2018). Nutritional value of duckweed species of the genus Wolffia (Lemnaceae) as human food. Frontiers in Chemistry, 6(483). https://dx.doi.org/10.3389%2Ffchem.2018.00483 Burel, C., Boujard, T., Kaushik, S.J., Boeuf, G., Van Der Geyten, S., Mol, K.A., Kühn, E.R., Quinsac, A., Jrouti, M., and Ribaillier, D. (2000). Potential of plant-protein sources as fish meal substitutes in diets for turbot (Psetta maxima): growth, nutrient utilization and thyroid status. Aquaculture, 188(3-4), 363-382. https://doi.org/10.1016/S0044-8486(00)00342-2 Cheng, J.J., and Stomp, A.M. (2009). Growing duckweed to recover nutrients from wastewaters and for production of fuel ethanol and animal feed. Clean, 37(1), 17-26. https://doi.org/10.1002/clen.200800210

Culley, D.D. Jr., Rejmankova, E., Kvet, J., and Frye, J.B. (1981). Production, chemical quality, and use of duckweeds (lemnaceae) in aquaculture, waste management, and animal feeds. Journal of World Mariculture Society, 12(2), 27-49. https://doi.org/10.1111/j.1749-7345.1981.tb00273.x Damry, H., Nolan, J.V., Bell, R.E., Thomson, E.S. (2001). Duckweed as a protein source for fine-wool Merino sheep: its edibility and effects on wool yield and characteristics. Asian- Australasian Journal of Animal Sciences, 14(4), 507–514. https://doi.org/10.5713/ajas.2001.507 De Pinto, J. V. and F. H. Verhoff. (1977). Nutrient regeneration from aerobic decomposition of green algae. Environmental Science and Technology, 11(4), 371-377. https://doi.org/10.1021/es60127a002 Du, T.H. (1998). Ensiled cassava leaves and duckweed as protein sources for fattening pigs on farms in Central Vietnam. Livestock Research for Rural Development, 10(25). Duwa, H., Saleh, B., and Igwebuike, J.U. (2014). The replacement of fish meal with maggot meal on the performance, carcass characteristic, hematological and serum biochemical indices of growing rabbits. Global Journal of Bio-Science and BioTechnology, 3(2), 215-220. Effiong, B.N., Sanni, A., and Fakunle, J.O. (2009). Effect of partial replacement of fishmeal with duckweed (Lemna pauciscostata) meal on the growth performance of Heterobranchus longifilis fingerlings. Report and Opinion, 1(3), 76–80. Flores-Miranda, M.C., Luna-González, A., Cortés-Espinosa, D.V., Alvarez-Ruiz, P., Cortes- Jacinto, E., Valdez-Gonzalez, F.J., Escamilla-Montes, R., and Gonzalez-Ocampo, H.A. (2015). Effects of diets with fermented duckweed (Lemna sp.) on growth performance and gene expression in the Pacific white shrimp, Litopenaeus vannamei. Aquaculture International, 23, 547-561. https://doi.org/10.1007/s10499-014-9835-x Food and Agriculture Organization of the United Nations (2011). Mapping supply and demand for animal-source foods to 2030. Animal Production and Health Working Paper No. 2. Rome, Italy. Gorissen, S.H.M., Crombag, J.J.R., Senden, J.M.G., Huub Waterval, W.A., Bierau, J., Verdijk, L.B., and van Loon, L.J.C. (2018). Protein content and amino acid composition of commericially available plant-based protein isolates. Amino Acids, 50. 1685-1695. Guivarch, E., Voiriot, G., Rouze, A., Kerbrat, S., Van Nhieu, J.T., Montravers, P., Maitre, B., Dessap, A.M., Desmard, M., and Boczkowski, J. (2018). Pulmonary effects of adjusting tidal volume to actual or ideal body weight in ventilated obese mice. Scientific Reports, 8, Article No. 6439. https://doi.org/10.1038/s41598-018-24615-5 Gutiérrez, K., Sanginés, L., Pérez, F., and Martínez, L. (2001). Studies on the potential of the aquatic plant Lemna gibba for pig feeding. Cuban Journal of Agricultural Science, 35. 343-348.

Hanczakowski, P., Szymczyk, B., and Wawrzynski, M. (1995). Composition and nutritive-value of sewage-grown duckweed (Lemna minor L) for rats. Animal Feed Science and Technology, 52, 339-343. https://doi.org/10.1016/0377-8401%2894%2900729-S Haustein, A.T., Gilman, R.H., and Skillicorn, P.W. (1994). Performance of broiler chickens fed diets containing duckweed (Lemna gibba). Journal of Agricultural Science, 122(2), 285-289. https://doi.org/10.1017/S0021859600087475 Haustein, A.T., Gilman, R.H.; Skillicorn, P.W., Guevara, V., Diaz, F., Vergara, A., and Gilman, J.B. (1992). Compensatory growth in broiler chicks fed on Lemna gibba. British Journal of Nutrition, 68, 329-335. https://doi.org/10.1079/bjn19920092 Henchion, M., Hayes, M., Mullen, A.M., Fenelon, M., and Tiwari, B. (2017). Future protein supply and demand: strategies and factors influencing a sustainable equilibrium. Foods, 6(7), 53. https://doi.org/10.3390/foods6070053 Kabir, J., Islam, M.A., Ahammad, M.U., and Howlider, M.A.R. (2005). Use of duckweed (Lemna minor) in the diet of broiler. Indian Journal of Animal Research, 39(1), 31-35. Kaushik, S.J., Cravedi, J.P., Lalles, J.P., Sumpter, J., Fruconneau, B., and Laroche, M. (1995). Partial of total replacement of fish meal by soybean protein on growth, utilization, potential estrogenic or antigenic effects, cholesterolemia and flesh quality in rainbow trout, Oncorhynchus mykiss. Aquaculture, 133(3-4), 257-274. https://doi.org/10.1016/0044-8486(94)00403-B Khandaker, T., Khan, J., Shahjalal, and Rahman, M. (2007). Use of duckweed (Lemna perpusilla) as a protein source feed item in the diet of semi-scavenging Jinding Layer ducks. The Journal of Poultry Science, 44(3), 314-321. https://doi.org/10.2141/jpsa.44.314 Khanum, J., Chwalibog, A., and Huque, K.S. (2005). Study on digestibility and feeding systems of duckweed in growing ducks. Livestock Research for Rural Development, 17(5). Leng, R.A. (1999). Duckweed: a tiny aquatic plant with enormous potential for agriculture and environment. Food & Agriculture Organization, Animal Production and Health Division. Rome, Italy. Leng, R.A., Stombolie, J.H., and Bell, R. (1995). Duckweed – a potential high-protein feed resource for domestic animals and fish. Livestock Research for Rural Development, 7(1). Lundy, M.E., and Parrella, M.P. (2015). Crickets are not a free lunch: protein capture from scalable organic side-streams via high-density populations of Acheta domesticus. PLoS ONE, 10(4). https://doi.org/10.1371/journal.pone.0118785 Lux Research (2014). WhooPea: Plant sources are changing the protein landscape. State of the Market Report, Boston, MA. McCarthy, J., and Dekoster, S. (2020). Nearly one in four in U.S. have cut back on eat meat. Press release, Jan. 27. Gallup, Washington, D.C. gallup.com

