1 W ATER USE OF POTENTIAL BIOENERGY TALL GRASSES IN FLORIDA By ARKORN SOIKAEW A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 201 2
2 201 2 Arkorn Soikaew
3 T o my parents, for their endless support and encouragement
4 ACKNOWLEDGMENTS I would like to sincerely thank Dr. John Erickson, Chairman of my Supervisory Committee, for his guida nce, all time support, and patience throughout my progression toward the completion of my graduate study. His knowledge, philosophy and suggestions were ultimately invaluable for me and my future. Moreover, I would like to thank Dr. Lynn Sollenberger and D r. Jerry Bennett for their help and assistance, and for serving on my Supervisory Committee. I also would like to thank Andrew Schreffler and Jonathan Holland for their general assistance in the lab and answering my questions.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Current U nited S tates Bioethanol Production ................................ ......................... 12 Alternative Bioenergy Grass Crops for the Southeastern US ................................ 14 Gi ant Reed ( Arundo donax L. ) ................................ ................................ .......... 15 Elephantgrass ( Pennisetum purpureum Schumach.) ................................ ....... 16 Sugarcane and Energycane ( Saccharum spp.) ................................ ................ 17 Miscanthus x giganteus ................................ ................................ .................... 18 Switchgrass ( Panicum virgatum L.) ................................ ................................ .. 19 Sorg hum ( Sorghum bicolor L.) ................................ ................................ ......... 19 Bioenergy Crops and Water Consumption ................................ .............................. 22 Measuring Crop Evapotranspiration ................................ ................................ ........ 23 2 YIELD, WATER USE, AN D WATER USE EFFICIENCY OF TH REE PERENNIAL POTENTIAL BIOENERGY GRASS CROP S IN FLORIDA ................ 31 Background ................................ ................................ ................................ ............. 31 Materials and Methods ................................ ................................ ............................ 34 Experimental Site and Design ................................ ................................ .......... 34 Cultural Practices and Biomass Yi eld ................................ ............................... 34 Water Use and Water Use Efficiency ................................ ............................... 35 Leaf gas exchange ................................ ................................ ........................... 37 Data Analyses ................................ ................................ ................................ .. 38 Results ................................ ................................ ................................ .................... 38 Weather Data ................................ ................................ ................................ ... 38 Crop Morphology an d Biomass Yield ................................ ............................... 39 Seasonal Water Use and Water Use Efficiency ................................ ............... 39 Leaf Gas Exchange ................................ ................................ .......................... 40 Discussion ................................ ................................ ................................ .............. 40
6 3 PLANTING DATE AFFECT S DRY MATTER YIELD, TRANSPIRATION, AND WATER USE EFFICIENCY OF SW EET SORGHUM ( SORGHUM BICOLOR L.) GROWN IN FLORIDA ................................ ................................ ............................. 49 Background ................................ ................................ ................................ ............. 49 Materials and Methods ................................ ................................ ............................ 51 Experimental Site and Design ................................ ................................ .......... 51 Cultural Practices and Biomass Yield ................................ ............................... 52 Water Use and Water Use Efficiency ................................ ............................... 53 Data Analyses ................................ ................................ ................................ .. 54 Results ................................ ................................ ................................ .................... 55 Weather Data ................................ ................................ ................................ ... 55 Sweet Sorghum Biomass Yield ................................ ................................ ........ 55 Sweet Sorghum Water Use and WUE ................................ .............................. 56 Discussion ................................ ................................ ................................ .............. 56 4 SUMMARY AND CONCLUSI ONS ................................ ................................ .......... 64 LIST OF REFERENCES ................................ ................................ ............................... 65 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 73
7 LIST OF TABLES Ta ble page 1 2 Methods for measuring and es timating evapotranspiration in cropping systems ... 28 2 1 Air temperature, relat ive humidity, solar radiation rainfall, and irri gation during the growing season at Citra, FL ................................ ................................ .......... 44 2 2 Treatment means for dry matter yield, tiller density and diameter of peren nial grasses at harvest .. ................................ ................................ ............................ 44 2 3 Treatment means for seasonal crop transpira tion and water use efficiency of perennial grasses ................................ ................................ ............................... 45 3 1 Plantin g dates, harvest dates and number of days from plant date to harvest date for To pper 76 6 sweet sorghum ................................ ................................ .. 59 3 2 Relative humidity, solar radiation, air temperature, rainfall, and reference ev apotranspiration during each of the planting date growth inte rvals ................. 59 3 3 Main e ffect of year on sorghum population at harvest and water use efficiency of Topper 76 6 sweet sorghum ................................ ................................ .......... 60 3 4 Main e ffect of planting date on dry matter yield, sorghum population at harvest, and water use efficiency of Topper 76 6 sweet sorghum ................................ ... 60 3 5 Effect of year x planting date interaction on transpiration of Topper 76 6 sweet sorghum ................................ ................................ ................................ ............. 61
8 LIST OF FIGURES Figure page 1 1 R elations hip between stem diameter and plant area for corn ................................ 29 1 2 Comparison between transpiration by sap flow measurement compared with evapotranspiration by Bowen ratio method ................................ ........................ 30 2 1 Average daily relative humidity, solar radiation, air temperature, and total rainfall during at Citra, FL ................................ ................................ ................... 47 2 2 Perennial grass c rop transpi ra tion during the growing seasons ............................. 47 2 3 Perennial grass crop leaf gas exchange data of fully exte nded upper canopy leaves ................................ ................................ ................................ ................. 48 3 1 Average daily relative humidity, solar radiation, air temperature and total rainfall during at Citra, FL ................................ ................................ ................... 62 3 2 Average daily Topper 76 6 sweet sorg hum transpiration during the growing s easons as affected by planting date ................................ ................................ 63
9 LIST OF ABBREVIATION S CV. Cultivar E C C anopy t ranspiration E L L eaf transpiration E S Soil Evaporation ET Evapotranspiration ET C Crop evapotranspiration ET O Reference evapotranspiratio n K c Crop coefficient K canopy Crop transpiration coefficient PD Planting date Q Total solar irradiance T AIR Air temperature at 2 m WUE Water use efficiency expressed as dry matter produced per unit of water transpired
10 Abstract of Thesis Presented t o the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science WATER USE OF POTENTIAL BIOENERGY GRASSES IN FLORIDA By Arkorn Soikaew May 2012 Chair: John E. Erickson Major : Interdisciplinary Ecology Identification of suitable regional bioenergy crops in the US is currently underway in order to increase domestic energy production and satisfy growing demand for energy. While yield potential will be a critical component of an y potential biofuel crop, the effects of bioenergy crop production on regional water resources will also be important. The overall goal of these studies were therefore to gain a better understanding of yield, transpiration ( E C ) and water use efficiency (W UE) of potential bioenergy feedstocks in Florida. The first objective was to measure dry matter yield, E C and WUE of giant reed ( Arundo donax L. Pennisetum purpureum Schum.), and Saccharum spp.) during th e plant (2009) and first ratoon (2010) crop growing seasons at Citra, FL. Dry matter yields ranged from 35 40 Mg ha 1 yr 1 for energycane and elephantgrass, which were greater than 13 (plant) and 29 (first ratoon) Mg ha 1 yr 1 for giant reed. Total season al crop E C measured by sap flow sensors and scaled to the canopy level averaged 1150, 1090, 890 mm yr 1 for giant reed, energycane, and elephantgrass, respectively, across the two growing seasons. These yield and water use data resulted in elephantgrass and energycane having similar WUE, which was greater than giant reed. Taken together, results from this study
11 indicated that elephantgrass has the potential for comparatively high biomass yields coupled with relatively low E C compared to the other perennia l grasses evaluated in the study although E C was relatively high by all grasses, as they used the majority of rainfall and irrigation received during the two growing seasons. The second objective was to evaluate the effect of planting date on dry matter y ield, E C 76 Sorghum bicolor (L.) Moench ] for two growing seasons at Citra, FL. Total dry matter yield averaged 21.6 Mg ha 1 yr 1 for the late March and early May planting dates which was greater than the 17.2 Mg ha 1 yr 1 for the mid June planting date Total growing season E C measured by sap flow sensors and scaled to the canopy level was generally greater for the earliest planting date (772 mm) compared to the early May planting (608 mm), which was greater than th e mid June planting (435 mm), averaged across both growing seasons. These yield and E C data indicated that early May and mid June planting dates resulted in similar WUE, which wa s greater than planting in late March Thus results from this study indicated that water use by sorghum was comparable to other C4 grasses and that an early May planting date was optimal for sweet sorghum production when yield and water use were considered together.
12 CHAPTER 1 INTRODUCTION Current US Bioethanol Production I t is imperative that biofuel cropping systems are managed for multiple ecosystem services beyond crop production, especially in light of the diminished services that typically accompany intensive agricultural food production and the impact upon natural eco systems that ma y result from agricultural production (Carpenter et al., 1998). Agroecosystems dedicated to feedstock production for renewable energy are likely to expand rapidly in the coming years to meet federal mandates for renewable fuels (DOE, 2006). Under the Energy Independence and Security Act (EISA) of 2007, the Renewable Fuels Standard (RFS) has mandated 7.5 billion ga llons of biofuel production by 2012 and 36 billion ga llons by 2022 (Table1 1) (Sissine, 2007). Currently, the predominant renewable liquid fuel in the US is ethanol, which is derived almost entirely from corn grain. The inability of corn alone to meet federal mandates for renewable fuel production, the diversion of corn from feed and food to ethanol, and the adverse environmental impa cts of expanding corn production to marginal lands have led to an interest in other more sustainable biofuel cropping systems ( Farrell et al., 2006; Ragauskas et al., 2006; Cassman and Liska, 2007). Consequently, there is interest in other crops for the so utheastern US where the climate is favorable for growing highly productive perennial grasses and sorghum as biofuel crops. However, many of these crops have little production history in this region, particularly at the larger scales required for biofuel pr oduction. Thus, the impact of cropping systems on ecosystem services is not well understood and research optimizing services in addition to crop production is a critical need.
13 At present renewable energy contributes to around 8% of total US energy consu mption with the largest portion (~50%) credited toward biomass energy ( Center for S ustainable S ystems, 2010 ). While it is likely that a number of renewable energies (e.g. photovoltaic, wind, etc.) will be needed to satiate demand for energy, biofuels deriv ed from agroecosystem production are one of the most promising approaches to reduce US dependency on foreign energy sources in the short term and/or to meet the growing demand for energy in the future. Analyses have also shown that b ioenergy cropping syste ms can reduce environmental impacts ( e g ., greenhouse gas emissions) compare d to fossil fuel combust ion because only carbon derived from the atmosphere during the growing season was released as carbon dioxide back to the atmosphere ( Cherubini et al., 2009 ) T he overwhelming majority of liquid renewable fuel in the US is ethanol 10.7 billion ga llons were produced in 2009 ( Renewable Fuels Association, 2012 ) derived from corn grain. T he 2010 production of corn grain wa s estimated to be 12.447 billion bushels at an average yield of 152.8 bushels/acre (USDA NASS, 2012 ) Since one bushel of corn produces approximately 2.8 ga llons of ethanol, this results in about 428 ga llons ethanol per acre from corn grain. Thus, production of 11 billion ga llons of ethanol from corn require s approximately 25 million acres of corn production or about 1/3 of the US corn crop. Although this level production has begun to impact the demand traditional us e as a food and feed crop, along with the impacts of land conversion to grow corn for ethanol ( Searchinger et al., 2008 ). When tak ing into account land clearing and conversion and n itrogen fertilizer and pesticide s require d for corn production, these
14 facto rs significantly lower the net energy balance of corn derived biofuel (Pimentel, 2003 ) Moreover, high water consumption ( Evans and Cohen, 2009 ) also hinders the benefits of exploiting corn based ethanol in term s of sustainability. As a result, the EISA of 2007 has capped the production of ethanol from corn starch (conventional biofuel) at 15 billion ga llons starting from 2015, and thus the 21 of 36 billion ga llons of biofuel to be produced by 2022 must derive from cellulosic crops or non corn feedstock (a dvanced biofuel) ( Sissine, 2007 ). Thus, there are growing concerns related to the sustainability of using corn starch as biofuel feedstock ( Pimentel, 2003; Searchinger et al., 2008) and a strong interest in more productive and more sustainable crops for bi oenergy use Alternative Bioenergy Grass Crop s for the Southeastern US Select ed tall grass crops off er several advantages compared to other current and potential bioenergy crops First, tall grass crops establish quickly and provide at least an annual h arvest compared to woody crops. Many grass crops do not directly compete with food and feed Some of the t all grass crops tend to be drought tolerant or well adapted to regional water regimes and thus requir ing minimal irrigation water inputs (e.g., Lewan dowski, 2003; Knoll et al., 2011). They also may requi re relatively low fertilizer inputs and still be able to maintain high yield (Lewandowski et al., 2000), although if inputs are too low continued yield loss will occur (Knoll et al., 2011) M oreover, wh ile pest and disease problems are critical for food crop s necessitating crop protection inputs, many tall grass crops have relatively few pest and disease pests especially since many have limited if any reproductive structures to promote/support disease. Kort et al. (1998) also reported that perennial grass crops help in reducing soil erosion and sediment loss.
