Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations
2021
Estimation of soybean radiation use efficiency with the help of UAV imagery to evaluate the impact of frogeye leaf spot on yield and the effect of fungicides.
Xavier Alaric Phillips Iowa State University
Follow this and additional works at: https://lib.dr.iastate.edu/etd
Recommended Citation Phillips, Xavier Alaric, "Estimation of soybean radiation use efficiency with the help ofA U V imagery to evaluate the impact of frogeye leaf spot on yield and the effect of fungicides." (2021). Graduate Theses and Dissertations. 18586. https://lib.dr.iastate.edu/etd/18586
This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Estimation of soybean radiation use efficiency with the help of UAV imagery to evaluate the impact of frogeye leaf spot on yield and the effect of fungicides.
by
Xavier Phillips
A dissertation submitted to the graduate faculty
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
Major: Plant Pathology
Program of Study Committee: Daren Mueller, Major Professor Alison Robertson Leonor Leandro Matthew Darr Mark Licht
The student author, whose presentation of the scholarship herein was approved by the program of study committee, is solely responsible for the content of this dissertation. The Graduate College will ensure this dissertation is globally accessible and will not permit alterations after a degree is conferred.
Iowa State University
Ames, Iowa
2021
Copyright © Xavier Phillips, 2021. All rights reserved. ii
TABLE OF CONTENTS
Page
NOMENCLATURE ...... iv
ACKNOWLEDGMENTS ...... v
ABSTRACT ...... vi
CHAPTER 1. GENERAL INTRODUCTION ...... 1 Dissertation Format ...... 1 General Introduction ...... 1 Literature Review ...... 2 Soybeans ...... 2 Diseases of Soybean ...... 3 Frogeye Leaf Spot ...... 3 Fungicide Use ...... 6 Disease Severity Assessment and Relation to Yield Loss ...... 8 Area Under Disease Progress Curve ...... 9 Radiation Use Efficiency (RUE) ...... 9 Components of RUEgreen ...... 10 RUE of Soybeans ...... 14 UAVs in Agriculture ...... 15 Passive and Active Sensors ...... 16 Radiometric Calibration ...... 18 Indices ...... 19 Literature Cited ...... 20
CHAPTER 2. ESTIMATING SOYBEAN RADIATION USE EFFICIENCY USING A UAV IN IOWA ...... 33 Abstract ...... 33 Introduction ...... 34 Materials and Methods ...... 36 Results ...... 41 Discussion ...... 43 Conclusion ...... 46 Tables and Figures ...... 47 Literature Cited ...... 57 Appendix. Tables ...... 62
CHAPTER 3. A COMPARISON BETWEEN DISEASE SEVERITY OF FROGEYE LEAF SPOT AND RADIATION USE EFFICIENCY OF SOYBEAN ACROSS FUNGICIDE TREATMENTS ...... 64 Abstract ...... 64 Introduction ...... 65 Materials and Methods ...... 69 iii
Results ...... 76 Discussion ...... 79 Conclusion ...... 84 Tables and Figures ...... 85 Literature Cited ...... 96 Appendix. Tables and Figures ...... 103
CHAPTER 4. GENERAL CONCLUSION ...... 110 iv
NOMENCLATURE
GLAI Green Leaf Area Index
GLAIDx GLAI With Disease Severity Removed
LAI Leaf Area Index
NDVI Normalized Difference Vegetation Index
RUE Radiation Use Efficiency
SPAD Soil Plant Analysis Development
TLAI Total Leaf Area Index
UAV Unmanned Aerial Vehicle
v
ACKNOWLEDGMENTS
I cannot express how grateful I am to have had this opportunity here at Iowa State
University. The amazing Mueller Lab and IPM group that I found myself in, is composed of the most kind and helpful people that I have ever met. I thank all of you for making me feel welcome and for the invaluable help you provided me. I also thank my committee members Dr. Alison
Robertson, Dr. Leonor Leandro, Dr. Mark Licht, and Dr. Matt Darr for their support and guidance. I thank you for accepting the commitment of being on my committee.