Moss, B.S. (1999). Economics and feed value of integrating duckweed production with a swine operation. Submitted to the Graduate Faculty of Texas Tech. University in Master of Science, Texas University. Ngamsaeng, A., Thy, S., and Preston, T.R. (2004). Duckweed (Lemna minor) and water spinach (Ipomoea aquatic) as protein supplements for ducks fed broken rice as basal diet. Livestock Research for Rural Development, 16(3). Reid, D.W.S. (2004). Exploring duckweed (Lemna gibba) as a protein supplement for ruminants using the Boer goat (Capra hircus) as a model. A thesis submitted to the Graduate Faculty of North Carolina State University. Roman, B., and Brennan, R.A. (2019). A beneficial by-product of ecological wastewater treatment: An evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture. Ecological Engineering X, 1, 100004. https://doi.org/10.1016/j.ecoena.2019.100004 Roser, M., and Ritchie, H. (2017). Yields and land use in agriculture. Published online at OurWorldInData.org Scott-Thomas, C. (2015). What’s next for protein? Press release, July 20. Crawley, United Kingdom. foodnavigator.com Sharma, J., Clark, W.D., Shrivastav, A.K., Goswami, R.K., Tocher, R.D., and Chakrabarti, R. (2019). Production potential of greater duckweed Spirodela polyrhiza (L. Schleiden) and its biochemical composition evaluation. Aquaculture, 513(15), 734419. https://doi.org/10.1016/j.aquaculture.2019.734419 Sonta, M., Rekiel, A., and Batorska, M. (2019). Use of duckweed (Lemna L.) in sustainable livestock production and aquaculture – a review. Annals of Animal Science, 19(2), 257-271. https://doi.org/10.2478/aoas-2018-0048 Steinfeld, H., Gerber, P., Wassenaar, T., Castel, V., Rosales, M., and De Haan, C. (2006). Livestock’s long shadow: environmental issues and options. The Livestock, Environment and Development Initiative. Rome, Italy. Tavares, F.A., Rodrigues, J.B.R., Fracalossi, D.M., Esquivel, J., and Roubach, R. (2008). Dried duckweed and commercial feed promote adequate growth performance of tilapia fingerlings. Biotemas, 21(3), 91–97. https://doi.org/10.5007/2175-7925.2008v21n3p91 Van, B.H., Men, L.T., Son, V.V., Preston, T.R. (1997). Duckweed (Lemna spp.) as protein supplement in an ensiled cassava root diet for fattening pigs. Livestock Research for Rural Development, 9(1). Vladimirova, I.N., and Georgiyants, V.A. (2014). Biologically active compounds from Lemna minor S.F. gray. Pharmaceutical Chemistry Journal, 47, 599-601. https://doi.org/10.1007/s11094-014-1016-8

Witkowska, Z., Saeid, A., Chojnacka, K.W., Dobrzański, Z., Górecki, H., Michalak, I., Korczynski, M., and Opaliński, S. (2012). New biological dietary feed supplement for laying hens with microelements based on duckweed (Lemna minor). American Journal of Agricultural and Biological Science, 7(4), 482. https://doi.org/10.3844/ajabssp.2012.482.493 World Health Organization (2007). Protein and amino acid requirements in human nutrition. WHO Technical Report Series 935.

Chapter 6 – Conclusions, significance, and future work

This dissertation critically evaluated the use of wastewater-grown and nutrient media-grown duckweed as a protein source for animal feed by determining its protein production potential, microbiological safety, environmental impacts, and protein quality. These results were obtained through duckweed growth experiments, microbiological techniques, life-cycle assessment modeling, and an in vivo animal feeding study. The results for duckweed were compared to previously conducted duckweed studies and commonly used crops for livestock, and it was determined that wastewater-grown duckweed can be safely used as a sustainable protein supplement for livestock or fish feed.

The majority of the research done in this dissertation was done within or sampling from the Penn

State Eco-MachineTM (a 1,000 gpd ecological wastewater treatment system located in a greenhouse on the

University Park Campus). Since this facility is fed with wastewater collected from the Penn State

Wastewater Treatment Plant, there is a high variability in the organic matter and nutrient concentrations throughout the year (i.e., in the summer when there are few students, the wastewater becomes very

‘weak’). This variability within the treatment system makes it difficult to produce duckweed biomass with a consistent growth rate and protein content year-round. In fact, in summer when duckweed should be growing at its fastest rate, the nutrient concentrations in the Eco-MachineTM drop significantly, which could underestimate the quantity of protein that duckweed can produce. The duckweed grown in this facility is a polyculture of Lemna japonica/minor and Wolffia columbiana, which is a naturally occurring co-culture found in the region. Since the amino acid composition of different genera of duckweed are similar, it is expected that other duckweed species would have comparable protein content when grown in the Eco-MachineTM (assuming that the nutrient concentrations and environmental conditions remain similar).

Although this dissertation focused on duckweed as a feed/supplement for animals, duckweed is commonly consumed by humans in parts of the world. Specifically, the genus Wolffia, named “Khai- nam”, or “eggs of the water”, is sold in vegetable markets in several Asian countries as an ingredient for

93 salads, omelets, etc. Wolffia is the smallest of the duckweed species, making it ideal to be used as a food ingredient without having to grind or mill the plant into a small, uniform size. Moreover, Wolffia is the only duckweed genera that does not store its oxalate content in the form of calcium oxalate crystals, which can cause kidney stones (Appenroth et al., 2017). Thus, Wolffia seems to be the most applicable to be used as human food. In fact, several companies have started selling duckweed as a ‘super food’, to be added to meals or used as a protein supplement. In addition to humans, it may be beneficial to use Wolffia over other duckweed genera for longer living domestic animals (ex., cattle and sheep can develop kidney or bladder stones; Newsom, 1938). However, for shorter living animals, like broiler chickens with an average lifespan of 40 days, using native duckweed species to the area is likely sufficient, but this an area that future duckweed feeding research should examine in more detail.

In addition, a better understanding of the interaction between microbial communities in the duckweed rhizosphere when grown on wastewater should be explored in future research. Natural wastewaters have complex microbial communities, which utilize nutrients within the wastewater for their own growth. Hence, it is possible that bacteria in the duckweed rhizosphere could be competing with duckweed for aqueous nutrients. Conversely, bacteria in the duckweed rhizosphere could assist in mobilizing nutrients that would otherwise be trapped within solids, increasing the nutrient availability for duckweed. Similar microbe-plant interactions have been shown to have substantial effects on the productivity of soil-grown plants.

To improve the overall understanding of utilizing ecological wastewater treatment systems coupled with protein production, it is recommended that pilot-scale studies on individual farms be conducted. These studies should examine the efficacy of growing duckweed (or other protein-rich aquatic plants) on various agricultural waste streams (fertilizer runoff or animal manure) to be utilized as a source of protein for the animals raised on that farm. A short manuscript addressing the potential of agricultural operations utilizing wastewater-grown duckweed as a protein source for dairy cattle is included in

Appendix A. In summary, duckweed is theoretically capable of producing 5.8x more protein and assimilating 6.6x N and 5.8x P than soybean per acre when grown on dairy-cow manure. These results

94 suggest that utilizing duckweed over soybean within livestock operations could increase the amount of protein produced at the point of use, providing economic benefits to the farmers, while simultaneously reducing the detrimental environmental impacts that come with agricultural operations. In addition, utilizing duckweed over soybean can reduce the nutrient load to receiving waters from livestock operations, improving the overall environmental health the watershed. Future work should be conducted to determine how these wastewater-to-food systems scale with treatment capacity, and how the growth and development of animals are affected by increasing the percentage of duckweed protein in their diet.

In addition to utilizing wastewater-grown duckweed for agriculture, ecological wastewater treatment systems producing protein in the form of plant biomass can overcome two major issues in the developing world: access to sewage treatment and malnutrition. Although some sustainable wastewater treatment systems have been implemented in the developing world, many of these systems are left unmaintained and fail. The reasons for these failures vary by location, from cultural/social issues in parts of Africa to poor policies and infrastructure in South America (UN-Water, 2017). To combat these failing systems, the Dublin-Rio Principles were implemented ~25 years ago, which recommends the use of sustainable planning for successful implementation of wastewater treatment systems in developing communities. Creating a source of protein from a wastewater treatment system may provide additional incentive for the local regions to install and maintain their wastewater treatment system.

The work done in this dissertation is the first to evaluate coupling ecological wastewater treatment with protein production via duckweed. The results indicate that these waste-to-food systems can greatly reduce the environmental impacts of conventional wastewater treatment and food production methods. In order to sustain a growing food demand without compromising the health of the environment, wastewater treatment and food production methods must become inherently interconnected to create a sustainable society.

References

Appenroth, K.J., Sree, K.S., Bohm, V., Hammann, S., Vetter, W., Leiterer, M., and Jahreis, G. 2017. Nutritional value of duckweeds (Lemnaceae) as human food. Food Chemistry, 217:266-273. https://doi.org/10.1016/j.foodchem.2016.08.116 Newsom, I.E. 1938. Urinary calculi with special reference to cattle and sheep. Journal of the American Veterinary Medical Association, 92:495-502. UN-Water 2017. The United Nations World Water Development Report. https://reliefweb.int/sites/reliefweb.int/files/resources/247153e.pdf

Appendix A – Duckweed grown on dairy cow manure - theoretical protein production and nitrogen and phosphorous removal in the Chesapeake Bay Watershed

The following analyses were done in the support of a proposal to the NSF ECO-CBET program that was subsequently awarded

Analysis conducted by Ben Roman

Proposal PI: Rachel Brennan. Co-PIs: Costello, Curtis, Hristov, McPhillips.