15 With a long growing season, high precipitation, favorable temperature s and marginal land availability, the southeastern US is now becoming an act ive region in perennial tallgrass development for biofuel resources. ( McKendry, 2002 a ; Prine et al 2000; and Woodard and Prine, 1993). In particular energycane ( Saccharum spp.), napiergrass or elephantgrass ( Pennisetum purpureum Schum. ), giant reed ( Aru ndo donax L. ) and sorghum ( Sorghum bicolor L.) have received much attention in the region as candidate bioenergy grasses ( Woodard and Prine, 1993; Lewandowski et al 2003; McKendry, 2002 b ; and Samson et al., 2005 ; Knoll et al., 2011; Fedenko et al., 2011 ) While these grasses are known to exhibit relatively high productivity in the region, little is yet known about their water use and potential impact on water resources in the region. Giant Reed ( Arundo donax L. ) Giant reed is a C3 tall grass species in t he Arundinoideae subfamily of the Poaceae (grass) family believed to be native to South East Asia and/or Mediterranean area. It is now being considered as a potential bioenergy crop due to its rapid biomass accumulation w ith moderate input requirement s dr ought tolerance, high heating value of harvestable component (up to 18.8 MJ kg 1 in dried stem Lewandowski et al. 2003), and the ability to thrive across a wide range of soil and climatic conditions. Giant reed is propagated vegetatively either by cuttin g the ma ture stem, rhizome regeneration or tissue culture since commercial seed is not available. Giant reed grass has hollow stem s, but can grow up to 9m and can yield up to 100 t ha 1 under optimal water condition s ( Lewandowski et al. 2003 ). It has an e xtensive root system that consist s of underground compact, dense rhizomes and extensive fibrous root s deep below the ground. However, a potential major problem of producing giant reed grass for bioenergy
16 is that it is a documented invasive weed species tha t can disrupt native plant species ( Davis et al. 2010 ), although it is not currently listed as an invasive plant in Florida Elephantgrass ( Pennisetum purpureum Schumach.) Elephantgrass is an erect C4 warm season perennial bunchgrass that produces large s tiff stems at maturity. Although dwarf and intermediate strains exist, tall growing selections produce the highest biomass yield when allowed to grow the entire warm growing season. From April through October, the coarse stems increase in length and weight accumulating dry matter up to 5 6 m above the ground. Spike (cattail) seed heads emerge from late October through November for most selections. Seeds are small, quickly lose viability, and are genetically diverse; therefore elephantgrass is propagated ve getatively. While the highest production has been reported on moderately drained, lowland soils, elephantgrass is drought tolerant and can produce high yields on excessively drained, deep sands (Woodard and Prine 1993 ) Though clump sections can be dug a nd used for propagation, it is primarily propagated with 3 to 4 node (bud) pieces cut from the hardened portion of the stalk (Woodard and Sollenberger, 2008) Pieces planted horizontally in furrows should be covered with no more than 5 to 7.5 cm of soil. Three node stalk pieces can also be placed into the soil at a 45 degree angle with the uppermost node exposed. At the northern limits of its range, stalk pieces can be planted in August or from late October until the first killing freeze (Woodard and Soll enberger, 2008) When planted in the summer, elephantgrass should be allowed to accumulate biomass without harvest throughout the remaining growing season to ensure winter survival. The planting time is less critical, though, at the warmer, more southern r egions of its range. In both cases, the seedbed should be thoroughly prepared and rows should
17 be 1 to 1.5 m apart. For large operations, machinery used to plant sugarcane stalk pieces can be adapted for elephantgrass. Both summer and fall plantings can be harvested for biomass during the time following growing season. Elephantgrass generally compares favorably in biomass yield when compared to other candidate bioenergy grasses for the Caribbean and US Gulf Coast regions ( Woodard and Prine, 1993 ). In the sou theastern US dry biomass yield for a full 1 ( Woodard and Prine, 1993; Knoll et al., 2011 ), depending on inputs and management practices. Elephantgrass can also be harvested twice yearly in the r egion. Annual yield will generally be 10 to 20% lower with two harvests, but the biomass composition may be more favorable for conversion to biofuels depending on the conversion process Sugarcane and E nergycane ( Saccharum spp.) Sugarcane and e nergycane ( S accharum spp.) are C4 tropical grasses in the Saccharum genus and consist of a number of species derived from hybridization through time. Sugarcane S. officinarum L., is believed to have originated in New Guinea (Barber, 1920) or in the region of common f rontiers of India, Burma, and China (Mukherjee, 1957). It has been cultivated for more than 4,000 years and has been distributed along human migration route s throughout many reg ions of the world In the US sugarcane has been primarily planted for producin g sugar and molasses while Brazil produces sugarcane for use in making both sugar and e thanol. Energycane is a group of Saccharum species hybrid s characterized by high biomass productivity and fiber production rather than sucrose accumulation in stalks (R ainbolt and Gilbert, 2008 ) Currently, sugarcane is being used to produce both sugar and lignocellu losic feedstocks and the bagasse can be also used to produce electricity to power sugar mills or ethanol
18 conversion facilities Sugarcane is now being develo ped and utilized as a biofuel feedsto c k, since its total above ground biomass can be converted to ethanol and energy. Generally, sugarcane has a relatively high seasonal water requirement (1500 2500 mm yr 1 ) and has only moderate drought tolerance. It is a lso quite susceptible to cold stress. In Florida, sugarca ne is widely grown commercially in and around the Ever glades Agricultural A rea for cane sugar production. In fact, Florida is currently the largest s ugarcane sugar producer in the US (Baucum and Rice 2009 ). Sugarcane cultivar CP89 2143 has been studied in Florida because it is one of the most productive in terms of harves table biomass and sugar content. It also possesses good tillering ability that underlies its high yield (reported up to 38 dry t ha 1 yr 1 ( Rainbolt and Gilbert, 2008 ) and it quickly develops a full canopy that helps to suppress weed growth and competition Energycane cultivar L79 1002 is the hybrid progeny of CP52 68 sugarcane and a female parent with Tainan (a c lone of Saccha rum spontaneum ) resulting from the cooperation between Louisiana State University, USDA ARS (Agricultural Research Service), and the American Sugarcane League, Inc. L79 1002 possesses yield potential of up to 60 dry t ha 1 yr 1 ( Bischoff et al. 2008 ), rap id growth and development, and good ratooning ability which has made this energycane cultivar a potential bio f uel crop in the southeastern US and a check for comparison with ne w and other potential crops. With expansion of bioenergy crops in the Southeast smut disease has become a problem for L79 1002, greatly reducing realized yields. Miscanthus x giganteus Giant miscanthus ( Miscanthus x giganteus ) is a C4 rhizomatous warm season grass in the family Poaceae and is native to Asia and Africa It has receiv ed much
19 attention as a potential bioenergy feedstock in the US. It can be grown as a loose or dense bunchgrass reaching heights of up to 4 m tall It is propagate d mainly via underground rhizomes. Miscanthus possesses high water use efficiency ( Clifton Bro wn and Lewandowski, 2000 ) and high yield potential up to 44 t DM ha 1 yr 1 (Lewandowski et al. 2000 ). Miscanthus can be grown in a wide range of soil types but has low tolerance to drought. It thus requires a c ertain amount of water (750 1200 mm yr 1 ) to maintain high yield. However, not much is known on growing miscanthus in Florida compared to other grasses in ter ms of production and management, but its high tolerance to cold weather (can be rejuvenated from dormant rhizome s in winter when the average te mperature is as low as 10 C) may represent an advantage in becoming a better alternative in higher latitude region s in Florida or in US. Switchgrass ( Panicum virgatum L.) Switchgrass is a C4 tall grass in the subfamily Panicoideae of the Poaceae family na tive to North America. It has been used primarily for animal pasture in the US It can reach a 3 m height and thri ve across a wide range of soil pH. Switchgrass is drought tolerant with a deep rooting system to maintain survival, allowing the crop to achie ve relatively high yield s even on marginal lands How ever, with a current yield potential of only arou nd 15 dry t ha 1 yr 1 (McLaughlin and Kszos, 2005), it is not competitive enough when compared to other tall grass candidates in the southeastern US. Sorg hum ( Sorghum bicolor L .) Sorghum is a highly productive C4 grass species that is well adapted to warm climates and tolerant of intermittent drought. Although, sorghum is technically a perennial crop in its native tropical environment, it is typically plant ed from seed and managed as an annual crop. Sorghum exhibits tremendous variability in growth
20 characteristics, which makes it a very attractive potential biofuel crop suitable for multiple biofuel conversion processes. Sorghums are commonly separated into different groups based on the amount of different carbohydrates they produce (Rooney et al., 2007; DOE, 2011) Grain sorghum hybrids produce large quantities of grain (~50% of total aboveground biomass) that are high in starch. Forage sorghum hybrids tend to produce higher biomass and relatively less grain, and have been optimized for ruminant nutrition (e.g., low lignin brown midrib mutants). Sweet sorghums are unique in that they accumulate high concentrations of soluble sugars in the stalk and relatively produce low grain yield. Finally, dedicated biomass sorghums have emerged recently that have been bred and selected for high lignocellulosic biomass production. Biomass sorghums produce low soluble sugars and because they are highly photoperiod sensitive they do not flower until late in the growing season, allowing for maximal accumulation of biomass and little if any grain production (DOE, 2011) Sorghum height, thus, ranges from less than 1 m for grain sorghum types to over 5 m tall for sweet and biom ass sorghums (Vermerris et al., 2011) The thickness of stalks also varies, ranging from 1 to 4 cm. Brace or prop roots often grow from the lower nodes. Juice content of stalks at maturity is typically lower in grain, forage, and biomass types compared to sweet types. Seeds are produced by self pollination from a panicle that emerges at the top of the plant and contains both the male and female inflorescences. Sorghum seeds are small, round and may be white, yellow, brown, or red in color. After harvesting the stalks, most varieties will regrow or ratoon. The ability to form a ratoon enables multiple harvests per season in certain environments although