I am thankful to have had Dr. Daren Mueller as my advisor throughout my graduate studies. Your confidence helped me foster my own self-efficacy. I appreciate the continued support and mentorship you provided over all of the challenging adventures I encountered.
Special thanks to my parents for their support. I appreciate your trips to visit us and for understanding when we were unable to get back. Finally, I thank my wife and children for taking care of me. Your smiles, laughter, and love get me through each day. vi
ABSTRACT
Radiation use efficiency (RUE) is a parameter that is directly associated to yield, but difficult to estimate and unreasonable to estimate on a small plot research scale using traditional techniques. It has the potential to be used for evaluation of in-season management techniques and disease impact. During 2018 and 2019, field experiments located in north central and south western Iowa were used to investigate the estimation of RUE of soybean (Glycine max (L.)
Merr.) and the effects of foliar fungicides across varying levels of frogeye leaf spot severity caused by the pathogen Cercospora sojina K. Hara. The objective of this research was to (i) demonstrate that it is possible to estimate RUE for soybean based on reflectance data derived from consumer-grade UAV imagery, (ii) determine how foliar fungicides affected frogeye leaf spot, remotely sensed plant health indicators, and soybean yield, and (iii) compare the impact of foliar fungicides and frogeye leaf spot on RUE estimated by using UAV reflectance data with a reduced sampling approach across foliar fungicide treatments. RUE of soybean was estimated by using UAVs to capture imagery on a high spatial and temporal scale. Values ranged from 0.98 to
1.07 and 0.96 to 1.12 across the entire season with 14 sampling dates and the period post- fungicide application with eight sampling dates, respectively. These values fall within the range of previously published soybean RUE values. The highest levels of frogeye leaf spot severity were recorded in Lewis 2018, and were just under 20%. At this location, the application of flutriafol + fluoxastrobin resulted in 19% more yield (P = 0.080) than the non-treated control. In addition, the frogeye leaf spot severity of the flutriafol + fluoxastrobin treatment was significantly (P = 0.010) less than the severity of the fluxapyroxad + pyraclostrobin treatment and non-treated control. Applications of foliar fungicides significantly (P = 0.012) increased canopy cover, but NDVI, SPAD, and RUE values did not differ between fungicide treatments. vii
All RUE values estimated in Lewis 2018, Kanawha 2018, and Kanawha 2019 fell within the range of known values even with the reduced sampling approach. Pathogen specific impacts on soybean may confound the variables needed to estimate RUE. Using RUE to estimate the impact of disease on yield remotely may be a valuable resource; however, confounding factors will require additional work to use RUE within certain pathosystems. 1
CHAPTER 1. GENERAL INTRODUCTION
Dissertation Format
This Dissertation is organized into four chapters. Chapter one is a literature review of frogeye leaf spot, radiation use efficiency (RUE), and application of unmanned aerial vehicles
(UAVs) in agriculture. Chapter two is an experiment carried out to demonstrate soybean RUE estimation on a small plot scale partially derived from UAV imagery. Chapter three investigates how foliar fungicides affect frogeye leaf spot, remotely sensed plant health indicators, and soybean yield and to compare the impact of foliar fungicides and frogeye leaf spot on RUE across foliar fungicide treatments. Chapter four is a general conclusion of the findings.
General Introduction
Soybean (Glycine max (L.) Merr.) is one of the most important and highest selling commodity crops in the U.S. and around the world. It contains a relatively high level of oil and protein compared to other legumes. Due to these compositional characteristics, soybeans are essential for human consumption, animal feed, and industrial components such as biodiesel. As a result of the increasing population, the need for soybean and soy products will grow. However, soybean yield is constantly reduced by pests and disease. Soybean foliar diseases alone cause an estimated annual economic loss of nearly 4.5 billion dollars in the U.S.