The Chesapeake Bay Watershed is largest watershed in the United States, spanning 64,000 square miles

(165,000 km2) and encompassing parts of six states in the northeast. The watershed includes 11,684 miles

(18,800 km) of shoreline, 150 major rivers and streams, and is home to over 18 million people. The land- to-water ratio of the Chesapeake Bay Watershed is the largest in the world at 14:1, meaning the actions on land have incredibly large impacts on the Bay’s environmental health (Chesapeake Bay Program, 2020).

Nutrient pollution is the leading cause of the Bay’s poor health, which causes algal blooms that create hypoxic environments and is responsible for declining populations of submerged grasses, blue crab, oysters, and fish species (US EPA, 2017). It is estimated that the annual load of N into the Bay is roughly

290 million lbs (130 million kg) and the annual load of P is roughly 21 million lbs (9.7 million kg)

(Moyer & Blomquist, 2020). N and P pollution come from a range of sources, including sewage treatment plants, agricultural fields, lawns, etc. Among these sources, manure from livestock operations is responsible for 19% of the total N and 26% of the total P load to the Bay (Eney, 2009). Nutrient management for livestock operations varies, but commonly the manure is applied on site as fertilizer to a forage crop (i.e., soybean) that is used to feed the livestock (Higgins and Wightman, 2012). This small modeling study compares the potential of duckweed and soybean for sequestering nutrients from dairy cow liquid manure and upcycling them into protein in the northeastern United States.

N and P removal from the Chesapeake Bay Watershed

Known N and P uptake rates of duckweed and soybean were used to calculate the theoretical growth area required to remove various percentages of N and P from livestock manure entering the Chesapeake Bay

(Figure A.1; see Table A.S1 for spreadsheet). Duckweed and soybean were both assumed to have a 7- month growing period, where duckweed was harvested weekly and soybean was harvested once per year.

The amount of N and P released from manure into the Chesapeake Bay was estimated using an average total N loading of 7.0 lb/acre-yr and total P loading of 0.5 lb/acre-yr for the entire watershed (Eney,

2009), and that 19% of the total N and 26% of the total P load was generated from manure (Moyer and

Blomquist, 2020). The results indicate that duckweed can sequester ~6x more N and P than soybean when grown on the same land area.

100% Duckweed P 90% Soybean 80% 70% N 60% 50% 40% 30% P Chesapeake Bay from manurefromBay Chesapeake 20% % N and P removed from entering the entering from removed P and % N 10% N 0% 0 10,000 20,000 30,000 40,000 Area (acre)

Protein requirement met, and N and P removal based on number of animals and growth area

The following analysis compares the percentage of cow protein requirement met, and N and P removed, by growing duckweed or soybean using dairy cow manure as a growth medium or fertilizer. Duckweed, soybean, and dairy cow manure characteristics are listed in Table A.1. Similar to N and P removed, duckweed is capable of producing 5.8x more protein than soybean per area (Figure A.2; see Table A.S2 for spreadsheet). This analysis uses the number of animal units (AU; equivalent to 1,000 pound dairy cow) divided by the available plant growth area (AU/acres) to encompass varying sizes of dairy farms, and to allow farmers to see how much protein can be produced and N and P removed depending on the area allowed for duckweed or soybean growth. The following equations were used to calculate the data presented in Figure A.2.

Table A.1. Characteristics of dairy cow manure, duckweed and soybean used for this analysis. Characteristic Value Citation Typical dairy cow weight 1,550 lb (1.55 AU) Penn State Extension, 2017 Liquid manure production 13 gal/AU-day Penn State Extension, 2017 Liquid manure N content 0.028 lb N/gal Penn State Extension, 2017 Liquid manure P content 0.013 lb P/gal Penn State Extension, 2017 Duckweed growth rate 40.5 kg DM/acre-day Leng, 1999 Duckweed protein content 40% DM Leng, 1999 Duckweed N uptake 2.23 kg/acre-day Zimmo et al., 2004 Duckweed P uptake 0.49 kg/acre-day Leng, 1999 Soybean growth rate 1,340 kg/acre-harvest USDA NASS, 2017 Soybean protein content 44% DM USDA NASS, 2017 Soybean N uptake 72.2 kg N/acre-harvest Gaspar et al., 2017 Soybean P uptake 17.8 kg P/acre-harvest Silva, 2017

Equation A.1:

% 퐴푈 푝푟표푡푒푖푛 푟푒푞푢푖푟푒푚푒푛푡

푘푔 푝푟표푡푒푖푛 푝푙푎푛푡 푝푟표푡푒푖푛 푦푖푒푙푑 푎푐푟푒 − 푦푟 = ∗ 100 푘푔 푝푟표푡푒푖푛 퐴푈 퐴푈 푝푟표푡푒푖푛 푟푒푞푢푖푟푒푚푒푛푡 ∗ 퐴푈 푝푒푟 푝푙푎푛푡 푔푟표푤푡ℎ 푎푟푒푎 퐴푈 − 푦푟 푎푐푟푒

Equation A.2:

% 푁 푎푛푑 푃 푟푒푚표푣푒푑

푘푔 푁 표푟 푃 푝푙푎푛푡 푁 푎푛푑 푃 푢푝푡푎푘푒 푎푐푟푒 − 푦푟 =1− 1− 푘푔 푁 표푟 푃 퐴푈 푚푎푛푢푟푒 푁 푎푛푑 푃 푐표푛푡푒푛푡 ∗ 퐴푈 푝푒푟 푝푙푎푛푡 푔푟표푤푡ℎ 푎푟푒푎 퐴푈 −푦푟 푎푐푟푒

∗ 100

100%

80%

60%

40% removal, and P removal P and removal, 20% % of AU protein requirement met, N N met, requirement protein AU% of

0% 10 20 30 40 50 60 70 80 90 100 AU/growth area (AU/acre) Figure A.2. Comparison of duckweed (DW) and soybean (SB) for fraction of protein requirements met for a 1000 dairy cow herd, with N- and P- uptake from manure fertilizer by both crops based on number of AU per duckweed/soybean growth area (assumes a 7-month growing season for duckweed and one harvest per year for soybean).

Techno-economic analysis of a theoretical 1000 dairy cow herd with 0-20 acres of plant growth area.

A techno-economic analysis (TEA) was performed to compare growing duckweed vs. soybean on dairy cow manure to produce protein for the cows’ diet (Figure A.3 A: see Table A.S3 for spreadsheet). Land purchasing costs were ignored, since it is assumed this will occur on an existing farm. Costs for excavating a lagoon for duckweed was set to $2,000/acre, but ignored for soybean. Harvester and dryer

100 purchase costs were determined by estimating the required sizes and pricing units with these dimensions:

$60,000 for harvester (JuLong aquatic weeds harvester, Shandong, China); and $50,000 for dryer (Sunrise drying oven, Henan, China). The cost of ground/surface water used for diluting the dairy manure was set to $40/acre irrigated. Drying costs were set to $10/MT. The market price for duckweed and soybean was assumed to be equal, which was $360/MT. The results indicate that even with the additional cost of excavation, duckweed outproduces soybean in both protein production and monetary value, and well as removes more N and P from the dairy manure (Figure A.3 B).

$350k 20% $80k DW DW $300k SB SB $250k % cow protein 15% $60k 10 acre $200k % N removed $150k % P removed $100k 5 acre $ Value 10% $40k $50k

production $k 10 acre

-$50k N removal, and removalP 5% $20k 5 acre -$100k % of cow protein requirement met, -$150k Net present Net worth present of duckweed/soybean -$200k 0% $k 0 5 10 15 20 0 5 10 15 20 Year of operation Growth area (acre) Figure A.3. Duckweed vs. soybean production on equivalent land areas: A) techno-economic analysis; B) fraction of protein requirements met for a 1000 dairy cow herd, with N- and P- uptake from manure fertilizer by the two crops. Analysis assumes a 7-month growing season in NE USA, and a crop value of $360/MT.