21 yields typically decrease in ratoon crops (Erickson et al., 2011). Sugar concentration in the juice increas es with maturity, and is low prior to seed development. Sorghum is a warm season crop that tolerates drought and high temperatures better than many crops, but it does not grow well under low temperatures ( Newman et al., 2010) Optimal planting times will v ary among locations, but soil temperatures at planting should be above 18C. Late planted crops will mature more rapidly, but should be planted early enough to ensure that the crop matures before the first expected killing frost. Sweet sorghum can be produ ced in a wide variety of soil types, but yields are typically highest in deep, well drained soils with good fertility (Doggett, 1988) Sorghum grown in shallow soils or soils very low in organic matter may be more prone to drought stress. Although sorghum is more tolerant of drought stress than many other crops, ample moisture during the growing season is important for good yields of stalks and juice. Sweet sorghum is typically seeded in widely spaced rows (76 cm) using a corn planter or a vacuum planter at 6 to 8 cm within row spacing. If plant populations are too high, the canes will be spindly, contain less juice than an equal tonnage of larger diameter canes, and be more susceptible to lodging. Sweet sorghum can be harvested with a forage harvester or a sugarcane type harvester that cuts the stalks into billets, although recent studies have shown significant and rapid deterioration of sugars in forage harvested material compared to billets or whole stalks ( Lingle et al., 2012 ). Sweet sorghum biomass and carbohydrate yields are generally greater than many annual crops and have even out yielded perennial grass crops ( Hallam et al., 2001 ). Erickson et al. (2011) reported dry matter yields for sweet sorghum around 20 Mg ha 1
22 Yields of greater than 30 Mg ha 1 for biomass sorghum have been reported (Rooney et al., 2007), which are comparable to some of the best performing perennial biofuel grass crops in the Southeast (Knoll et al., 2011). Bioenergy Crops and W ater C onsumption Currently, the agricultural sect or contributes to around 70% of global water use by humans ( FAO, 2012 ), especially by irrigated cropping systems. The trend of agricultural water withdrawal is likely to continue to grow in the near future to supplement the rising demand resulting from a r apid ly growing population and this will contribute to the reduction in per capita domestic water availability (Berndes, 2002). In the s outheastern US however, rain fed cropping system s will be more preferable due to abundant rainfall, and thus would bec ome more sustainable and cost efficient than irrigated system s Still, the impacts of large scale production of bioenergy crops on water resources are not well understood. As biomass yield of crops increases an inevitable consequence is the loss of water through stomata via evapotranspiration (explained in next section). Crops that are able to maintain high yield s per unit water use are more preferable. This concept is known as water use efficiency (WUE) and is often defined as a measurement of the yield (either biological yield or economic yield) per unit of water use (either evapotranspiration or applied irrigation). However, there are several factors affecting results fr om the measurement of crop WUE, such as the variable quantity of the harvestable pa rt of the biomass (harvest index) and the crop growing season and climatic conditions during which WUE was measured ( Berndes, 2002). In general, C4 plant s tend to exhibit higher WUE than C3 plant s ( Sage, 2000). Stanhill (1986) c ompared the difference betwe en WUE of various C4 versus C3 species and found a
23 high er effici ency among 14 C4 plants with an average of 3.1 g kg 1 compared wi th 51 C3 plants species which average d 1.6 g kg 1 However, significant variation was also seen within both C3 and C4 species, sometimes by as much as 50% and WUE has also been shown to vary by genotype within species. Even a 10% difference in WUE could result in either higher yields or a savings of water of over 100 mm per year. In addition to WUE, water use rates in the field wi ll be influenced by water availability and crop phenology. Thus, selection of crop species has the potential to impact crop water use, WUE, and water resources in the region, yet little is known in this regard for the species (genotypes) under consideratio n. Measuring C rop E vapotranspiration In general, water is lost to the atmosphere by plant transpiration a process by which leaf water vapor is released to the atmosph ere via the stomatal openings and by evaporation primarily f rom the soil. The term evapot ranspiration (ET) is hence coined as the sum of leaf transpiration ( T L eaf ) and soil evaporation ( E s ). ET can be affected by several factors such as environmental conditions (e.g. soil characteristics, soil management, soil water con centration ground cove r or potential diseases), weather conditions (e.g. temperature, wind speed, and relative humidity), and crop characteristics (e.g. crop variety, crop height, canopy and root characteristics, and developmental stages). Early during vegetative growth and d evelopment, crops tend to have less leaf area resulting in the majority of ET being E s Following canopy closure E will become the dominant component of ET as soil evaporation is reduced due to more light interception by crop leaves are completely cover ing the soil surface
24 Evapotranspiration can either be measured directly or estimated from model s ( Rana and Keterji, 2000). ET measurement can be divided into sub categories based on purposes and objectives of those measurements (Table 2 1 ). W eighing lysi meters are often regarded as one of the most practical and precise methods to measure plant ET directly. Weighing lysimeters are soil filled tanks with planted crops in which ET is measured by means of mass balance (weighing lysimeters) or volume balance of water. However, lysimetry also has a number of potential limitations for determining ET under field condition s. First the difference between vegetation inside and outside vegetation (such as crop height and density) of lysimeters can result in overest imation or underestimation due to aerodynamic and radiative transfer as well as difference in microclimate s due to solar radiation and other environmental factors Second, lysimeters determine the result in just a small area and thus do not always represen t the ET of the whole field. Third, the metallic compartment rim of lysimeters can be heated and cause advection of sensible heat to the plant, and also be the potential cause, in the case of relatively tall rims to the crops, of modification of plant micr oclimates, such as wind and radiation. Moreover, root compaction due to limited space for root growth and surface cracks around the lysimeter border s (which will cause ET overestimation from extra bare soil evaporation and which is the site of water accumu lation after rainfall or irrigation) are also the possible source s of misleading ET measurement s Crop E can also be measured in situ by the heat balance method using sap flow sensors. Originally developed by Sakuratani (1981), the stem heat balance method is a thermic approach to measure sap flow in an intact plant stem that is widely used in the study of plant and water relation s In brief, a constant heat input from an external source
25 will be applied to the annular heater that encircles a plant stem. An improved version with a digital datalogger ( Baker and van Bavel 1987) allowed the prolonged measurement s of crop transpiration, making this method more suitable for field condition s According to Sakuratani (1981), t he energy balance of a heated stem can be defined as Q = Q v + Q r + Q f + S where Q represents the heat supplied, Q v is the conduction of heat upward and downward from the heating section, Q r is the radial transfer of heat perpendicular to the stem axis, Q f is heat energy transport v ia mass flow of water, and S is the rate of change in heat storage. By coupling these measurements with the stem area, we can essentially estimate the amount of water moving through the crop stem and thus crop E If we assume E s to be essentially negligib le following canopy closure we can obtain the evapotranspiration of a specific crop (ET C ) by multiplying reference evapotranspiration (ET O ) by crop coefficient factor ( K c ). ET O defined as the rate of evapotranspiration of a well watered hypothetical gras s surface of known characteristics (assumed crop height of 0.12 m, an albedo of 0.23, and fixed surface resistance of 70 s m 1 ) can be obtained via numerous public database accesses, such as University of Florida Institute of Food and Agricultural Science (UF IFAS) Florida Automated Weather Network (FAWN : http://fawn.ifas.ufl.edu/ ), United States Geological Survey (USGS) hydrological data web portal ( http://www.sflorida.er.usgs.gov/ ) and the National Climatic Data Center (NCDC) webpage ( http://www.ncdc.no aa.gov ) The empirically determined K c will account for the difference between an actual crop and a reference crop (typically
26 a short grass ), as it incorporates the indiv idual crop type characteristics, such as height, radiation absorbed rate, canopy resis tance, and eva poration from bare soil surface In order to obtain ET C data from single plant transpiration measured via sap flow, measurements of plant density and/or plant stem area are required to scale up to daily and seasonal crop ET C This can be comp leted by analyzing either plant leaf area spatial variability or plant stem diameter variability ( Bethenod et al., 2000 ). These authors showed that scaling up from stem to stand ET C using stem diameter to choose to be installed plant with sap flow sensor e xhibited a strong relationship (R 2 = 0.77) between stem diameter and plant area (Figure 1 1 ). The strong correlation indicated that the use of stem diameter was a suitable way for scaling to ET C In order to appropriately implement the scaling up informatio n, it is important to select representative sampling units (Bethenod et al., 2000). Thus, when properly implemented, t he sap flow method has shown a strong agreement to ET C data determined from the Bowen Ratio (ratio of soil heat flux to sensible heat flux ) as illustrated in Figure 1 2 (Bethenod et al., 2000).
27 Table 1 1 Summary of the R enewable F uel S tandard set in E nergy I ndependence and S ecurity A ct of 2007 Year Conventional Biofuel (billion ga ) Advanced Biofuel (billion ga ) Cellulosic Biofuel (billio n ga ) Total RFS (billion ga ) 2008 9.00 9 2009 10.50 0.60 11.10 2010 12.00 0.95 0.10 12.95 2011 12.60 1.35 0.25 13.95 2012 13.20 2.00 0.50 15.20 2013 13.80 2.75 1.00 16.55 2014 14.40 3.75 1.75 18.15 2015 15.00 5.50 3.00 20.50 2016 15.00 7.25 4.25 22.25 2017 15.00 9.00 5.50 24.00 2018 15.00 11.00 7.00 26.00 2019 15.00 13.00 8.50 28.00 2020 15.00 15.00 10.50 30.00 2021 15.00 18.00 13.50 33.00 2022 15.00 21.00 16.00 36.00
28 Table 1 2 Methods for measuring and estimating evapotranspirati on (ET) in cropping systems (adapted from Rana and Katerji, 2000). ET measurement Hydrological a pproaches Micrometeorological a pproaches Plant physiology approaches Soil water balance Energy balance and Bowen ratio Sap flow method Weighing lysimeters A erodynamic method Chamber system s Eddy covariance ET estimation Analytical a pproaches Empirical a pproaches Penman Monteith model Methods based on crop coefficient approach Methods based on soil water balance modeling
29 Figure 1 1 Correlati on between stem diameter and plant area for corn in order to identify the selection of plant s to equip with sap flow sensor s ( reprinted from Bethenod et al. 2000) Note: Y = 0.5921 X 3 + 18.70 X 2 + 74.63 X (r = 0.877, n = 35).