Frogeye leaf spot caused by the pathogen Cercospora sojina K. Hara, is a foliar disease of soybean reported to cause significant yield loss. More prevalent in the southern U.S., it is becoming increasingly more impactful in the northern U.S. due to warmer temperatures associated with climate change. Management of frogeye leaf spot includes fungicide application; however, resistant fungal isolates are seen across the soybean growing states. 2
Determining the impact of foliar diseases on yield is difficult. Historically, disease severity collected at multiple instances has been plotted to derive the area under the disease progress curve (AUDPC). A larger AUDPC value will be associated with less yield. This is inconsistent due to the fact that disease severity and yield are not directly related. Another measure to estimate disease impact on yield is radiation use efficiency (RUE). RUE is a measure of photosynthetic efficiency that correlates the amount of dry mass accumulated to absorbed radiation. Foliar disease severity must be removed from the green leaf area index (GLAI) used for estimating RUE. This parameter is directly related to yield, but it is difficult to collect on a small plot scale. However, the increased availability of unmanned aerial vehicles (UAVs) has provided researchers with a method to collect spatial and temporal data at high resolution and frequency.
RUE is theoretically a more appropriate predictor of the impact of foliar disease on yield.
However, RUE of soybean has yet to be partially estimated using consumer grade UAV imagery.
The goal of this dissertation is to more accurately capture the impact of foliar disease on soybean yield. The estimation of RUE from UAV reflectance data will be used to evaluate the effect fungicides and frogeye leaf spot have on soybean.
Literature Review
Soybeans
Arguably the most versatile row crop, soybean is one of the most important in the world
(Bandera et al., 2020; Savary et al., 2019). Worldwide, it is estimated to be grown on 6% of arable land and has been increasing more than any other crop since the 1970s (Hartman et al.,
2011). Soybean is comprised of approximately 40% protein and 20% oil (Clemente and Cahoon
2009). The majority of soybeans are crushed or processed into soybean oil for consumption and industrial products and meal for animal feed due to the relatively high protein content (Ali, 2010; 3
Clemente and Cahoon, 2009; Raghuvanshi and Bisht, 2010). Industrial uses of soybean oil include plastics, cosmetics, inks, solvents, and biodiesel (Mofijur et al., 2013; Raghuvanshi and
Bisht, 2010).
Diseases of Soybean
Soybeans are impacted by many diseases and pests that result in yield losses and quality reductions (Allen et al., 2017; Bandara et al., 2020). Bandara et al. (2020) reported that from
1996 to 2016, soybean diseases resulted in a total of $95.48 billion in losses. During this timeframe, foliar diseases of soybean caused an average loss of approximately 1.4 million metric tons and $435, 956,122 annually (Crop Protection Network, 2020). Major pathogens of soybean come from the fungal, bacterial, oomycete, and nematode domains (Roth et al., 2020).
Disease severity and economic loss may vary across the regions. Bandara et al. (2020) reported that disease cost the northern U.S. $80.89 billion compared to the southern U.S. loss of
$14.59 billion. Pathogens may be found across soybean growing regions or restricted because of environmental conditions (Roth et al., 2020). Pathogens that may be found throughout the U.S. include soybean cyst nematode, charcoal rot and seedling diseases (Roth et al., 2020).
Changing climate patterns and the increased amounts of precipitation and warming temperatures will make environments additionally conducive for certain pathogens and is expected to increase the risk of disease (Hatfield et al. 2011; Karl et al., 2009; Prein et al., 2017;
Tebaldi et al., 2006). Roth et al. (2020) suggests that increasing temperatures will support the overwintering survival of pathogens such as charcoal rot and frogeye leaf spot in the north.