Conclusions and future work

The results from this study suggest that dairy farmers should utilize duckweed over soybean as a nutrient management and protein production strategy. Future studies should investigate: 1) the capability of duckweed to produce safe and high quality protein when grown on dairy manure; 2) the economic and environmental impacts of using large areas of standing water (required for duckweed growth) compared

101 to soil (for soybean growth); 3) and vertical farming techniques for growing duckweed to reduce the overall footprint of duckweed growth operations.

102

References Chesapeake Bay Program (2020) “Watershed” Published online at: chesapeakebay.net. Retrieved from: https://www.chesapeakebay.net/discover/watershed Eney, L. (2009) "What are the main sources of pollution to the Bay?" Chesapeake Bay Program. Gaspar, A.P.; Laboski, C.A.M.; Naeve, S.L.; and Conley, S.P. (2017) “Dry matter and nitrogen uptake, partitioning, and removal across a wide range of soybean seed yield levels.” Crop Science, 57:2170-2182. Higgins, S.; and Wightman, S. (2012) “Nutrient management concepts for livestock producers.” University of Kentucky College of Agriculture, Food, and Environment. Biosystems and Agricultural Engineering. Cooperative Extension Service. Lexington, KY. Leng, R.A., 1999. Duckweed: a tiny aquatic plant with enormous potential for agriculture and environment. Food & Agriculture Organization, Animal Production and Health Division. Rome, Italy. Moyer, D.L.; and Blomquist, J.D. (2020) “Summary of nitrogen, phosphorous, and suspended-sediment loads and trends measured at the Chesapeake Bay Nontidal Network Stations for water years 2009-2018.” USGS CBRIM. Penn State Extension (2017) “Manure management plan. Nutrient balance worksheet user guide.” The Pennsylvania State University. University Park, PA. Silva, G. (2017) “Nutrient removal rates by grain crops.” Michigan State University Extension. United States Department of Agriculture (2017) “USA Soybean yield trends.” Quick Stats. National Agricultural Statistics Service. United States Environmental Protection Agency (2017) “Addressing nutrient pollution in the Chesapeake Bay.” Published online at: epa.gov. Retrieved from: https://www.epa.gov/nutrient-policy-data/addressing- nutrient-pollution-chesapeake-bay Zimmo O.R., Van der Steen N.P., and Gijzen H.J. (2004) “Nitrogen mass balance across pilot-scale algae and duckweed-based wastewater stabilization ponds.” Water Research, 38(4):913–20.

103

Supplemental information

Table A.S1. Spreadsheet calculations of Chesapeake Bay Watershed theoretical N and P removal Duckweed Soybean % of Area Chesapeake N removed N removed P removed P removed N removed P removed P removed (acre) watershed (kg) (%) (kg) (%) N removed (kg) (%) (kg) (%) 0 0.00% 0 0% 0 0% - 0% - 0% 10,000 0.02% 4,739,042 19% 1,033,973 41% 721,594 3% 178,171 7% 20,000 0.05% 9,478,083 38% 2,067,945 82% 1,443,187 6% 356,343 14% 30,000 0.07% 14,217,125 57% 3,101,918 123% 2,164,781 9% 534,514 21% 40,000 0.10% 18,956,167 77% 4,135,891 164% 2,886,375 12% 712,685 28% 50,000 0.12% 23,695,208 96% 5,169,864 205% 3,607,969 15% 890,856 35% 60,000 0.15% 28,434,250 115% 6,203,836 246% 4,329,562 17% 1,069,028 42% 70,000 0.17% 33,173,292 134% 7,237,809 288% 5,051,156 20% 1,247,199 50% 80,000 0.20% 37,912,333 153% 8,271,782 329% 5,772,750 23% 1,425,370 57% 90,000 0.22% 42,651,375 172% 9,305,755 370% 6,494,344 26% 1,603,542 64% 100,000 0.24% 47,390,417 191% 10,339,727 411% 7,215,937 29% 1,781,713 71% 110,000 0.27% 52,129,458 211% 11,373,700 452% 7,937,531 32% 1,959,884 78% 120,000 0.29% 56,868,500 230% 12,407,673 493% 8,659,125 35% 2,138,056 85% 130,000 0.32% 61,607,542 249% 13,441,645 534% 9,380,719 38% 2,316,227 92% 140,000 0.34% 66,346,584 268% 14,475,618 575% 10,102,312 41% 2,494,398 99% 150,000 0.37% 71,085,625 287% 15,509,591 616% 10,823,906 44% 2,672,569 106%

104

Table A.S2. Spreadsheet calculations of percentage of cow protein requirement, N removed, and P removed in AU/acre % protein % N removal % P removal AU/acre DW SB DW SB DW SB 1 1444% 247% 786% 120% 370% 64% 5 289% 49% 157% 24% 74% 13% 10 144% 25% 79% 12% 37% 6% 15 96% 16% 52% 8% 25% 4% 20 72% 12% 39% 6% 18% 3% 25 58% 10% 31% 5% 15% 3% 30 48% 8% 26% 4% 12% 2% 35 41% 7% 22% 3% 11% 2% 40 36% 6% 20% 3% 9% 2% 45 32% 5% 17% 3% 8% 1% 50 29% 5% 16% 2% 7% 1% 55 26% 4% 14% 2% 7% 1% 60 24% 4% 13% 2% 6% 1% 65 22% 4% 12% 2% 6% 1% 70 21% 4% 11% 2% 5% 1% 75 19% 3% 10% 2% 5% 1% 80 18% 3% 10% 1% 5% 1% 85 17% 3% 9% 1% 4% 1% 90 16% 3% 9% 1% 4% 1% 95 15% 3% 8% 1% 4% 1% 100 14% 2% 8% 1% 4% 1%

105

Table A.S3. Spreadsheet calculations for percentage of cow protein requirement, N removed, and P removed for 1000 cow herd. % of cow protein Land area requirement % N removal % P removal Monetary Value (acre) Duckweed Soybean Duckweed Soybean Duckweed Soybean Duckweed Soybean 0 0% 0% 0% 0% 0% 0% $ - $ - 5 5% 1% 2% 0% 1% 0% $ 15,297 $ 2,404 10 9% 2% 5% 1% 2% 0% $ 30,594 $ 4,808 15 14% 2% 7% 1% 4% 1% $ 45,891 $ 7,212 20 19% 3% 10% 2% 5% 1% $ 61,189 $ 9,615