30 Figure 1 2 Comp arison between transpiration by sap flow measurement and evapotranspiration by the Bowen ratio method for both daily (left, r 2 =0.90, n=45) and hourly (right, r 2 =0.88, n=456) measurements ( reprinted from Bethenod et al. 2000 )
31 CHAPTER 2 YIELD, WATER USE, AND WATER USE EFFICIENCY OF TH REE PERENNIAL POTENTIAL BIOENERGY GRASS CROPS IN FLORI DA Background Agroecosystems currently represent over 920 million acres of land (> 40% of land area) in the US These agroecosystems have been managed for years to provide food, fuel and fiber. Demand for these services especially to produce renewable liquid transportation fuels has increased in recent years, coinciding with federal mandates to have 30% of all fuel consumed in 2030 to be renewable (DOE, 2006 2011 ). Altho ugh the US produced about 13 billion ga llons of renewable ethanol in 2010 ( Renewable Fuels Association, 2012 ) it was derived from almost 1/3 of the m idwest ern corn grain crop. This has led to an interest in producing energy from other more sustainable cro pping systems like perennial grasses that do not directly compete with our food systems (Knoll et al., 2011). While many perennial tall grass crops have the potential for high productivity (Woodard and Prine, 199 3 ) these high yielding bioenergy cropping s ystems have the potential to greatly diminish less visible ecosystem services (Hill et al., 2006) especially hydrological regulation at the expense of crop primary production. While p roduction of crops with readily fermentable sugars (i.e. starch, sucro se, and simple sugars) can immediately increase renewable liquid fuel production and lignocellulosic bioenergy cropping systems will clearly be needed to meet renewable fuel mandates ( DOE, 2006, 2011 ). In addition to woody biomass, highly productive perenn ial grasses such as energycane ( Saccharum spp.) e lephantgrass ( Pennisetum purpureum Schum.) giant reed ( Arundo donax L.), Miscanthus x giganteus reed canarygrass ( Phalaris arundinacea L.) and switchgrass ( Panicum virgatum L.) have
32 been evaluated as pot ential lignocellulosic bioenergy crops ( Woodard and Prine, 1993; Angelini et al., 2005 ; Cassman et al., 2006 ; Christian et al., 2008; Knoll et al. 2011 ; Kering et al., 2012 ). In the southeastern US, elephantgrass and energycane, both warm season grasses, h ave generally been among the most productive perennial grass crops, producing dry matter yields of 20 40 Mg ha 1 (Woodard and Prine, 1993; Knoll et al., 2011; Fedenko, 2011). While the C4 grasses like elephantgrass and energycane are generally efficient in terms of water and nutrient use ( S tanhill 1986 ; Beale et al., 1999 ) many have poor cold tolerance and do not perform optimally at higher latitudes in the southeastern US ( Berry and Bjorkman, 1980 ). Therefore, productive C3 grasses like giant reed (Burval l, 1997; Angelini et al., 2005) have also been evaluated for bioenergy use in the region (Knoll et al., 2011). Dry matter yields of about 30 Mg ha 1 have been reported in temperate climates for giant reed (Angelini et al., 2005). Under certain environmenta l conditions in the southeastern US, dry matter yields of giant re ed have been comparable to elephantgrass and energycane (Fedenko, 2011), but giant reed yields are generally less (Knoll et al., 2011). Taken as a whole, limited data on biomass production have indicate d that energycane, e lephantgrass and giant reed are among the most promising perennial grass row crops for combustion and lignocellulosic conversion to ethanol in the southeastern US However, the impact of these cropping systems on water reso urces is not well understood. Agroecosystems in general have become a major global consumer of water. In fact, agriculture consumes about 70% of the water used by humans globally, and irrigated croplands (approx. billion ha globally) in particular consum e significantly more water than the ecosystems they replace ( FAO, 2012) R ain fed
33 cropping systems are more sustainable than irrigated systems, although they might consume more or less water than the plant communities they replace. Despite abundant annual rainfall, the implications of bioenergy cropping systems for the southeastern US remains a big concern (Evans and Cohen, 2009 ), especially since bioenergy cropping systems are likely to displace relatively low input (i.e., no irrigation) pasture lands Thu s, there is an ever increasing need to achieve greater crop production with less water use (Marris 2008). This is important for all crops, but is especially needed for bioenergy crops to allow for production on marginal lands and to minimize competition w ith food crops. However, an intrinsic property of plant photosynthesis is that water is lost from the plant through stomata to the atmosphere as carbon dioxide is taken up from the atmosphere and assimilated by the plant to be used for biomass synthesis ( W ong et al., 1979 ) Nevertheless, there is substantial variation in water use efficiency ( WUE; g biomass produced kg 1 of water transpired ) both within and across crops ( Jrgensen and Schelde, 2001 ). In particular crops that possess the C4 photosynthetic p athway tend to have a higher water use efficiency ( WUE ) compared to C3 crops. In a comprehensive review, Stanhill (1986) reported a mean WUE of 1.6 g kg 1 for 51 C3 plants and a mean WUE of 3.1 g kg 1 for 14 C4 plants, however values have been reported as high as 9 g kg 1 for the C4 grass m iscanthus ( Beale et al., 1999 ). Despite the potential implications of bioenergy crop production on transpiration and water resources, few data are available on transpiration of bioenergy crops grown in the field Therefo re, the objective of this study was to compare dry matter yield, water use and WUE of elephantgrass, energycane, and giant reed, grown under field conditions in North Central Florida. We hypothesized that perennial grass species would
34 differ in total tran spiration and temporal patterns of water use which could help lead to improved selection and use of perennial grass crops for bioenergy production in the region. Materials and Methods Experimental Site and Design As part of a larger experiment to identify potential tall grass bioenergy crops well adapted to the southeastern US, a replicated field experiment was conducted in North Central Florida at the University of Florida Plant Science Research and Education Unit on a very deep, excessively drained fine Candler sand ( h yperthermic, uncoated Lamellic Quartzipsamments). The previous crop was bahiagrass ( Paspalum notatum Flugge) followed by winter fallow. Weather data including reference ET (ET O ) were collected from the Florida Auto mated Weather Network (FAWN) weather station located less than 0.5 km N of the field site (FAWN, 2012) The experimental design was a randomized complete block design with four replicates. The main treatment factor was species and included energycane (cv. and giant reed (wild Florida population) Cultural Practices and Biomass Yield In November 2008, plots were established from stem cuttings. Each plot was 6 rows of 6 m length with plant spacing in the row of ~ 0.5 m for all species. Plots were fertilized with 280 kg N ha 1 yr 1 using a 16 4 8 blended granular fertilizer that included minor nutrients in split applications that supplied 90 kg N ha 1 yr 1 in mid April and 190 kg N ha 1 yr 1 in June. Known quantitie s of irrigation (Table 2 1) were applied to plots during establishment (2009) via overhead irrigation with a linear move system but thereafter only at sign of early visual drought stress (e.g., leaf rolling). Weeds were
35 mechanically removed during establ ishment by rotary hoe and subsequently by hand as needed. To estimate biomass yields in 2009 and 2010, plots were harvested once per year in the fall around late November, prior to anticipated frost. A 4 m section (4 m 2 ) from the middle of one of the two inner rows was cut at a 7.5 cm stubble height using a gasoline powered trimmer (Echo, Inc., Lake Zurich, IL) and harvested by hand. The 4 m section was immediately weighed green in the field to provide estimates of green yield. The total number of stalks from the 4 m harvested section was counted and used to determine tiller population at harvest. Additionally a four stalk whole plant subsample was collected, weighed fresh in the field and then dried at 50C until a constant dry weight was achieved to de termine dry matter concentration at harvest and dry biomass yield. The remaining biomass in each plot was mechanically harvested with a forage harvester Water Use and Water Use Efficiency Whole plant c rop w ater use (i.e., transpiration) was measured throu ghout the growing season on each of the three species using s ap flow sensors (Dynamax, Houston, TX) installed in situ on selected intact plant stems ( Sakuratani, 19 81; Sakuratani, 1987; and Baker and van Bavel, 1987) Although time and labor intensive, thi s heat balance method measures water use under actual field conditions without altering the microclimate (e.g., chamber methods) or the soil profile (e.g., lysimeter methods), and has been used to accurately measure crop water use for sorghum (Dugas et al. 1998), corn (Bethenod et al., 2000), sugarcane ( Saliendra and Meinzer, 1992 ) and other herbaceous crop species like soybean and cotton (Tan and Buttery, 199 5 ; Dugas et al. 1990; Dugas et al., 199 4 ).
36 Approximately every 2 to 3 wk three to four representa tive tillers from each species were selected from the inner 2 m 2 of the plot for sensor installation (maximum of 16 sensors) Prior to installation, any leaf sheath tissue was removed and stem diameter was measured in two directions (N S and E W) at sensor height (equidistant between two internodes) and averaged to estimate stem sap flow area Before placing sensors on the stem, the area was sp rayed with canola oil to maximize sensor contact with the stem. The foam insulated sensor was then placed on the st em and wrapped with an aluminum covered bubble wrap to shield the sensor and stem from solar radiation. Finally, a conical shaped plastic piece was wrapped around the stem above the sensor unit to prevent irrigation or rain water from moving downward along the stem toward the sensor. Once installed, all s ensors were left on the stem for 5 to 7 days A 12 volt deep cycle battery, a CR1000 data logger (Campbell Scientific, Logan, UT), and Dynamax software program ( Dynagage flow 32 1k ver 220.127.116.11 ) were used t o heat the sensors and to monitor thermocouple temperatures from each of the sensors. The temperature data were recorded at 15 s intervals and averaged every 15 min and stored by the datalogger. Using the input stem area and measured temperature data, the software program directly calculated average tiller water use in g h 1 over the 15 min interval and this was also stored by the datalogger for each tiller. M easurements were repeated approx imately every 3 wk during the growing season until harvest in mid N ovember. Measured w hole plant sap flow data were then scaled to estimate d aily crop canopy tr anspiration ( E C ) per unit ground area for each of the plots Thus, E C was calculated based on the product of the mean measured sap flow (g hr 1 cm 2 stem area)
37 a verage stem area (cm 2 ), and tiller density (no. per m 2 ) per plot Stem diameter data were collected monthly on 40 randoml y (every 4 th tiller in the inner plot where water use was measured) selected tillers at sensor height to obtain the mean tiller diamete r for each plot. The number of tillers in the inner 2 middle rows of the plot (8 m 2 ) was also counted monthly to determine tiller density For each day where sap flow was measured, c rop canopy transpiration coefficients ( K c anopy ) were calculated as the quo tient of E C and ET O from the nearby FAWN Netw ork For days where sensor sap flow was not measured daily E C was estimated as the product of ET O and K c anopy where v alues of K c anopy were linearly interpolated for each day across the measured values, and fro m K c anopy = 0 at crop emergence to the K c anopy value approximately 3 weeks after emergence each year. Total seasonal E C was then estimated as the sum of daily E C from emergence to harvest. Water use efficiency for each species was calculated from the quoti ent of harvested dry biomass and total seasonal E C Leaf gas exchange In the central two rows of each plot, three fully expanded leaves were chosen at random for leaf gas exchange measurements during mid June in each of the 2009 and 2010 growing seasons Light saturated net CO 2 exchange ( A sat mol m 2 s 1 ), stomatal conductance ( g s mol m 2 s 1 ), intercellular CO 2 concentration ( C i mol mol 1 ), and transpiration efficiency (TE, mol CO 2 mol 1 water) were measured on 6 cm 2 leaf area using a LI 6400XT port able open flow photosynthesis systems (LI COR Inc., Lincoln, NE) Measurements were made between 1100 and 1300 h under cloud free conditions at 2000 mol m 2 s 1 photosynthetic photon flux density. Reference CO 2 concentration was set at 400 mol CO 2 mol 1 air and flow rate at 500 mol s 1 Temperature was
38 maintained at 28 C and relative humidity between 55 and 65%, similar to environmental conditions in the field when measured. Data within species did not differ statistically across years and were thus pool ed and presented as means across both years. Data Analyses Statistical analyses on species effects within year were performed using analysis of variance procedures in the GLIMMIX procedure of SAS (SAS Institute, 2009). Species was treated as a fixed effect and block was treated as a random effect in the model. R esiduals from each model fit were analyzed for homogeneity of variance visually and for normality visually and with the Shapiro Wilk W test. Degrees of freedom were determined using the Kenward Roger method. Where significant ( P < 0.05) fixed effects were seen, treatment mean pairwise comparisons were made using the LSMEANS statement with the TUKEY method. Results Weather Data Average daily air temperature ( T AIR ) and solar radiation level ( Q ) were gre ater and RH lower during the 2009 growing season (April to November) compared to the 2010 season (Table 2 1 Figure 2 1 ). These differences were associated with reduced rainfall seen d uring the 2010 growing season especially during the summer months that are typically wetter in the region Radiation was about 50% higher during the peak summer months compared to the s pring and f all months. Since 2009 was an establishing year, irrigation inputs were greater during 2009 compared to 2010 even though rainfall w as also more abundant in 2009. In total, water inputs were 1327 and 990 mm for the 2009 and 2010 growing seasons, respectively
39 Crop Morphology and Biomass Yield Elephantgrass and energycane were quick to establish and produced relatively high dry matter y ields during the 2009 growing season compared to giant reed which was slower to establish (Table 2 2 ). No difference in biomass yield was seen between elephantgrass or energycane during either the 2009 or 2010 growing seasons. However, both energycane and elephantgrass yields were greater than giant reed, but they were only about 37 % greater in 2010 compared to 17 5% in 2009, which was the first growing season following establishment. During 2009, tiller densities were greatest in energycane, intermediate i n elephantgrass and least in giant reed (Table 2 2 ). During 2010, however, no difference in tiller density at harvest was seen among any of the species. Tiller diameter w as greater in elephantgrass compared to giant reed and energycane was intermediate b etween the two (Table 2 2 ). Seasonal Water Use and Water Use Efficiency Total seasonal water use E C during the first year following planting (2009) was greatest for energycane lowest for elephantgrass, and intermediate for giant reed, which did not diff er from either of the other two grasses (Table 2 3). During 2010, E C was greatest for giant reed, intermediate for energycane, and lowest for elephantgrass (Table 2 3). Seasonal patterns in daily E C (Figure 2 1) indicated earlier emergence during 2010, the second full growing season, especially for giant reed. Thus E C of giant reed tended to be relatively greater earl y in the growing season, while energy cane tended to be relatively greater late in the growing season. Water use efficiency was greater for el ephantgrass and energycane, which did not differ, compared to giant reed in both the 2009 and 2010 growing seasons (Table 2 3) In 2010, WUE was generally greater for all species compared to the 2009 growing season.