Frogeye Leaf Spot
Frogeye leaf spot caused by the fungal ascomycete pathogen Cercospora sojina is a common polycyclic foliar disease of soybean with the potential to cause yield loss of up to 60%
(Dashiell and Akem, 1991). It was first discovered in Japan in 1915 and may now be found 4 worldwide infecting soybeans (Akem et al., 1992; Athow and Probst, 1952; Bernaux, 1979;
Lehman, 1928; Ma, 1994). C. sojina is capable of infecting the soybean plant at any stage of development at any time during the season. Young expanding leaves are most vulnerable to infection, with susceptibility diminishing as leaves mature (Mian et al., 2008).
Frogeye leaf spot displays characteristic foliar lesions, making it relatively easy to identify. The foliar lesions begin as small water-soaked spots, developing into brown spots, surrounded by a dark reddish-brown margin (Mian et al., 2008). These angular/circular lesions range from 1 to 5 mm in diameter (Grau et al., 2004). As severity increases, lesions may fuse, coalescing into an irregularly shaped spot, and appearing like blight (Phillips, 1999). Frogeye leaf spot may cause premature defoliation with severity greater than 30% (Mian et al., 2008). C. sojina is also able to infect soybean pods and stems (Sinclair and Backman, 1989). Pod and stem lesions are not as easily distinguishable as foliar lesions. Lesions on the stem begin as reddish and darken as they mature (Westphal et al., 2006). Lesions on the pods are reddish-brown, slightly depressed circular/oblong (Westphal et al., 2006). Lesions on both the stems and pods lack the characteristic reddish-brown margin that surrounds the foliar lesions. The fungus also has the ability to infect soybean seeds through the pod walls (Phillips, 1999). Infected seeds may display areas of light to dark gray spots, with seed coats that may crack or flake (Bisht and
Sinclair, 1985; Mian et al., 2008).
The pathogen overwinters in infested crop residue which remains on the soil surface and infected seed (Sherwin and Kreitlow, 1952). It was previously thought that C. sojina was incapable of surviving the cold winters of the north-central United States (Walker et al., 1994).
However, Yang et al. (2001) has reported the increase of frogeye leaf spot incidence in Iowa. In addition to Iowa, Ohio, and Wisconsin have also seen an increase in frogeye leaf spot (Cruz and 5
Dorrance, 2009; Mengistu et al., 2002). The increase of frogeye leaf spot may be due in part by the adoption of no-till farming practices and milder winters (Cruz and Dorrance, 2009; Dorrance et al., 2010; Roth et al., 2020; Wrather and Koenning, 2006; Yang et al., 2001).
Warm and humid environmental conditions are conducive to frogeye leaf spot.
Temperatures in the range of 25-30°C combined with humidity levels >90% may lead C. sojina to sporulate profusely (Mian et al., 2008). Long, hyaline conidia are the inoculum produced on infected soybean residue and infested seed (Mian et al., 2008). With appropriate temperatures and leaf wetness, conidia may germinate within an hour (Mian et al., 2008). Frequent rainfalls increase humidity and also help spread the disease. The conidia may be dispersed by rain-splash and wind facilitating secondary infections. This may create a layered effect of disease within the soybean canopy, with layers of diseased foliage telling of previously conducive conditions
(Mengistu et al., 2014).
Yield loss due to this disease is attributed to the reduction of green leaf area and premature defoliation that may occur (Dashiell and Akem, 1991). This reduction in photosynthetic leaf area may result in a shriveled grain and reduced seed weight, lowering yield
(Dashiell and Akem, 1991). Yield loss has been shown to be the greatest when infection occurs during vegetative or flowering (R1-R3) growth stages (Dashiell and Akem, 1991; Mian et al.,
2008). In the U.S., yield reductions due to frogeye leaf spot averaged 290,000 metric tons annually for 1996 to 2019 and an average reduction of $97,805,451 (Crop Protection Network,
2020). In the northern U.S., frogeye leaf spot caused an average yield loss of 150,000 metric tons and $51,944,385 annually during that same time frame, while the southern U.S. had an average yield reduction of 140,000 metric tons and $45,861,066 annually (Crop Protection Network, 6
2020). In Iowa during 2018 and 2019, frogeye leaf spot caused a loss of 368,000 and 146,000 metric tons, and $119,496,208 and $46,574,466, respectively (Crop Protection Network, 2020).