106

Appendix B – Chapter 2 additional files

107

Table B.1. Harvested duckweed masses, densities, and growth rates over the course of the experiment. Mass harvested (g/m2/day) Dry Density (g/m2) Growth rate (g/m2/day) Date (2017) CA1 CA2 OA1 OA2 OA3 CA1 CA2 OA1 OA2 OA3 CA1 CA2 OA1 OA2 OA3 Jan. 24 0.7 0.6 0.8 0.6 0.8 26.1 26.5 22.4 16.8 23.7 2.8 2.6 3.0 2.3 3.4 Jan. 31 0.5 0.4 0.7 0.7 0.7 21.1 13.1 18.7 19.6 20.3 2.2 1.6 2.7 2.8 2.9 Feb. 7 0.5 0.6 0.9 0.8 0.9 21.3 18.0 24.3 23.5 24.6 2.0 2.5 3.5 3.4 3.5 Feb. 14 0.8 0.8 1.2 0.9 0.9 23.9 22.4 32.4 24.5 23.8 3.1 3.1 4.6 3.5 3.4 Feb. 21 1.0 1.0 1.2 0.9 1.1 29.0 28.0 34.9 26.3 30.5 4.0 4.0 5.0 3.8 4.4 Feb. 28 1.2 1.2 1.5 1.0 1.1 32.9 34.9 42.1 28.3 32.0 4.7 5.0 6.0 4.0 4.6 Mar. 7 1.2 1.2 1.6 1.3 1.3 34.2 35.0 43.5 35.4 35.3 4.9 5.0 6.2 5.1 5.0 Mar. 14 1.8 1.5 2.0 1.5 1.3 49.0 42.3 54.9 41.9 35.4 7.0 6.0 7.8 6.0 5.1 Mar. 19 2.0 2.2 2.3 1.8 1.7 40.0 43.9 45.1 35.3 33.6 8.0 8.8 9.0 7.1 6.7 Mar. 24 2.6 2.2 2.5 1.9 1.7 52.1 44.8 49.0 38.3 33.1 10.4 9.0 9.8 7.7 6.6 Mar. 29 2.0 2.1 2.2 1.9 1.5 40.0 42.6 44.4 37.8 30.5 8.0 8.5 8.9 7.6 6.1 Apr. 3 2.1 2.3 2.2 1.7 1.5 41.6 45.1 44.4 33.9 29.5 8.3 9.0 8.9 6.8 5.9 Apr. 8 2.4 2.4 2.4 1.9 1.6 47.4 48.4 48.5 37.2 32.4 9.5 9.7 9.7 7.4 6.5 Apr. 13 2.6 2.0 3.1 2.2 2.0 51.4 40.2 62.3 43.9 40.2 10.3 8.0 12.5 8.8 8.0 Apr. 18 3.3 2.5 3.2 2.9 2.0 66.2 49.8 63.4 58.8 40.2 13.2 10.0 12.7 11.8 8.0 Apr. 23 3.0 2.5 3.3 2.7 2.5 60.6 49.3 65.3 53.9 50.4 12.1 9.9 13.1 10.8 10.1 Apr. 28 6.3 6.0 7.0 5.9 4.1 63.5 59.7 70.0 59.0 40.8 12.7 11.9 14.0 11.8 8.2 May 3 4.8 3.3 4.0 3.7 4.2 48.4 33.4 39.7 37.1 41.7 9.7 6.7 7.9 7.4 8.3 May 8 4.8 3.0 3.5 3.2 3.7 47.9 30.1 34.8 31.6 36.7 9.6 6.0 7.0 6.3 7.3 May 13 4.4 2.8 3.5 3.2 4.0 43.6 28.0 35.1 32.0 40.0 8.7 5.6 7.0 6.4 8.0 May 18 4.2 3.3 3.8 3.2 4.7 42.2 32.6 38.0 32.1 47.0 8.4 6.5 7.6 6.4 9.4 May 23 4.0 3.3 3.6 3.1 4.5 39.8 32.9 35.6 30.7 45.1 8.0 6.6 7.1 6.1 9.0 May 28 4.1 2.8 3.2 2.7 5.0 40.6 27.9 32.2 27.2 49.7 8.1 5.6 6.4 5.4 9.9 Avg. 2.6 2.1 2.5 2.1 2.3 41.3 34.6 41.5 34.1 35.7 7.7 6.5 7.8 6.4 6.7 St. Dev 1.4 0.9 1.0 0.9 1.5 12.8 11.2 13.1 10.4 9.1 3.4 2.7 3.1 2.5 2.3

108

109

110

111

112

Figure B.1. Duckweed protein content results from Cumberland Valley Analytical Services.

113

Table B.2. Duckweed biomass metal concentrations from ICP-MS performed by Penn State Laboratory for Isotopes and Metals in the Environment. Element/Species Al Cr Fe Ni Cu Zn As Ag Cd Pb units ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL 35-MilliQ Blank A 5.4 0.029 0.577 nd 2.703 0.280 nd 0.041 0.002 0.016 36-MilliQ Blank B 5.7 nd 0.614 nd 3.034 0.283 nd 0.013 0.001 0.017 37-MilliQ Blank C 6.6 nd 0.695 nd 3.072 0.312 nd 0.005 0.001 0.015 38-Distilled Blank A 6.6 0.074 1.225 nd 5.495 0.495 nd nd 0.001 1.471 39-Distilled Blank B 5.8 0.034 1.128 nd 5.422 0.418 nd nd 0.001 1.469 40-Distilled Blank C 6.8 nd 5.725 nd 5.621 0.441 nd nd 0.001 1.444 31-9-15 Blank clear 2.5 0.116 0.854 nd 0.200 0.237 nd 0.014 nd 0.016 32-9-23 Blank stain 2.18 0.131 1.012 nd 0.076 0.285 nd nd 0.001 0.007 33-9-30 Blank stain2 1.55 0.129 1.090 nd 0.139 0.227 nd nd nd 0.013 34-Peach STD clear 334 1.266 321.2 0.879 5.731 26.46 nd nd 0.041 1.203 01-2-28 CA1 A 158 0.646 440.8 4.793 20.58 63.75 0.3 0.008 0.157 2.987 02-2-28 CA1 B 146 0.625 420.5 4.503 19.38 60.81 0.3 0.009 0.024 2.238 03-2-28 CA2 A 148 0.635 635.3 2.984 27.10 68.20 0.5 0.014 0.034 3.734 04-2-28 CA2 B 151 0.775 621.3 2.847 29.98 63.90 0.4 0.017 0.038 4.811 05-2-28 OA1 A 96.0 0.723 433.4 1.875 22.01 99.40 0.4 0.014 0.040 4.542 06-2-28 OA1 B 204 0.765 444.3 1.857 20.94 94.39 0.3 0.013 0.042 5.11 07-2-28 OA2 A 375 0.901 357.2 2.103 30.04 156.3 0.3 0.016 0.075 5.193 08-2-28 OA2 B 144 0.913 315.7 1.990 28.92 152.7 0.4 0.018 0.072 4.677 09-2-28 OA3 A 108 0.856 190.3 1.274 15.17 43.24 0.2 0.014 0.027 6.478 10-2-28 OA3 B 107 0.971 189.2 1.317 14.32 40.77 0.2 0.014 0.02 8.485 11-3-29 CA1 A 91.1 0.587 510.3 2.551 24.97 51.64 0.3 0.067 0.063 2.816 12-3-29 CA1 B 95.6 0.642 528.1 2.546 24.88 52.46 0.3 0.028 0.029 2.814 13-3-29 CA2 A 93.9 0.671 516.1 2.939 19.24 49.25 0.3 0.015 0.023 2.625 14-3-29 CA2 B 104 0.620 551.4 2.975 20.05 52.26 0.3 0.016 0.026 2.688 15-3-29 OA1 A 74.4 0.419 435.3 1.434 17.43 69.84 0.3 0.010 0.043 2.019 16-3-29 OA1 B 77.0 0.420 424.3 1.395 16.64 70.65 0.2 0.008 0.026 1.768 17-3-29 OA2 A 78.7 0.376 239.3 1.113 18.73 83.88 0.2 0.005 0.041 1.621 18-3-29 OA2 B 94.1 0.391 244.6 1.170 19.59 85.11 0.2 nd 0.035 1.522 19-3-29 OA3 A 114 0.578 198.1 1.136 15.53 50.80 0.2 0.007 0.02 2.682 20-3-29 OA3 B 104 0.503 176.1 1.035 12.79 45.92 0.1 0.006 0.019 2.618 21-4-8 CA1 A 88.6 0.679 481.5 2.162 21.98 44.58 0.3 0.018 0.023 4.568 22-4-8 CA1 B 83.3 0.628 469.0 2.027 21.60 43.77 0.3 0.015 0.019 4.503 23-4-8 CA2 A 86.9 0.674 559.7 2.027 17.60 50.52 0.3 0.012 0.023 1.589 24-4-8 CA2 B 77.2 0.542 487.6 1.751 15.55 44.17 0.3 0.009 0.02 1.564 25-4-8 OA1 A 51.5 0.352 367.6 1.213 15.51 66.91 0.2 0.005 0.022 1.313 26-4-8 OA1 B 57.4 0.379 370.4 1.219 16.75 66.35 0.2 0.005 0.021 1.687 27-4-8 OA2 A 78.9 0.430 235.4 1.025 17.63 69.60 0.2 nd 0.025 1.637 28-4-8 OA2 B 72.4 0.363 228.6 1.028 16.87 66.88 0.2 nd 0.032 1.506 29-4-8 OA3 A 225 0.636 206.6 1.204 11.83 41.44 0.1 0.008 0.015 2.478 30-4-8 OA3 B 276 0.631 198.3 1.175 10.51 42.45 0.2 0.008 0.014 2.385 45-9 DDI A 2.16 nd 1.182 nd 0.608 0.199 nd nd 0.002 0.037 46-9 DDI B 2.62 nd 0.965 nd 0.539 0.183 nd nd 0.001 0.032 47-9 DDI C 2.57 nd 1.146 nd 0.573 0.213 nd nd 0.002 0.035 48-128 DDI A 2.41 nd 0.366 nd 2.082 0.328 nd nd 0.002 0.036