40 Leaf Gas Exchange Light saturated leaf net carbon assimilation A sat w as greatest for energycane, approaching 50 mol m 2 s 1 and lowest for giant reed, approximately 30 mol m 2 s 1 ( Figure 2 2A). Leaf stomatal conductance g s was greatest for giant reed, intermediate for energycane, and l owest for elephantgrass (Figure 2 2B). This resulted in elephantgrass possessing the greatest leaf TE followed by energycane, and then giant reed (Figure 2 2 C ). Finally, leaf C i was lowest in elephantgrass, intermediate in energycane, and highest in giant reed (Figure 2 2D). Discussion A number of recent studies have raised concerns over water resources with regard to production of biofuel cropping systems (Berndes, 200 2 ; Evans and Cohen, 2009). In the present study, E C was relatively high, especially for giant reed, the C3 grass. Daily E C of approximately 8 to 10 mm d 1 was not uncommon following precipitation events. Although precipitation was historically low during both growing seasons, E C was well above rainfall received. Scrutiny of tradeoffs in E C wi th biofuel crop production is needed, but must be placed in the context of current land use patterns. In the southeastern US this includes native pine flatwoods ecosystems, managed pine plantations, managed pastureland, and potentially citrus groves. Annu al ET of a natural pine flatwoods in Florida was 812 mm ( Bracho et al., 2008 ), managed slash pine systems ranged from 951 to 1110 mm ( Gholz and Clark, 2002 ), a pasture system was 787 mm ( Sumner and Jacobs, 2005 ), and a citrus grove was 920 mm ( Fares and Al va, 2000 ). For the most part, ET by these ecosystems ranged from 70 to 90 % of annual precipitation (see summary in Bracho et al., 2008). Thus, seasonal E C of 850 1150 mm by the perennial grass crops in the present study are not that far out of line with cu rrent
41 land uses, but were on the high side, especially given the relatively low rainfall received during the study. Daily E C by the grasses was greatest early in the growing season during peak vegetative growth coinciding with high radiation and high evapo rative demand. Daily water use rates of corn grown near Ames, IA, were also greatest early in the growing season during growth stages R1 and R2 (Wiggans et al., 2012). During this period, daily E C of 5 to 7 mm d 1 was reported (Wiggans et al., 2012), which w as comparable to the daily E C for the C4 grasses in the present study. However, the C4 perennial grasses in the current study remain vegetative for much longer periods, resulting in relatively high total seasonal E C High transpiration rates in the prese nt study were comparable to those reported for corn (Jara et al., 1998), which averaged about 120 mm mo 1 and pearl millet (Azam Al ia et al., 19 84), which ranged from 77 to 100 mm mo 1 depending on row spacing. Thus, whereas a corn crop can be produced on about 500 mm of water, the C4 perennial grasses in the present study used between 850 to 1100 mm of water. Duration of growing season and management practices are therefore important for E C and ET as evidenced by total annual ET of intensively managed bahi agrass and St. Augustinegrass grown in South Florida, which ranged from 1200 to 1500 mm (McGroary et al., 2011 ). Data from the present study do not include evaporation, and actual crop ET would be expected to be even higher than the numbers reported here, but soil evaporation from tall grass crops with a long growing season is generally minimal. For example, E C was estimated to contribute to over 90% of ET of a fully matured crop (Allen et al.
42 1998). Thus, the numbers presented are likely to be generally r epresentative of actual ET. Actual crop ET depends on climatic factors, crop type and crop growth stage. While ET O provides the climatic influence on crop water use, the effect of crop type and management is addressed by crop ET. Factors affecting crop ET such as ground cover, canopy properties and aerodynamic resistance for a crop are different from the factors affecting the reference crop (grass or alfalfa); therefore, crop ET differs from ET O These factors that distinguish field crops from the reference crop are commonly integrated into a cr op factor or crop coefficient (Allen et al., 1998), which is then used to determine the actual E C for any crop as was done in the present study using K canopy Measured v alues of K c anopy in the present study were commo nly in excess of 1.0 and generally closer to 2.0 during mid growing season (data not shown). High K c anopy values close to 2.0 have been reported for other crops, including sugarcane (Chabot et al., 2005) and melon (Lie et al 2003). These high values coul d be due to differences in canopy aerodynamic resistance, which could have been exacerbated by relatively small plot sizes. However it has been suggested that smaller buffer and fetch areas are typically needed for humid and sub humid conditions (Fougerou z e, 1966) In general, giant reed transpired more water than elephantgrass and energycane during both growing seasons, which was expected given that elephantgrass and energycane are C4 grasses and giant reed is a C3 grass. Additionally, however, elephantgr ass E C was generally lower than that of energycane. These daily and seasonal findings were supported by leaf level gas exchange data, which indicated that g s was greatest in giant reed, intermediate in energycane and lowest in elephantgrass.
43 The differenc e in E C between giant reed and the C4 grasses was not as great as might be expected, and in fact did not differ from energycane during the plant crop growing season. While differences in seasonal E C were modest between the C3 and C4 grasses, differences in biomass yield between giant reed and the two C 4 grasses were large as dry biomass yield was substantially greater for the C4 grasses during both growing seasons, especially during 2009 (Table 2 2). Energycane and elephantgrass did not differ in dry matte r yields during either season and were consistent with reported values for biomass yield for these species (Sollenberger and Woodard, 2008; Knoll et al., 2011; Woodard and Prine 1993) Given comparable or lower seasonal E C and greater biomass yields, WUE was greater for the C4 grasses compared to giant reed the C3 grass. It has long been known that C4 grasses use water more efficiently to produce dry matter (see review in Stanhill, 1986), which was confirmed by this study. Values of WUE ranging from 3.3 t o 4.6 g kg 1 for the perennial C4 grasses in the present study were comparable to median value s of 4.2 reported for grain sorghum, 3.9 for pearl millet, and 4.8 g kg 1 for maize ( Connor et al., 2011 ). While C4 grasses tend to use water efficiently to produ ce biomass, this efficiency should not necessarily be construed as low total seasonal E C as discussed above. The implications are that approx imately twice the land area or twice the water would be needed for giant reed to produce yields competitive with t he C4 grasses. Therefore, in areas where consumptive water use is a concern, our results indicated that elephantgrass would be best suited for bioenergy crop production, but in regions such as frequently flooded marginal lands giant reed might achieve high er yields and make a suitable bioenergy crop.
44 Table 2 1 Average daily air temperature ( T AIR ) at 2 m, relative humidity (RH), solar radiation (Q), total rainfall, and total irrigation for the 2009 and 2010 growing seasons (Apr il November) at Citra, FL Year T AIR RH Q Rainfall Irrigation C % W m 2 mm mm 2009 23.6 79.1 195 897 430 2010 24.4 77.0 213 585 405 Table 2 2 Treatment means for dry matter yield, tiller d ensity and diameter data of perennial grasses collected at harvest during the 2009 and 2010 growing seasons Species include giant reed ( GR ), energycane ( EC ), and elephantgrass ( EG ). Species Dry Matter Yield Tiller Density Tiller Diameter 2009 2010 2009 2010 2009 2010 ------Mg ha 1 -------------per m 2 ----------------mm --------G R 13.4B 29.1B 14.5C 24.9A 13.5B 14.8B EC 38.4A 40.9A 25.3A 26.7A 15.0AB 15.0AB EG 35.5A 38.9A 19.3B 20.6A 17.1A 16.3A s e 2.61 2.75 1.55 2.23 0.73 0.58 Numbers followed by the same letter within a column do not differ ( P > 0.05) ; Standard error of differences of species means
45 Table 2 3 Treatment means for seasonal crop transpiration ( E C ) and water use efficiency (WUE) of perennial grasses during the 2009 and 2010 growing seasons. Species include giant reed ( GR ), energycane ( EC ), and elephantgras s ( EG ). Species E C WUE 2009 2010 2009 2010 --------mm ----------------g kg 1 --------GR 11 13 A B 1 177 A 1.1 9 B 2. 47 B EC 1151A 1035B 3.35A 3.96 A EG 930B 856 C 3.84A 4.57 A s.e. 77.1 33.3 0.24 0.32 Numbers followed by the same letter within a colum n do not differ ( P > 0.05) ; Standard error of differences of species means
46 Figure 2 1. Average daily relative humidity (RH), solar radiation ( Q ), air temperature at 2 m ( T AIR ), and total rainfall for 2009 (A) and 2010 (B) at Citra, FL.
47 Fig ure 2 2 Daily c rop transpiration ( E C ) during the 2009 (A) and 2010 (B) growing seasons at Citra, FL Species include giant reed ( GR ), energycane ( EC ), and elephantgrass ( EG ).
48 Figure 2 3 A) Light saturated net carbon exchange ( A sat ); B) stomatal conductance ( g s ); C) transpiration e fficiency (TE); and D) intercellular CO 2 concentration ( C i ) of fully extended upper canopy leaves averaged across the 2009 (n=3) and 2010 (n=3) growing seasons. Species include giant reed ( GR ), energycane ( EC ), and elep hantgrass ( EG ).