Management of frogeye leaf spot includes planting resistant cultivars and applications of fungicides, specifically quinone outside inhibitors (QoI) (Cruz and Dorrance, 2009; Dorrance et al., 2010; Mengistu et al., 2014). However, QoI resistant C. sojina populations have been reported in most of the soybean growing regions (Zhang et. al. 2018). QoI resistant isolates were first collected from Tennessee in 2010 (Zhang et al., 2012). Zhang et al. (2018) reported that C. sojina isolates collected across Alabama, Arkansas, Delaware, Illinois, Indiana, Iowa, Kentucky,
Louisiana, Mississippi, Missouri, North Carolina, Ohio, Tennessee, and Virginia had QoI fungicide resistance. Mathew et al. (2019) has also reported QoI resistant isolates in South
Dakota.
Fungicide Use
Applying fungicides to reduce foliar diseases has become an integral component to crop production systems (Gossen et al., 2014). Foliar fungicide use has increased in field crops because of the need to manage soybean rust, the obligate fungal parasite Phakopsora pachyrhizi
Sydow (Mueller et al., 2013; Schneider et al., 2005; Wise and Mueller, 2011). Fungicides are the only effective management strategies for soybean rust (Koch et al., 2010). In addition to disease management, certain fungicides have been used to provide additional plant health benefits (Wise and Mueller, 2011) even in the absence of foliar diseases. Applications in the absence of disease or based on risk of disease are prophylactic applications.
Pyrazole-carboxamide fungicides are succinate dehydrogenase inhibitors (SDHI) and are grouped in FRAC Code Group 7. This group acts by binding to ubiquinone-binding site (Q-site) of the mitochondrial complex II and inhibits fungal respiration (Avenot and Michailides, 2010).
These fungicides may have a protectant, translaminar, or systemic activity depending on the host 7 and pathogen (McKay et al., 2011). Newer generation of SDHIs, which includes fluxapyroxad, have activity against a broad range of fungal pathogens (McKay et al., 2011). These fungicides have a medium to high risk of resistance (McKay et al., 2011).
Demethylation inhibitors (DMI), including triazole fungicides, are grouped in FRAC
Code Group 3. This is the largest class of fungicides and may be applied pre- and post-infection
(Monton and Staub, 2008; Mueller et al., 2013). These fungicides have a single site mode of action by inhibiting sterol production (Mueller et al., 2013). Sterols are a key component to membrane structure and function. Due to the single mode of action and multiple applications during a season, resistance development is an issue (Klittich, 2008). Resistance to triazole fungicides was first identified in powdery mildews (Kuck et al., 1987). The use of mixtures has helped maintain successful activity against most fungal pathogens (Klittich, 2008).
QoI fungicides are commercially available products that were derived from the analogues of the natural compound produced by Strobilurus, a genus of Basidomycete (Balba, 2007). These fungicides have a single site of action (Mueller et al., 2013). These compounds include the strobilurin fungicides. A part of FRAC Code Group 11, QoI fungicides inhibit mitochondrial respiration of the fungus by obstructing electron transfer (Leroux, 1996). These fungicides include pyraclostrobin and fluoxastrobin. This single site of action makes this group susceptible to resistance development in fungal pathogens. Continuous application of fungicides exerts a selective pressure on the pathogen’s population, selecting for those that are resistant, and encourages their proliferation (Miles et al., 2012).