114

49-128 DDI B 2.67 nd 0.984 nd 2.170 0.384 nd nd 0.001 0.039 50-128 DDI C 2.16 nd 0.452 nd 2.148 0.376 nd nd 0.001 0.039 41-10-21 Blank 3.30 nd 1.850 0.043 0.177 0.665 nd nd 0.001 0.019 42-10-28 Blank 2.05 0.071 2.408 nd 0.170 0.341 nd nd 0.001 0.019 43-11-2 Blank nd 0.159 1.529 nd 0.096 0.229 nd nd 0.001 0.010 44-Peach STD 407 1.705 345.7 1.065 5.650 28.62 0.1 0.020 0.038 1.379 1-2-14 CA1 A 158 0.995 600.0 5.168 55.85 79.52 0.4 0.024 0.083 6.254 2-2-14 CA1 B 97.7 0.564 395.0 6.110 30.94 49.33 0.3 0.010 0.033 5.824 3-2-14 CA2 A 99.6 0.802 323.4 3.183 27.92 69.51 0.3 nd 0.037 5.914 4-2-14 CA2 B 148 0.754 299.2 1.764 25.97 55.69 0.3 nd 0.029 6.210 5-2-14 OA1 A 137 0.717 268.3 1.529 14.60 52.75 0.2 nd 0.027 6.024 6-2-14 OA1 B 56.2 0.662 271.0 1.437 15.23 53.52 0.3 nd 0.035 6.100 7-2-14 OA2 A 95.0 0.918 242.3 1.472 18.68 71.59 0.2 nd 0.026 8.205 8-2-14 OA2 B 104 0.898 248.6 1.477 18.57 70.94 0.3 nd 0.032 7.271 9-2-14 OA3 A 212 1.877 228.0 1.417 23.25 48.60 0.2 nd 0.030 6.760 10-2-14 OA3 B 284 1.991 231.6 1.635 23.27 49.95 0.2 0.010 0.026 7.040 11-4-18 CA1 A 88.9 0.442 366.4 2.145 20.11 49.20 0.2 0.013 0.036 1.588 12-4-18 CA1 B 89.2 0.457 358.9 2.015 20.07 48.81 0.2 0.011 0.021 1.347 13-4-18 CA2 A 43.0 0.439 424.3 1.888 12.83 41.45 0.3 nd 0.014 1.624 14-4-18 CA2 B 45.1 0.494 423.7 1.738 12.78 40.71 0.2 nd 0.016 1.502 15-4-18 OA1 A 27.0 0.275 358.3 1.141 13.10 57.64 0.2 0.020 0.021 1.509 16-4-18 OA1 B 27.1 0.272 358.5 1.038 13.16 57.88 0.2 nd 0.018 1.741 17-4-18 OA2 A 26.3 0.244 224.3 0.886 14.08 58.28 0.1 nd 0.023 5.301 18-4-18 OA2 B 27.6 0.263 217.0 0.807 13.59 55.43 0.2 nd 0.024 4.264 19-4-18 OA3 A 64.9 0.450 205.9 0.827 12.69 46.18 0.2 nd 0.020 2.708 20-4-18 OA3 B 78.3 0.376 206.4 0.812 13.00 48.89 0.1 nd 0.024 2.666 21-4-28 CA1 A 137 0.637 722.8 2.662 28.54 65.51 0.3 0.014 0.044 2.927 22-4-28 CA1 B 145 0.584 747.4 2.739 29.59 67.88 0.3 0.020 0.044 3.046 23-4-28 CA2 A 106 0.528 660.2 2.455 25.49 56.80 0.4 0.011 0.043 2.781 24-4-28 CA2 B 114 0.503 655.9 2.482 25.02 56.62 0.3 nd 0.050 2.377 25-4-28 OA1 A 60.6 0.468 586.0 1.426 27.14 66.43 0.2 nd 0.028 3.625 26-4-28 OA1 B 83.7 0.494 554.3 1.425 26.93 70.66 0.2 nd 0.029 2.361 27-4-28 OA2 A 55.5 0.331 289.8 0.981 24.78 76.68 0.1 nd 0.033 3.761 28-4-28 OA2 B 55.3 0.300 283.0 0.991 23.79 76.83 0.2 nd 0.038 3.724 29-4-28 OA3 A 103 0.411 227.2 1.008 14.57 49.67 0.2 nd 0.018 4.016 30-4-28 OA3 B 111 0.386 234.7 1.031 14.13 50.60 nd nd 0.018 3.596 31-5-8 CA1 A 98.3 0.415 584.3 2.204 21.86 53.57 0.3 nd 0.021 3.434 32-5-8 CA1 B 98.9 0.409 631.7 2.363 23.06 55.82 0.2 nd 0.023 2.987 33-5-8 CA2 A 67.6 0.390 453.2 1.797 16.76 51.90 0.3 nd 0.020 3.849 34-5-8 CA2 B 85.1 0.411 510.9 2.026 19.71 52.73 0.3 nd 0.025 4.552 35-5-8 OA1 A 62.0 0.345 368.2 1.038 16.27 54.72 0.2 nd 0.023 3.684 36-5-8 OA1 B 58.4 0.327 354.7 1.071 15.62 56.41 0.2 nd 0.023 4.646 37-5-8 OA2 A 48.4 0.294 286.0 1.007 20.90 71.18 0.2 nd 0.035 4.681 38-5-8 OA2 B 48.8 0.271 283.0 0.901 20.60 71.88 nd nd 0.039 4.997 39-5-8 OA3 A 82.9 0.400 241.1 1.151 13.37 50.19 0.2 nd 0.018 3.692 40-5-8 OA3 B 100 0.806 255.4 1.268 14.19 51.87 0.1 nd 0.015 3.656 Instrum. Det. Limit 1.0 0.020 0.300 0.150 0.013 0.100 0.1 0.005 0.001 0.006 CA1 = anaerobic tank CA2 = anoxic tank OA1 = aerobic 1 tank

115

OA2 = aerobic 2 tank OA3 = aerobic 3 tank A&B designate replicates

Appendix C – Chapter 3 additional files

Table C.1. Plate counts of total coliforms, fecal coliforms, and E.coli from duckweed grown in the Penn State Eco-MachineTM at different drying temperatures. Drying Total coliform colonies/g duckweed time Air (21.5±0.9 oC) 40oC 60oC (hr) a b c a b c a b c 0 4,494 8,514 7,316 4,494 8,514 7,316 4,494 8,514 7,316 2 4,206 5,334 5,074 179 0 0 4 777 175 259 0 0 0 6 175 355 344 0 0 0 8 356 537 867 0 0 0 12 344 701 87 0 0 0 24 516 1,949 1,574 173 87 87 0 0 0 48 697 1,243 793 0 0 0 0 0 0 72 175 690 448 0 0 0 0 0 0 96 259 87 86 0 0 0 0 0 0 120 177 88 440 0 0 0 0 0 0 144 87 265 267 0 0 0 0 0 0 168 0 88 174 0 0 0 0 0 0

Drying Fecal coliform colonies/g duckweed time Air (21.5±0.9 oC) 40oC 60oC (hr) a b c a b c a b c 0 13,923 14,394 12,516 13,923 14,394 12,516 13,923 14,394 12,516 2 5,408 6,034 7,481 89 0 0 4 518 262 86 0 0 0 6 175 177 86 0 0 0 8 89 537 693 0 0 0 12 86 88 87 0 0 0 24 172 1,861 962 86 0 0 0 0 0 48 697 710 353 0 0 0 0 0 0 72 349 432 269 0 0 0 0 0 0 96 259 87 172 0 0 0 0 0 0 120 88 88 88 0 0 0 0 0 0 144 87 88 89 0 0 0 0 0 0 168 175 0 261 0 0 0 0 0 0

Drying E.coli colonies/g duckweed time Air (21.5±0.9 oC) 40oC 60oC (hr) a b c a b c a b c 0 1,146 1,053 1,410 1,146 1,053 1,410 1,146 1,053 1,410 2 1,631 1,049 1,290 89 0 0 4 345 262 173 0 0 0 6 88 0 86 0 0 0 8 712 0 87 0 0 0 12 172 88 0 0 0 0 24 258 532 437 86 0 87 0 0 0 48 261 355 441 0 0 0 0 0 0 72 87 259 179 0 0 0 0 0 0 96 173 87 258 0 0 0 0 0 0 120 88 0 0 0 0 0 0 0 0 144 0 88 0 0 0 0 0 0 0 168 87 0 0 0 0 0 0 0 0