49 CHAPTER 3 PLANTING DATE AFFECTS DRY MATTER YIELD, TRANSPIRATION AND WATER USE EFFICIENCY OF SWEET SORGHUM ( SORGHUM BICOLOR L.) GROWN IN FLORIDA Background Sorghum is a unique C4 grass in that it is capable of high biomass, grain, and s ugar yields ( Rooney et al., 2007; Zhao et al., 2009). Thus it has received considerable attention in recent years for use as a biofuel crop (Amaducci et al., 2004; Goff et al., 2010) Sweet sorghum, in particular, offers p otential advantages over other ca ndidate biofuel crops because it stores readily fermentable sugars in the stalk that can be converted to liquid fuels and bio based products using current technologies. It is also an annual crop that can be integrated into existing crop rotations, and it i s readily established from seed. Moreover, it does not have to compete directly with food or feed. Thus numerous recent studies have demonstrated the ability of sweet sorghum to be used as a biofuel crop in the southeastern US, with estimated potential eth anol yields equivalent to or greater than corn ethanol yields from agroecosystems in the Midwest ( Erickson et al., 2011 ). Although these studies have demonstrated that sweet sorghum has good productivity and is well adapted to the Southeast relatively lit tle is known about the potential implications of sweet sorghum bioenergy cropping systems for other ecosystem services, especially hydrological cycling Thus, there is a need for a better understan ding of sweet sorghum water use in the Southeast especiall y with regard to management effects on water use and water productivity Sweet sorghum biomass yields have been variable within and across studies due to cultivar, environment and management practices. Wortmann et al. (2010) reported dry stalk yields fro m 8 to 48 Mg ha 1 across a range of N fertilization rates, three
50 cultivars, and multiple plant populations in Nebraska. Soileau and Bradford (1985) reported dry biomass yields ranging from 6 to 18 Mg ha 1 across a range of fertilization and liming treatmen ts in northern Alabama. Similarly Tamang et al. (2011 ) reported total dry matter yields of 9 to 18 Mg ha 1 81E sweet sorghum cultivars across a range of N fertilization rates in the Southern High Plains Miller and Ottman (2010) r eported no effect of irrigation frequencies on M 81E sweet sorghum dry matter yields, which ranged from 20 to 31 Mg ha 1 in Arizona Although sweet sorghum biomass yield data are abundant for some growing regions, there is limited information available on water use of sorghum in general, but especially of sweet sorghum Sweet sorghum has been shown to use less water and N than maize for similar ethanol yields (Keeney and DeLuca, 1992; Geng et al., 1989 ) Sorghum has also been shown to use water effi ciently Mastr orilli et al. (1999) reported water use efficiency ( WUE ) of up to 6 g kg 1 when sorghum was grown under water stress at differing phe nological stages in Italy. Additionally, St e d u to et al. (1997) concluded that sweet sorghum exhibited very high canopy WUE in s outhern Europe due to the small proportion of carbon loss through dark respiration. W hile limited data indicate that sorghum may be relatively efficient with water use, crop management practices such as cultivar, sowing date plant population, contro lling pests and disease, and avoiding periods of high evaporative demand can impact water use and WUE ( Connor et al., 2011 ) Planting date in particular could be possibly manipulated in the Southeast to reduce water use and optimize WUE by avoiding crop pr oduction during times of the year when evaporative demand is especially high For example, by sowing sunflower in Dec ember compared to Mar ch at Cordoba, Spain,
51 average daily ET was reduced during maximum vegetative growth and overall WUE was significantly increased, although total seasonal ET was higher due to the longer growing season ( Gimeno et al., 1989 ). Compared to maize for example, the use of sweet sorghum as a bioenergy feedstock will require a greater seasonal distribution of feedstock availabilit y, since juice cannot be as readily stored as grain, thus making cultivar and planting date important management tools. The use of both early and late maturing cultivars along with multiple planting dates can increase the length of time when stalks are ava ilable to be milled (Broadhead, 1969), but the consequences for yield and water use are not well known. The objectives of this research were therefore to evaluate the eff ects of planting date on total dry matter yield, water use and water use efficiency of a commercially ava ilable sweet sorghum cultivar grown on a well drained sandy soil in North Central Florida. Given the seasonal patterns in evaporative demand, t he hypothesis was that later planting date s would decrease water use and increase water use efficiency for sweet sorghum production in the region. Materials and Methods Experimental Site and Design A replicated field experiment was conducted at the University of Florida (UF ) Plant in 2009 and 2010. The study was conducted on a very deep, excessively drained fine sand of the Candler series ( h yperthermic, uncoated Lamellic Quartzipsamments). The previous crop was bahiagrass ( Paspalum notatum Flugge) followed by winter fallow. Weather dat a were collected from Florida A utomated Weather Network weather station and averaged [ air temperature at 2 m ( T AIR ), relative humidity (RH) and radiation ( Q )] or totaled monthly
52 (rainfall) (Figure 3 1). The experimental arrangement was a randomized complet e block design with four replicates. The treatment s consisted o f three planting dates [ l ate M arch (PD1), e arly May (PD2) and m id June (PD3)], which were randomly assigned. Cultural Practices and Biomass Yield Following cultivation to prepare a seed bed Topper 76 6 sweet sorghum seed (source: Mississippi Agricultural and Forestry Experiment Station ( MAFES ) Foundation Seed Stocks, MS State, MS) was planted in 0.76 m rows having a within row spacing of 6 to 8 cm and planted at approximately a 2.5 cm depth Topper 76 6 is a mid season sweet sorghum adapted for the southeastern USA with low lodging potential (Day et al., 1995). Each plot consisted of six rows that were approx. 7 m long. The plots were sown and harvested according to the schedule presented Tab le 3 1. Harvest date was based on soft dough stage, which is optimal for sugar yields (Tarpley et al., 1994). In total, all plots were fertilized with 135, 39, and 48 kg ha 1 yr 1 of N, P, and K, respectively. Liquid fertilizer (11 37 0) was applied at pl anting along with Counter, a systemic insecticide nematicide (Terbufos: S[[(1,1 dimethylethyl)thio]methyl]0,0 diethyl phosphorodithioate) at 9.4 kg ha 1 Weeds were removed by rotary hoe if needed. Irrigation was applied as needed, typically when rainfall was insufficient to provide approx. 25 mm of water per week. At harvest, a 4 m length of one of the inner two rows of ea ch subplot was harvested at 7.5 cm stubble height and weighed fresh in the field to determine fresh biomass yield. A subsample of six r epresentative tillers was collected and weighed fresh in the field. The subsample was partitioned into leaves, stems, and panicles plus grain (grain heads). These partitioned samples were weighed fresh in the f ield, and then
53 oven dried at 50 C to a constan t weight, which was used to determine dry matter concentration and total dry matter yield Water Use and Water Use Efficiency Sweet sorghum water use (i.e., plant transpiration) was measured throughout the growing season for each of the three planting dat es using sap flow sensors (Dynamax, Houston, TX) installed in situ on selected intact plant stems (Sakuratani, 19 81; Sakuratani, 1987; and Baker and van Bavel, 1987). Although time and labor intensive, this heat balance method measures water use under actu al field conditions without altering the microclimate (e.g., chamber methods) or the soil profile (e.g., lysimeter methods), and has been used to accurately measure crop water use for sorghum (Dugas et al., 1998), corn (Bethenod et al., 2000), sugarcane (S aliendra and Meinzer, 1992) and other herbaceous crop species like soybean and cotton (Tan and Buttery, 1994; Dugas et al., 1998). Approximately every 2 to 3 wk three to four representative sweet sorghum tillers for each PD were selected from the inner 2 m 2 of the plot for sensor installation (maximum of 16 sensors). Prior to installation, any leaf sheath tissue was removed and stem diameter was measured in two directions (N S and E W) at sensor height (equidistant between two internodes) and averaged to e stimate stem sap flow area. Before placing sensors on the stem, the area was sprayed with canola oil to maximize sensor contact with the stem. The foam insulated sensor was then placed on the st em and wrapped with an aluminum covered bubble wrap to shield the sensor and stem from solar radiation. Finally, a conical shaped plastic piece was wrapped around the stem above the sensor unit to prevent irrigation or rain water from moving downward along the stem toward the sensor.
54 Once installed, all sensors were left on the stem for 5 to 7 days. A 12 volt deep cycle battery, a CR1000 data logger (Campbell Scientific, Logan, UT), and Dynamax software program ( Dynagage flow 32 1k ver 18.104.22.168 ) were used to heat the sensors and to monitor thermocouple temperatures fr om each of the sensors. The temperature data were recorded at 15 s interva ls and averaged every 15 min and stored by the datalogger. Using the input stem area and measured temperature data, the software program directly calculated average tiller water use in g h 1 over the 15 min interval and data for each tiller were stored by the datalogger for each tiller. Measurements were repeated approx. every 2 3 wk during growing seas on until harvest Daily E C for each of the plots was then calculated based on the product of the mean measured sap flow (g hr 1 cm 2 stem area), average stem area (cm 2 ), and tiller density per plot. Stem diameter data were collected monthly on 40 randomly selected tillers at sensor height to obtain the mean tiller diameter for each plot The number of tillers in 8 m 2 of the inner two rows of the plot were also counted monthly. For days whe n sensor sap flow was not measured, daily E C was estimated as the product of ET O and K c anopy where values of K c anopy were linearly interpolated for ea ch day across the measured values, and from K c anopy = 0 at crop emergence to the K c anopy value approximately 3 weeks after emergence each year. Total seasonal E C was then estimated as the sum of daily E C from emergence to harvest. Water use efficiency for each PD was calculated from the quotient of harvested dry biomass and total seasonal E C Data Analyses Statistical analyses on year and planting date effects within year were performed using analysis of variance procedures in the GLIMMIX procedure of SAS (SAS Institute,
55 2009). Planting date and year were treated as fixed effect s and block was treated as a random effect in the model. R esiduals from each model fit were analyzed for homogeneity of variance visually and for normality visually and with the Shap iro Wilk W test. Degrees of freedom were determined using the Kenward Roger method. Where significant ( P < 0.05) fixed effects were seen, treatment mean pairwise comparisons were made using the LSMEANS statement with the TUKEY method. Results Weather Data Weather variables differed across planting date growing intervals (Figure 3 1 ; Table 3 2 ). Daily air temperature ( T AIR ) and RH were generally similar for PD2 and PD3, both of which were greater than PD1 during both the 2009 and 2010 growing seasons (Table 3 2). In 2010, T AIR was greate r for each of the respective PD s compared to 2009. Average daily solar radiation was greatest for PD1, intermediate for PD2, and least for PD3 during both growing seasons, but radiation was higher for 2010 compared to 2009 (Ta ble 3 2). Although PD3 essentially coincided with the beginning of the wet season in the region (mid June), total rainfall received during the PD3 growth cycle was less than either PD1 or PD2 during both 2009 and 2010 (Table 3 2) Additionally, rainfall wa s greater for PD2 compared to PD1 in both 2009 and 2010. Finally, ET O was greatest for PD2, a little less for PD1, and considerably less for PD3 during both 2009 and 2010 (Table 3 2). Sweet Sorghum Biomass Yield Across all PD, sorghum stand populations at harvest were 19% greater in 2009 compared to 2010 despite the same seeding rate for all plots (Table 3 3). Across both growing seasons, plant populations were also greater for PD1 compared to PD2 and
56 PD3, which did not differ (Table 3 4). Total so rghum dry matter yields which included all abo ve ground biomass components, were above 20 Mg ha 1 for PD1 and PD2, which did differ (Table 3 4). However, total dry matter yields decreased by about 20% for PD3 compared to the earlier two PD Sweet Sorghum Water Us e and WUE Total seasonal E C in 2009 was similar between PD1 and PD2, which were greater than PD3 (Table 3 5). However, in 2010 E C was greatest for PD1, intermediate for PD2, and least for PD3. Within PD no significant difference in E C was seen across the t wo growing seasons. Seasonal patterns in E C indicated relatively high E C shortly after emergence during vegetative growth followed by a decreasing trend towards maturity (Figure 3 2). Additionally, relatively high E C was evident in the early summer growing season coinciding with long days and high evaporative demand. Given the seasonal patterns in E C and biomass production, WUE was lowest for PD1 compared to PD2 and PD3, which did not differ (Table 3 3). Discussion I t is frequently suggested that sweet sorg hum can grow on marginal soils and is efficient in nutrient and water use compared to other crops. (e.g., Evans and Cohen, 2009). However, to date there have been few credible data to support these assertions. Over the present study, total season sorghum E C ranged from about 350 to 750 mm per crop. These data along with other recent studies indicate that water use of sorghum is likely to be similar to that of other C4 grasses, although this is not to say that sorghum cannot survive and tolerate water defici t better than many other C4 crops. For example, studies from the US Southern High Plains show that forage sorghum could be grown on less water than corn but WUE (i.e. biomass p roduced per unit water use) was
57 comparable to corn (Howell et al., 2008). Simil ar studies at the same location have shown that forage sorghum yield increased linearly with a slope of 0. 7 3 Mg biomass per hectare per cm of water (from 380 to 760 mm applied water) ( Bean and McCollum, 2006 ) Thus, while sorghum may possess the advantage of being able to survive under severe water deficit and perhaps under potential watering restrictions, E C by intensively managed sorghum for optimal yield is likely to be similar to other C4 grass crops. That being said, management practices can be used to help minimize effects of sorghum crop production on water resources. Total seasonal E C of sweet sorghum decreased with each delay in planting date from PD1 to PD3 in both seasons, although the difference was not significant between PD1 and PD2 in 2009 (Ta ble 3 5). The lower E C of PD3 in both years was associated with reduced radiation, reduced rainfall, and reduced total dry matter accumulation. Similar or greater E C of PD1 compared to PD2 was consistent with greater average daily radiation and lower RH. R eference ET was lower for PD3, which was consistent with our results, but, interestingly, ET O was greater for PD2 compared to PD1 (not to the same magnitude as differences for PD3), which was not consistent with our data. Overall, we found about a 20% red uction in total dry matter yield when sorghum was planted late (PD3) compared to PD1 and PD2 (Table 3 4). Similar reductions in yield with relatively late planting have also been reported by Blum (1972) who found increased grain yield when grain sorghum wa s planted in late March compared with a mid April planting in Israel, and by Allen and Musick (1993) who reported greater grain yield of sorghum in Texas under adequate irrigation when the crop was planted in May compared to those planted in June. Broadhea d (1969) also reported lower stalk and
58 June in Mississippi compared with April and May planting date. The reduction in yield with PD3 was most likely associated with reduced average daily radiati on, which was more than 10% lower for PD3 over the whole interval, but also could have been due to reduced rainfall for PD3 compared to PD1 and PD2, although irrigation was also used throughout the study. Given similar yields between PD1 and PD2 and crop transpiration for PD1 compared to PD2 (at least in 2010), water use efficiency of PD1 was relatively low compared with PD2 and PD3 (Table 3 3). Significantly lower crop transpiration for PD3 compared to PD2 did not result in greater WUE for PD3 due to corr espondingly lower dry matter yields for PD3 compared to PD2. In contrast, Blum (1972) reported differences in WUE, but not in transpiration (inferred from pan evaporation) for sorghum planted in late March and mid April in coastal Israel because of higher yield associated with the earlier sowing date. Overall, WUE ranged from about 2.7 to 4.1 g kg 1 for sorghum in the present study, which is consistent with other C4 grasses (Stanhill, 1986; Connor et al., 2011 ; Chapter 2). Taken together, the results from the present study indicated that PD2 ( e arly May) was optimal for both sorghum production and water resources. Of the three PD in the study, PD2 possessed high total dry matter yields, moderate overall transpiration, and relatively high WUE. In contrast, PD 1 had similar dry matter yields as PD2, but had higher transpiration and lower WUE and PD3 had low transpiration compared to PD2, but also lower total dry matter yields and therefore no difference in WUE.