In the absence of disease, QoI fungicides have been promoted to increase yield and improvements in yield have been shown, though inconsistent (Bradley and Sweets, 2008; Kandel et al. 2016; Kyveryga et al., 2013; Wise and Mueller, 2011). This yield increase may be 8 attributed to a phenomenon called the “greening effect”, which is described to extend the duration of green leaf area, maintaining photosynthetic efficiency, and allowing continuation of dry matter accumulation (Balba, 2007; Bartlett et al., 2002; Kumudini et al., 2001; Morrison et al., 1999). Joshi et al. (2014) found that application of pyraclostrobin 10 and 20 days after emergence benefited the formation of root nodules. Subsequently, nitrogen fixation was enhanced resulting in improved growth and yield attribute benefits (Joshi et al., 2014). A review of the physiological effects of QoI fungicides concluded that the benefits are due to an increase in net photosynthesis (Amaro et al., 2019). In addition, the applications delay senescence by impacting hormonal balance (Amaro et al., 2019). However, Phillips et al. (2017) demonstrated that QoI fungicides may result in an undesirable increase of green stem disorder (GSD), as defined by Harbach et al. (2016), under certain conditions.
Disease Severity Assessment and Relation to Yield Loss
The amount of disease of a host population is known as the disease intensity (Nutter et al., 2006). Disease intensity is commonly measured in three ways: prevalence, incidence, and severity (Nutter et al., 2006). Prevalence and incidence are used interchangeably, but are not the same thing (Nutter et al., 2006). Prevalence refers to the macro geographical sampling units
(farms, counties, states, etc.) that the pathogen or disease has been detected, divided by the total sampled (Campbell and Madden,1990; Nutter et al., 1991; Nutter and Gaunt, 1996; Zadoks and
Schein, 1979). In contrast, disease incidence is the number of units on the plant scale (leaves, pods, stems, etc.) that display disease and divided by the total sample number (Nutter et al.,
1991; Nutter, 1997, 1999). Disease severity is defined as the amount of disease symptoms displayed on the observed sampling unit (Nutter et al., 1991; Nutter, 1997). These three assessment measurements may be used to monitor epidemics and develop yield loss models
(Guan and Nutter, 2001, 2004). 9
Area Under Disease Progress Curve
The area under the disease progress curve (AUDPC) has been historically used by plant pathologists to examine the progress of disease severity over time and can be used to determine the impact of disease on yield (Filho et al., 1997; Kumudini et al., 2008; Madden and Nutter,
1995). The AUDPC derived from severity ratings across multiple dates may be calculated using trapezoidal integration (Campbell and Madden, 1990; Madden et al., 2007; Shaner and Finney,
1977). However, the relationship between the AUDPC and yield is inconsistent (Carretero et al.,
2010; Serrago et al., 2009; Waggoner and Berger, 1987). Yield has more consistently been predicted by healthy leaf area duration and the absorption of solar radiation (Bryson et al., 1997;
Leite et al., 2006; Waggoner and Berger, 1987).
Radiation Use Efficiency (RUE)
Radiation use efficiency (RUE) is simply defined as the amount of dry matter accumulated per unit of photosynthetically active radiation absorbed by green tissues of a plant.
RUE is based on the relationship between the accumulation of biomass and the time integral of intercepted light (Warren-Wilson, 1967). RUE is synonymous with light use efficiency (LUE). It is a measure of photosynthetic efficiency that relates the rate at which dry mass is produced per amount of intercepted photosynthetically active radiation (PAR) (Ceotto and Castelli, 2002;
Monteith, 1977; Muchow et al., 1993). RUE is used as a parameter to determine the impact of different management inputs, the efficiency of resource capture, and the potential for yield improvements (Monteith, 1977; Muchow et al., 1993; Sinclair and Muchow, 1999; Stöckle and
Kemanian, 2009).