116

Appendix D – Chapter 4 additional files

Table D.1. SimaPro Characterization of the Penn State Eco-MachineTM operated for 30 years using IMPACT 2002+ V2.15. Non- Respiratory Ionizing Ozone layer Respiratory Aquatic Terrestrial Impact category Carcinogens carcinogens inorganics radiation depletion organics ecotoxicity ecotoxicity Unit kg C2H3Cl eq kg C2H3Cl eq kg PM2.5 eq Bq C-14 eq kg CFC-11 eq kg C2H4 eq kg TEG water kg TEG soil Air compressor 204.5 780.8 23.4 2,941,973.6 0.0 5.8 3,188,014.8 588,925.5 External recycle pump 0.6 2.2 0.1 8,166.7 0.0 0.0 8,849.7 1,634.8 Holding tank pump 3.0 11.6 0.3 43,584.0 0.0 0.1 47,229.0 8,724.7 Humidifier 3.3 12.8 0.4 48,039.7 0.0 0.1 52,057.3 9,616.6 Internal recycle pump 6.1 23.1 0.7 87,168.0 0.0 0.2 94,458.0 17,449.3 CA tank 3.3 6.1 0.1 796.8 0.0 0.0 131,707.3 52,129.4 Furnace 671.5 612.8 56.6 1,316,656.4 0.0 43.2 8,585,570.0 2,025,126.0 Greenhouse glass 31.9 17.6 2.3 6,702.0 0.0 0.3 162,449.4 39,288.7 OA tank 11.1 20.4 0.4 2,655.7 0.0 0.1 438,977.3 173,746.1 Solar panel -180.1 -699.1 -20.8 -2,641,358.7 0.0 -5.2 -2,844,341.1 -523,860.6 Wetland gravel 6.4 9.8 0.9 4,887.3 0.0 0.3 56,814.9 26,256.0 Concrete block foundation 22.8 25.6 1.5 15,012.3 0.0 0.5 122,281.9 51,223.8 Wastewater delivery truck 50.2 61.2 5.4 174,716.2 0.0 4.8 989,997.2 218,347.0 Wetland liner 788.5 22.2 2.7 1,768.0 0.0 7.1 128,851.9 37,357.8 Clarifier 12.1 0.3 0.0 27.2 0.0 0.1 1,981.1 574.4 Holding tank 6.3 15.5 0.8 5,589.7 0.0 0.2 51,852.6 21,274.9 Steel beams 160.6 125.4 4.6 8,177.6 0.0 3.4 499,375.9 122,810.4 Valves and piping 0.7 2.9 0.2 111.9 0.0 0.5 25,746.5 7,650.8 Filter for UV 5.9 0.1 0.0 11.3 0.0 0.0 800.5 258.6 UV disinfection 4.3 14.0 0.4 43,938.5 0.0 0.1 54,294.9 16,161.0 Irrigation water -508.6 -604.5 -32.4 -242,126.8 0.0 -10.3 -1,579,925.7 -941,343.9 Duckweed, tilapia feed -17.5 13.7 -1.8 -4,644.4 0.0 -0.7 -50,809.8 48,904.6 Wastewater, from residence -1,360.3 -4,241.3 -20.0 -61,866.7 0.0 -6.9 -2,168,127.6 -1,155,716.8

117

Total -73.5 -3,767.0 25.7 1,759,986.1 0.0 43.8 7,998,105.9 846,539.1

Terrestrial Land Aquatic Aquatic Global Non-renewable Mineral Impact category acid/nutri occupation acidification eutrophication warming energy extraction Unit kg SO2 eq m2org.arable kg SO2 eq kg PO4 P-lim kg CO2 eq MJ primary MJ surplus Air compressor 453.8 358.9 92.7 3.8 21,813.4 759,060.0 914.8 External recycle pump 1.3 1.0 0.3 0.0 60.6 2,107.1 2.5 Holding tank pump 6.7 5.3 1.4 0.1 323.2 11,245.1 13.6 Humidifier 7.4 5.9 1.5 0.1 356.2 12,394.7 14.9 Internal recycle pump 13.4 10.6 2.7 0.1 646.3 22,490.3 27.1 CA tank 0.8 1.5 0.2 0.0 45.0 739.7 2.0 Furnace 1,020.5 474.6 358.5 16.1 47,467.9 3,125,089.5 427.8 Greenhouse glass 50.1 15.4 14.3 0.2 1,588.8 19,683.2 105.1 OA tank 2.6 4.9 0.6 0.1 150.0 2,465.6 6.6 Solar panel -401.8 -320.1 -81.6 -3.4 -19,406.8 -679,383.1 -809.9 Wetland gravel 19.3 31.2 3.7 0.1 525.6 8,214.7 20.3 Concrete block foundation 32.3 51.7 6.2 0.3 1,705.7 14,697.3 127.1 Wastewater delivery truck 97.6 39.8 36.6 2.0 3,548.3 401,517.9 36.5 Wetland liner 64.8 10.6 18.1 0.1 4,713.7 201,401.5 4.5 Clarifier 1.0 0.2 0.3 0.0 72.5 3,096.6 0.1 Holding tank 16.7 25.3 3.2 0.1 1,075.5 7,838.8 22.1 Steel beams 48.1 35.5 12.4 0.5 2,858.8 32,392.7 256.1 Valves and piping 4.5 0.7 1.0 0.0 323.1 10,060.7 0.7 Filter for UV 0.4 0.1 0.1 0.0 32.0 1,304.5 0.0 UV disinfection 7.2 5.6 1.5 0.1 351.6 11,712.7 18.0 Irrigation water -219.8 -691.2 -59.5 -4.2 -13,950.7 -249,469.9 -901.4 Duckweed, tilapia feed -43.6 -1,602.6 -7.5 -0.5 -1,481.5 -9,585.0 -26.0 Wastewater, from residence -228.9 -293.2 -56.3 -2.6 -13,066.9 -159,439.1 -1,843.9 Total 954.4 -1,828.4 350.5 12.8 39,752.3 3,549,635.4 -1,581.1

118

Appendix E – Chapter 5 additional files

Table E.1. Body weight of mice over the course of the study. Group 1 – control (casein) Day Date Body weight (g) Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 0 28-Jan-20 33.7 31.6 34.1 30.7 32.9 2 30-Jan-20 34.2 32.0 35.0 30.9 33.2 7 4-Feb-20 35.5 33.7 36.9 32.1 33.3 9 6-Feb-20 36.6 33.8 37.1 33.8 34.1 14 11-Feb-20 38.2 35.1 39.2 34.3 34.3 16 13-Feb-20 38.7 35.1 39.7 34.9 35.0 21 18-Feb-20 39.8 35.6 40.2 35.2 35.0 23 20-Feb-20 40.0 35.2 40.5 35.0 35.1 29 26-Feb-20 40.1 35.7 42.8 35.2 35.8

Group 2 – 25% duckweed protein Day Date Body weight (g) Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 0 28-Jan-20 30.7 31.4 31.5 36.7 33.6 2 30-Jan-20 31.4 31.9 33.0 38.2 35.4 7 4-Feb-20 32.5 33.4 34.8 39.6 37.5 9 6-Feb-20 32.4 34 36 40.4 39.1 14 11-Feb-20 33.2 34.9 38.2 40.8 41.1 16 13-Feb-20 34.2 34.7 39.3 40.8 42.2 21 18-Feb-20 34.7 35.6 40.9 41.9 43.8 23 20-Feb-20 34.9 34.9 41.5 41.5 44.6 29 26-Feb-20 35.5 35.4 42.4 41.9 46.0

Group 3 – 25% duckweed protein Day Date Body weight (g) Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 0 28-Jan-20 36.6 33.8 34.3 29.8 31.6 2 30-Jan-20 37.0 33.8 34.2 30.4 30.9 7 4-Feb-20 38.6 35.8 35.2 31.5 32 9 6-Feb-20 39.2 36.3 35.9 32 32.6 14 11-Feb-20 41.8 37.9 37.1 33.2 33.2 16 13-Feb-20 42.5 38.9 37.9 34.2 34.0 21 18-Feb-20 43.3 39.9 38.8 34.9 34.1 23 20-Feb-20 44.0 40.9 40.0 35.6 34.5 29 26-Feb-20 46.0 41.9 40.0 36.0 35.5