59 Table 3 1 Planting dates (PD), harvest dates (H D), and number of days from planting date to harvest date (DTH) for Topper 76 6 sweet sorghum during the 2009 and 2010 growing seasons at Citra, FL Note: crop harvested at late soft dough stage. PD treatment 2009 2010 PD HD DTH PD HD DTH PD1 Mar 31 Jul 30 121 Apr 1 Jul 27 117 PD2 May 5 Sep 10 128 May 5 Sep 3 121 PD3 Jun 10 Oct 6 118 Jun 10 Oct 6 118 Table 3 2 Average daily relative humidity (RH), solar radiation ( Q ), air temperature at 2 m ( T AIR ), total rainfall, and reference evapotranspirat ion (ET O ) during each of the planting date (PD) growth intervals for 2009 (A) and 2010 (B) at Citra, FL. Year PD RH Q T AIR Rainfall ET O % W m 2 C mm mm 2009 PD1 77.3 227 24.7 529 439 PD2 81.7 210 25.9 597 455 PD3 81.8 197 26.0 375 385 2010 PD1 7 7.0 244 26.2 397 460 PD2 80.4 231 27.6 444 473 PD3 81.2 209 27.1 345 412
60 Table 3 3 Main e ffect of year on sorghum population at harvest and water use efficiency (WUE) of Topper 76 6 sweet sorghum grown during the 2009 and 2010 growing seasons at Ci tra, FL Pop. WUE N o. ha 1 g kg 1 2009 175,300 A 3.10B 2010 147,100B 3.88 A s.e. 7019 0.26 Numbers followed by the same letter within a column do not differ ( P > 0.05) ; Standard error of differences of species means Table 3 4 Main e ffect of pl anting date (PD) on dry matter yield, sorghum population at harvest, and water use efficiency (WUE) of Topper 76 6 sweet sorghum grown during the 2009 and 2010 growing seasons at Citra, FL PD treatment Dry matter yield Population WUE Mg ha 1 No. ha 1 g kg 1 PD1 20.7 A 197,700 A 2.68B PD2 22.4 A 150,100B 3.71 A PD3 17.2B 135,800B 4.07 A s.e. 1.16 8596 0.23 Numbers followed by the same letter within a column do not differ ( P > 0.05) ; Standard error of differences of species means
61 Table 3 5 Effect of year x planting date (PD) interaction on total seasonal transpiration ( E C ) of Topper 76 6 sweet sorghum grown during the 2009 and 2010 growing seasons at Citra, FL Year PD E C mm 2009 PD1 750AB PD2 652BC PD3 480DE 2010 PD1 794 A PD2 563CD PD3 389E s.e. 37.5 Numbers followed by the same letter within a column do not differ (P > 0.05) ; Standard error of differences of species means
62 Figure 3 1. Average daily relative humidity (RH), solar radiation ( Q ), air temperature at 2 m ( T AIR ), and total rainfall for 2009 (A) and 2010 (B) at Citra, FL.
63 Figure 3 2 Average daily Topper 76 6 sweet sorghum transpiration (E) during the 2009 (A) and 2010 (B) growing seasons as affected by planting date (PD)
64 CHAPTER 4 SUMMARY AND CONCL USIONS Results from the perennial grass study (Chapter 2) indicated higher yield and WUE of C4 elephantgrass and energycane compared with those of C3 giant reed. The higher yield of the C4 grasses was likely due in large part to their greater WUE, as highe r dry matter yields of giant reed have been observed with greater water availability. This clearly demonstrated the superior capability of these C4 grasses to use water more efficiently in producing biomass, which is a critical consideration when assessing candidate species to be used as biofuel crops. In regions of the southeastern US especially in Florida, where coarse textured, well drained soils are dominant, and in area s where water resources are relatively scare C4 grass es especially e lephantgrass, would be the most suitable bioenergy crop choice based on the results of the present study It should be noted though that transpiration was relatively hig h overall for all of the crops due to their long growing seasons. Results from the sweet sorghum pl anting date study (Chapter 3) indicated that, with adequate available water, water use of sorghum was comparable to other C4 grasses, including maize and the perennial grasses in Chapter 2. However, given the relatively short growing season for current com mercial sweet sorghum cultivars, planting date (PD) could be used to simultaneous ly optimize crop yield and water use. Sweet sorghum planted early in the growing season (PD1) produced optimal yields, but used more water than sweet sorghum planted on PD2. S weet sorghum planted on PD3 used less water than either PD1 or PD2, but yields were also about 20% lower for PD3 compared to PD1 and PD2. Thus, PD2 (early May) was identified as the best PD for Topper 76 6 sweet sorghum when considering both yield and wate r use together.
65 LIST OF REFERENCES Allen R.R., Musick J.T. (1993) Planting date, water management, and maturity length relationships for irrigated grain sorghum. Trans. Am. Soc. Agric. Eng. 36:1123 1129. Allen R.G., Pereira L.S., Raes D., Smith M. (1998) Crop evapotranspiration. Guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. Food and Agriculture Organization of United Nations. Rome, Italy. Amaducci S., Monti A., Venturi G. (2004) Non structural carbohydrates and fi bre components in sweet and fibre sorghum as affected by low and normal input techniques. Ind Crop Prod 20:111 118. Angelini L.G., Ceccarini L., Bonari E. (2005) Biomass yield and energy balance of giant reed (Arundo donax L.) cropped in central Italy a s related to different management practices. Eur J. Agron 22:375 389. Azam Alia S.N., Gregorya P.J., Monteitha J.L. (1984) Effects of p lanting d ensity on w ater u se and p roductivity of p earl m illet ( Pennisetum Typhoides ) g rown on s tored w ater. II. W ater u se, l ight i nterception and d ry m atter p roduction. Exp. Agric. 20:215 224. Baker J.M., van Bavel C.H.M (1987) Measurement of mass flow of water in the stems of herbaceous plants. Plant Cell Environ. 10:777 782. Barber, C.A. (1920) The origin of sugarcane. Int Sugar J 22:249 251. Baucum L.E., Rice R.W. (2009) An o verview of Florida s ugarcane Agronomy Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL. Beale C.V., Morison J.I.L., Long S.P. (1999) Water use efficiency of C4 perennial grasses in a temperate climate. Agr. Forest. Meterol. 96:103 115. Bean B. McCollum T. (2006) Summary of six years of forage sorghum variety trials. Texas Coop. Ext. Serv. Pub. SCS 2006 4. Colle ge Station, TX. Berndes G. (2002) Bioenergy and water -the implications of large scale bioenergy production for water use and supply. Global Environ. Chang. 12:253 271. Berry J., Bjorkman O. (1980) Photosynthetic r esponse and a daptation to t emperature in H igher Plants. Annu Rev Plant Phys 31:491 543. Bethenod O., Katerji N., Goujet R., Bertolini J.M., Rana G. (2000) Determination and validation of corn crop transpirationby sap flow measurement under field conditions. Theor. Appl. Climatol. 67:153 160.
66 Bischoff K.P., Gravois K.A., Reagan T.E., Hoy J.W., Kimbeng C.A., LaBorde C.M., Hawkins G.L. (2008) Registration of 'L79 1002' Sugarcane. J Plant Regis 2:211 217. Blum A. (1972) Effect of Planting Date on Water Use and Its Efficiency in Dryland Grain Sor ghum. Agron. J. 64:775 778. Bracho R., Powell T.L., Dore S., Li J., Hinkle C.R., Drake B.G. (2008) Environmental and biological controls on water and energy exchange in Florida scrub oak and pine flatwoods ecosystems. J. Geophys. Res. 113. Broadhead D.M. (1969) Sugar p roduction f rom s weet s orghum as a ffected by p lanting d ate, a fter r ipe h arvesting, and s torage. Agron. J. 61:811 812. Burvall J. (1997) Influence of harvest time and soil type on fuel quality in reed canary grass ( Phalaris arundinacea L.). Bio mass Bioenergy 12:149 154. Carpenter S.R., Caraco N.F., Correll D.L., Howarth R.W., Sharpley A.N., Smith V.H. (1998) N onpoint p ollution of s urface w aters with p hosphorus and n itrogen Ecol. Appl. 8:559 568. Cassman K., Eidman V., Simpson E. (2006) Converg ence of agriculture and energy: Implications for research and policy, CAST Commentary QTA 2006 3, Ames, IA. Cassman K.G., Liska A.J. (2007) Food and fuel for all: realistic or foolish? Biofpr. 1:18 23. Center for Sustainable Systems U.o.M. (201 0 ) US Renew able Energy Factsheet. Chabot R., Bouarfa S., Zimmer D., Chaumont C., Moreau S. (2005) Evaluation of the sap flow determined with a heat balance method to measure the transpiration of a sugarcane canopy. Agr. Water Manage. 75:10 24. Cherubini F., Bird N.D. Cowie A., Jungmeier G., Schlamadinger B., Woess Gallasch S. (2009) Energy and greenhouse gas based LCA of biofuel and bioenergy systems: Key issues, ranges and recommendations. Resour. Conserv. Recy. 53:434 447. Christian D.G., Riche A.B., Yates N.E. (2 008) Growth, yield and mineral content of Miscanthus x giganteus grown as a biofuel for 14 successive harvests. Ind Crop Prod 28:320 327. Clifton Brown J.C., Lewandowski I. (2000) Water u se e fficiency and b iomass p artitioning of t hree d ifferent Miscanth us g enotypes with l imited and u nlimited w ater s upply. Ann. Bot. 86:191 200. Connor D.J., Loomis R.S., Cassman K.G. (2011) Crop ecology: productivity and management in agricultural systems. 2 nd edition, Cambridge University Press, New York. 562 pp.
67 Davis A .S., Cousens R.D., Hill J., Mack R.N., Simberloff D., Raghu S. (2010) Screening bioenergy feedstock crops to mitigate invasion risk. Front. Ecol. Environ. 8:533 539. Day J.L., Duncan R.R., Raymer P.L., Lovell G.R., Thompson D.S., Garrett H.D., Zummo. N. (1 995) Top 76 6: A new sweet sorghum variety for sirup production., Research Report. pp. 4. Department of Energy (DOE) ( 2006 ) Breaking the biological barriers to cellulosic ethanol. U.S. Dept. of Energy, Washington D.C. Department of Energy (DOE) (2011) U.S. billion ton update: Biomass supply for a bioenergy and bioproducts industry R.D. Perlack and B.J. Stokes (Leads), ORNL/TM 2011/224. Oak Ridge National Laboratory, Oak Ridge, TN. 227 p. Doggett H ( 1988 ) Sorghum 2 nd edition, Wiley and Sons, New York. 512 pp. Dugas W.A. (1990) Comparative measurement of stem flow and transpiration in cotton. Theor. Appl. Climatol. 42:215 221. Dugas W.A., Heuer M.L., Hunsaker D., Kimbal B.A., Lewin K.F., Nagy J., Johnson M. (1994) Sap f low m easurements of t ranspiration fr om c otton g rown u nder a mbient and e nriched CO 2 c oncentrations. Agri. Forest Meteorol. 70:231 245. Dugas W.A., Prior S.A., Rogers H.H. (1998) Transpiration from sorghum and soybean growing under ambient and elevated CO 2 concentrations. Agr. Forest Meteorol. 83:37 48. Erickson JE, ZR Helsel, KR Woodard, et al. 2011. Planting date affects biomass and brix of sweet sorghum grown for biofuel across Florida. Agron J 103:1827 1833. Erickson JE, KR Woodard, and LE Sollenberger. 2012. Optimizing sweet sorghum prod uction for biofuel in the southeastern USA through nitrogen fertilization and top removal. BioEnergy Research 5:86 94. Evans J.M., Cohen M.J. (2009) Regional water resource implications of bioethanol production in the Southeastern United States. Glob. Chan ge Biol. 15:2261 2273. FAO. (2012) FAOSTAT. Available at http://www.fao.org/nr/water/aquastat/ water_use/index6.stm (verified 5 March 2012) FAO, Rome, Italy. Fares A., Alva A.K. (2000) Evaluation of capacitance probes for optimal irrigation of citrus thro ugh soil moisture monitoring in an entisol profile. Irrigation Science 19:57 64. Farrell A.E., Plevin R.J., Turner B.T., Jones A.D., O'Hare M., Kammen D.M. (2006) Ethanol Can Contribute to Energy and Environmental Goals. Science 311:506 508.