Group 4 – control (casein) Day Date Body weight (g) Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 0 28-Jan-20 34.2 32.9 35.6 28.3 32.4 2 30-Jan-20 34.0 33.5 36.5 28.1 32.8 7 4-Feb-20 35.6 35.3 38.1 28 34.4 9 6-Feb-20 36.1 36.2 39.1 29.4 35.6 14 11-Feb-20 36.7 37.8 41.2 32.3 37.4 16 13-Feb-20 36.5 38.2 41.5 32.6 36.9 21 18-Feb-20 36.9 40.0 42.4 33.5 38.2 23 20-Feb-20 36.9 40.8 43.0 34.2 37.8 29 26-Feb-20 37.7 41.1 43.2 34.3 38.6

Group 5 – 10% duckweed protein Day Date Body weight (g) Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 0 28-Jan-20 34.8 32.6 36.4 33.7 37.3 2 30-Jan-20 35.7 32.6 36.7 34.2 37.6 7 4-Feb-20 38.3 34.1 38.4 35.7 40.1 9 6-Feb-20 39.4 34.9 39.4 37 41.3 14 11-Feb-20 41.8 35.8 40.6 38.3 42.7 16 13-Feb-20 42.2 36.5 41.7 39.2 42.8 21 18-Feb-20 43.7 37.0 41.7 40.4 43.4 23 20-Feb-20 44.4 37.6 42.6 41.2 43.6 29 26-Feb-20 43.8 38.0 43.1 41.1 45.4

Group 6 – 10% duckweed protein Day Date Body weight (g) Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 0 28-Jan-20 28.8 35.6 32.2 33.4 36.6 2 30-Jan-20 29.6 35.4 31.9 33.4 37.2 7 4-Feb-20 30.6 36.7 33.5 34.6 38.5 9 6-Feb-20 31.8 38.1 33.7 35.6 39 14 11-Feb-20 33.1 38.8 34.3 36 39.4 16 13-Feb-20 34.2 39.6 35.0 36.9 40.3 21 18-Feb-20 35.2 40.2 35.4 37.7 40.3 23 20-Feb-20 36.1 41.3 36.0 38.2 41.1 29 26-Feb-20 36.5 42.0 36.4 39.2 41.4

Table E.2. Food intake rate of each group over the course of the study (DWP = duckweed protein) Day Date Food intake (g/day) Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 (control) (25% DWP) (25% DWP) (control) (10% DWP) (10% DWP) 2 30-Jan-20 22.3 24.0 24.8 23.0 23.5 22.9 7 4-Feb-20 21.8 23.0 22.4 20.8 23.6 22.0 9 6-Feb-20 18.4 20.6 19.9 20.0 23.2 21.0 14 11-Feb-20 20.5 21.2 21.2 20.9 21.9 20.1 16 13-Feb-20 19.5 19.3 21.7 18.9 22.8 21.2 21 18-Feb-20 19.7 20.4 21.2 20.1 20.2 20.3 23 20-Feb-20 17.6 18.6 21.5 19.5 22.6 21.7 29 26-Feb-20 19.7 20.6 22.0 19.7 21.0 20.2

120

Table E.3. Organ weights of each mouse. Group 1 – control (casein) Mouse Liver Spleen Lungs Kidneys Heart Epididymal fat Colon length (g) (g) (g) (g) (g) (g) (cm) 1 2.00 0.13 0.24 0.62 0.19 1.50 6.0 2 1.97 0.15 0.15 0.54 0.20 0.79 8.0 3 2.09 0.15 0.15 0.72 0.18 1.24 8.0 4 1.71 0.09 0.09 0.59 0.16 1.31 5.5 5 1.86 0.16 0.16 0.61 0.26 0.70 5.5

Group 2 – 25% duckweed protein Mouse Liver Spleen Lungs Kidneys Heart Epididymal fat Colon length (g) (g) (g) (g) (g) (g) (cm) 1 1.86 0.15 0.31 0.59 0.15 0.69 7.0 2 1.79 0.16 0.19 0.61 0.20 0.82 7.0 3 2.20 0.13 0.21 0.62 0.20 1.40 5.5 4 2.03 0.14 0.28 0.76 0.22 1.23 6.0 5 2.17 0.16 0.29 0.71 0.28 2.50 8.5

Group 3 – 25% duckweed protein Mouse Liver Spleen Lungs Kidneys Heart Epididymal fat Colon length (g) (g) (g) (g) (g) (g) (cm) 1 2.60 0.20 0.24 0.73 0.26 2.07 7.0 2 2.54 0.18 0.28 0.75 0.20 1.77 6.0 3 2.58 0.15 0.28 0.87 0.25 0.88 7.0 4 2.26 0.13 0.22 0.59 0.19 1.05 6.2 5 2.34 0.14 0.20 0.61 0.22 0.81 8.3

Group 4 – control (casein) Mouse Liver Spleen Lungs Kidneys Heart Epididymal fat Colon length (g) (g) (g) (g) (g) (g) (cm) 1 2.51 0.14 0.21 0.70 0.19 0.65 8.0 2 2.37 0.15 0.20 0.70 0.25 1.41 7.9 3 2.57 0.17 0.21 0.74 0.23 1.13 8.0 4 2.15 0.22 0.24 0.52 0.20 1.07 6.0 5 2.93 0.15 0.23 0.77 0.17 1.11 7.5

Group 5 – 10% duckweed protein Mouse Liver Spleen Lungs Kidneys Heart Epididymal fat Colon length (g) (g) (g) (g) (g) (g) (cm) 1 2.00 0.15 0.25 0.75 0.23 1.65 8.0 2 1.85 0.13 0.26 0.73 0.22 0.88 6.0 3 2.53 0.13 0.31 0.68 0.24 1.38 7.0 4 2.20 0.13 0.23 0.80 0.21 1.54 7.5 5 2.50 0.16 0.28 0.89 0.32 1.36 7.0

Group 6 – 10% duckweed protein Mouse Liver Spleen Lungs Kidneys Heart Epididymal fat Colon length (g) (g) (g) (g) (g) (g) (cm) 1 2.25 0.13 0.22 0.56 0.19 1.07 7.5 2 1.85 0.19 0.26 0.74 0.25 1.62 5.5 3 2.30 0.14 0.31 0.59 0.19 0.81 6.0 4 2.52 0.15 0.18 0.70 0.24 1.52 6.0 5 2.24 0.15 0.22 0.66 0.24 1.52 6.0

121

Benjamin Roman [email protected] ● 724-678-3414

Education The Pennsylvania State University, Department of Civil and University Park, PA Environmental Engineering Ph.D. in Environmental Engineering, GPA: 3.84/4.00 Expected Dec 2020

M.Eng., Environmental Engineering, GPA: 3.81/4.00 May 2015

B.Sc., Civil Engineering, GPA: 3.33/4.00 May 2014 Minor in Watersheds and Water Resources

Leadership Experience The Penn State Eco-MachineTM University Park, PA Lead graduate student 2016-Present  Led multidisciplinary research experiments in a pilot-scale ecological wastewater treatment system that focused on upcycling nutrients from wastewater into plants (duckweed) to be used for protein, biofuels, or fertilizers  Trained and assisted numerous undergraduate and graduate researchers simultaneously on various laboratory procedures and experiments  Maintained and trained all new users on an ion chromatograph and total organic carbon/total nitrogen analyzer

Environmental, Chemistry, and Microbiology Student Symposium University Park, PA Chair 2017  Chaired a two-day student-run symposium that hosted 3 distinguished keynote speakers and 39 student presenters  Managed a committee of 15 graduate and undergraduate students to run the symposium  Raised over $16,000 from numerous Penn State departments and organizations to fund the symposium

Publications Roman, B., and Brennan, R.A. (2019). A beneficial by-product of ecological wastewater treatment: An evaluation of wastewater-grown duckweed as a protein supplement for sustainable agriculture. Ecological Engineering X, 1, 100004.

Roman, B., Brennan, R.A., and Lambert, J.D. (2020). Duckweed protein supports the growth and organ development of mice: a feeding study comparison to conventional casein protein. Journal of Food Science, in review.

Roman, B., and Brennan, R.A. (2020). Life cycle assessment of an Eco-MachineTM indicates beneficial environmental impacts for upcycling wastewater nutrients into protein-rich duckweed. Journal of Environmental Management, in review.