68 FAWN. (2012) Available at http:// fawn.ifas.ufl.edu (verified 5 March 2012) University of Florida Gainesville, FL, USA Fedenko J.R. (2011) Biomass yield and composition of potential bioenergy feedstocks. Thesis. University of Florida, Gainesville, FL. Fougerouze J. ( 1966) Quelques problems de bioclimatologie en Guyanne Francaise. L'Agron. Tropicale 3:291 346. Geng S., Hills F.J., Johnson S.S., Sah R.N. (1989) Potential y ields and o n f arm Ethanol p roduction c ost of c orn, s weet s orghum, f odderbeet, and s ugarbeet. J. Ag ron Crop Sci 162:21 29. Gholz H.L., Clark K.L. (2002) Energy exchange across a chronosequence of slash pine forests in Florida. Agr Forest Meteorol 112:87 102. Gimeno V., Fernndez Martnez J.M., Fereres E. (1989) Winter planting as a means of drought escape in sunflower. Field Crops Res 22:307 316. Goff B.M., Moore K.J., Fales S.L., Heaton E.A. (2010) Double c ropping s orghum for b iomass. Agron. J. 102:1586 1592. Hallam A., Anderson C., Buxton D.R. (2001) Comparative economic analysis of perennial, a nnual, and intercrops for biomass production. Biomass Bioenergy 21:407 424. Hill J., Nelson E., Tilman D., Polasky S., Tiffany D. (2006) Environmental, economic, and energetic costs and benefits of biodiesel and ethanol biofuels. Proc. Nat. Acad. Sci. USA 103:11206 11210. Howell T.A., Evett S.R., Tolk J.A., Copeland K.S., Colaizzi P.D., Gowda P.H., USDA A. (2008) Evapotranspiration of c orn and f orage s orghum for s ilage, w orld Environmental and Water Resources Congress 2008 Ahupua'a, Bushland, TX. Jara J., S tockle C.O., Kjelgaard J. (1998) Measurement of evapotranspiration and its components in a corn ( Zea Mays L.) field. Agr. Forest. Meterol. 92:131 145. Jrgensen U., Schelde K. (2001) Energy crop water and nutrient use efficiency (prepared for the Internati onal Energy Agency IEA Bioenergy Task 17, Short Rotation Crops), Danish Institute of Agricultural Sciences (DIAS), Department of Crop Physiology and Soil Science Research Centre Foulum, Tjele, Denmark. Keeney D.R., DeLuca T.H. (1992) Biomass as an energy s ource for the midwestern U.S. Am J Alternative Agr 7:137 144. Kering M., Biermacher J., Butler T., Mosali J., Guretzky J. (2012) Biomass y ield and n utrient r esponses of s witchgrass to p hosphorus a pplication. BioEnergy Research 5:71 78.
69 Knoll J.E., And erson W.F., Strickland T.C. (2011) Low input production of biomass from perennial grasses in the Coastal Plain of Georgia, USA. BioEnergy Research. Kort J., Collins M., Ditsch D. (1998) A review of soil erosion potential associated with biomass crops. Bio mass Bioenergy 14:351 359. Lewandowski I., Clifton Brown J.C., Scurlock J.M.O., Huisman W. (2000) Miscanthus : European experience with a novel energy crop. Biomass Bioenergy 19:209 227. Lewandowski I., Scurlock J.M.O., Lindvall E., Christou M. (2003) The d evelopment and current status of perennial rhizomatous grasses as energy crops in the US and Europe. Biomass Bioenergy 25:335 361. Lingle S.E., T. L. Tew, Rukavina H., Boykin D.L. (2012) Post harvest changes in sweet sorghum I: Brix and sugars. Bioenergy R es. 5:158 167. Marris E. (2008) Water: More crop per drop. Nature 452:273 277. Mastrorilli M., Katerji N., Rana G. (1999) Productivity and water use efficiency of sweet sorghum as affected by soil water deficit occurring at different vegetative growth stag es. Eur. J. Agron. 11:207 215. McGroary P.C., Cisar J.L., Snyder G.H., Erickson J.E., Daroub S.H., Sartain J.B. (2011) Water Use of St. Augustinegrass and Bahiagrass under Varying Nitrogen Rates. Agron. J. 103:100 106. McKendry P. (2002 a ) Energy production from biomass (part 1): overview of biomass. Bioresour. Technol. 83:37 46. McKendry P. (2002 b ) Energy production from biomass (part 2): conversion technologies. Bioresour. Technol. 83:47 54. McLaughlin S.B., Walsh M.E. (1998) Evaluating environmental conse quences of producing herbaceous crops for bioenergy. Biomass Bioenergy 14:317 324. McLaughlin S.B., Adams Kszos L. (2005) Development of switchgrass (Panicum virgatum) as a bioenergy feedstock in the United States. Biomass Bioenergy 28:515 535. Miller A.N. Ottman M.J. (2010) Irrigation f requency e ffects on g rowth and e thanol y ield in s weet s orghum. Agron. J. 102:60 70. Mukherjee, S.K. (1957) Origin and distribution of Saccharum. Botanical Gazette 119:55 61. Newman Y.C., Erickson J.E., Vermerris W.E., Hanc ock D., Wright D. ( 2010 ) United s orghum c heckoff p rogram: East f orage p roduction g uide. United Sorghum Checkoff Program, Lubbock, TX. 101 pp.
70 Olufayo A., Baldy C., Ruelle P. (1996) Sorghum yield, water use and canopy temperatures under different levels of irrigation. Agr. Water Manage. 30:77 90. Pimentel D. (2003) Ethanol f uels: Energy b alance, e conomics, and e nvironmental i mpacts a re n egative. Natural Resources Research 12:127 134. Ragauskas A.J., Williams C.K., Davison B.H., Britovsek G., Cairney J., Ecke rt C.A., Frederick W.J., Hallett J.P., Leak D.J., Liotta C.L., Mielenz J.R., Murphy R., Templer R., Tschaplinski T. (2006) The p ath f orward for b iofuels and b iomaterials. Science 311:484 489. Rainbolt C., Gilbert R. (2008) Production of b iofuel c rops in F lorida: s ugarcane/ e nergycane, Agronomy Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL. Rana G., Katerji N. (2000) Measurement and estimation of actual evapotranspiratio n in the field under Mediterranean climate: a review. Eur. J. Agron. 13:125 153. Renewable Fuels Association. (2012) Industry statistics. Available at http://ethanolrfa.org (verified 5 March 2012) Renewable Fuels Association, Washington DC, USA. Rooney W.L ., Blumenthal J., Bean B., Mullet J.E. (2007) Designing sorghum as a dedicated bioenergy feedstock Biofpr. 1:147 157. Sage R.F. (2000) C3 versus C4 photosynthesis in rice: E cophysiological perspectives, in: P. L. M. J.E. Sheehy and B. Hardy (Eds.), Studi es in Plant Science, Elsevier. pp. 13 35. Sakuratani T. (1981) A h eat b alance m ethod for m easuring w ater f lux in the s tem of i ntact p lants. Agr. Meteorol. 37:9 17. Sakuratani T. (1987) Studies on e vapotranspiration from c rops (2) s eparate e stimation of t ra nspiration and e vaporation from a s oybean f ield without w ater s hortage. Agr. Meteorol. 42:309 317. Saliendra N.Z., Meinzer F.C. (1992) Genotypic, d evelopmental and d rought i nduced d ifferences in r oot h ydraulic c onductance of c ontrasting s ugarcane c ultivars J Exp Bot 43:1209 1217. Samson R., Mani S., Boddey R., Sokhansanj S., Quesada D., Urquiaga S., Reis V., Ho Lem C. (2005) The potential of C 4 p erennial g rasses for d eveloping a g lobal BIOHEAT Industry. Crit. Rev. Plant Sci. 24:461 495.
71 Searchinger T. Heimlich R., Houghton R.A., Dong F., Elobeid A., Fabiosa J., Tokgoz S., Hayes D., Yu T. H. (2008) Use of US c roplands for b iofuels i ncreases g reenhouse g ases t hrough e missions from l and u se c hange. Science 319:1238 1240. Sissine F. (2007) Energy i ndepen dence and s ecurity a ct of 2007: A s ummary of major p rovisions. CRS Report for Congress:27. Soileau J.M., Bradford B.N. Biomass and s ugar y ield r esponse of s weet s orghum to l ime and f ertilizer. Agron. J. 77:471 475. Stanhill G. (1986) Water Use Efficiency, in: N. C. Brady (Ed.), Advances in a gronomy, Academic Press. pp. 53 85. Steduto P., Katerji N., Puertos Molina H., UÂ¨nluÂ¨ M., Mastrorilli M., Rana G. (1997) Water use efficiency of sweet sorghum under water stress conditions Gas exchange investigations a t leaf and canopy scales. Field Crops Res. 54:221 234. Sumner D.M., Jacobs J.M. (2005) Utility of Penman Monteith, Priestley Taylor, reference evapotranspiration, and pan evaporation methods to estimate pasture evapotranspiration. J. Hydrol 308:81 104. Ta mang P.L., Bronson K.F., Malapati A., Schwartz R., Johnson J., Moore Kucera J. (2011) Nitrogen r equirements for e thanol p roduction from s weet and p hotoperiod s ensitive s orghums in the s outhern h igh p lains. Agron. J. 103:431 440. Tan C.S., Buttery B.R. (19 95) Determination of the water use of two pairs of soybean isolines differing in stomatal frequency using a heat balance stem flow gauge. Can. J. Plant Sci. 75:99 103.. Tarpley L., Lingle S.E., Vietor D.M., Andrews D.L., Miller F.R. Enzymatic c ontrol of n o nstructural c arbohydrate c oncentrations in s tems and p anicles of s orghum. Crop Sci. 34:446 452. USDA NASS. (2012) Crop acreage and value: Corn. Available at http://www.nass.usda.gov (verified 5 March 2012) USDA, Washington DC, USA. Vermerris W.E., Erickso n J.E., Wright D., Newman Y., Rainbolt C. (2011) Production of b iofuel c rops in Florida: Sweet s orghum, Agronomy Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL. Wiggans D.R., Singer J.W., Moore K.J., Lamkey K.R. (2012) Maize w ater u se in l iving m ulch s ystems with s tover r emoval. Crop Sci. 52:327 338. Wong S.C. (1979) Elevated atmospheric partial pressure of CO 2 and plant growth. Oecologia 44:68 74.
72 Woodard K.R., Prine G .M. (1993) Regional performance of tall tropical bunchgrasses in the Southeastern US A. Biomass Bioenergy 5:3 21. Woodard K.R., Sollenberger L.E. (2008) Production of b iofuel c rops in Florida: Elephantgrass, Agronomy Department, Florida Cooperative Extensio n Service, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL. Wortmann C.S., Liska A.J., Ferguson R.B., Lyon D.J., Klein R.N., Dweikat I. (2010) Dryland p erformance of s weet s orghum and g rain c rops for b iofuel in Nebraska. Agron. J. 102:319 326. Zhao Y.L., Dolat A., Steinberger Y., Wang X., Osman A., Xie G.H. (2009) Biomass yield and changes in chemical composition of sweet sorghum cultivars grown for biofuel. Field Crops Res. 111:55 64.
73 BIOGRAPHICAL SK ETCH Arkorn Soikaew was born in 1985 in Kanchanaburi, Thailand and received a b achelor of s cience in b iology (distinction program) in 2006 from Mahidol University, Thailand. He joined the Master in Interdisciplinary Ecology program at the University of Flo rida in 2009 under the supervision of Dr. John Erickson of the Agronomy Department under the support of the Royal Thai Government Scholarship