There are multiple ways to calculate RUE. Gitelson and Gamon (2015) discuss these definitions and identify three widely used methods of determining RUE: 1) using the total amount of incoming PAR assuming it is all absorbed; 2) incorporating the fraction of absorbed 10 photosynthetically radiation (fAPAR); and 3) using the fAPAR on only the green leaf tissue to determine the total absorbed light. Method (3) uses the incoming photosynthetically absorbed radiation (PARinc) and only considers the absorbed photosynthetically active radiation (APAR) by the green, photosynthetically active tissue (APARgreen) (Gitelson and Gamon, 2015). This estimation method is termed RUEgreen. Gitelson and Gamon (2015) suggest making the RUEgreen method of RUE estimation the standard for consistency. Since only the green leaf area is considered, RUEgreen becomes sensitive to pigment changes during senescence (Gitelson et al.,
2015). Madden and Nutter (1995) also postulated that using the healthy leaf area and RUE could quantify the effects of disease on yield while the portion of disease-free area may be multiplied by the LAI. Tewes and Schellberg (2018) used the RUEgreen method to estimate the RUEgreen of corn (Zea mays L.) for the first-time by utilizing consumer grade imagery from an unmanned aerial vehicle (UAV) to obtain reflectance data. However, foliar disease was not accounted for in this instance.
Components of RUEgreen
RUEgreen is determined from three main factors which are the dry mass, fraction of absorbed photosynthetically active radiation of the green tissue (fAPARgreen), and the incoming photosynthetically active radiation (PARinc).
Dry mass normally consists of the above ground portion of the plant (Sinclair and
Muchow, 1999). Different methods are used to estimate the dry mass per unit ground area. Plants are cut at the soil line then dried until a constant weight is maintained. Muchow et al. (1993) harvested all of the plants from a 1.73 and 2-m2 quadrat, and took a 10-plant subsample for dry weight. Maitree and Toyota (2017) sampled either three or four soybean plants for dry mass estimations from each plot. To convert the dry weight value into values of per unit ground area,
Maitree and Toyota (2017) used the number of soybean plants sampled and the planting density. 11
Hatfield (2014) counted the soybean plant population of a 1-m row before destructively sampling for dry mass measurements. Adeboye et al. (2016) destructively sampled soybean plants from an area of 0.179-m2.
Leaf area index (LAI) may be defined as the one-sided surface area measured over a horizontal ground surface area (Watson, 1947). LAI is a unitless quantity and serves as the interface of energy conversion. It is essential for biogeochemical cycles, driving radiation extinction, carbon gas exchange, and influencing the below-canopy microclimate (Bréda, 2003).
LAI estimation methods may be either direct or indirect. LAI estimation is difficult to measure as a result of spatial and temporal variability (Bréda, 2003). Direct methods involve the physical collection of tissue of an area to be measured. Determining the LAI of crops is commonly done by destructive sampling and measuring all of the leaves in a measured area (Bréda, 2003). This method is time consuming and difficult to apply in large areas (Bréda, 2003). Indirect methods utilize the radiative transfer theory to measure transmission through the canopy. The indirect method relies on a statistical approach to determine the distribution and arrangement of foliage within the canopy (Bréda, 2003; Jones, 1992; Ross, 1981).
The LAI of soybean follows a logistic pattern until senescence begins. That is, a lag in
LAI early in the season after emergence, followed by a sharp increase until the maximum LAI is reached, then declining once senescence begins, and resulting in leaf drop (Setiyono et al., 2008).
Modern day soybean cultivars reach a LAI between 6 and 7 (Dermody et al., 2006; Tagliapietra et al., 2018). Muchow et al. (1993) harvested all of the plants from a 1.73 and 2-m2 quadrat, pending experiment location and took a subsample of 5-10 plants to determine leaf area. Maitree and Toyota (2017) sampled either three or four plants for leaf area estimations of each plot. To translate leaf aera to LAI m2 m-2, Maitree and Toyota (2017) used the number of plants sampled 12 and the planting density. Hatfield (2014) indirectly measured soybean LAI by using a LI-2200 leaf area meter (LI-COR Biosciences, Lincoln, NE, USA), taking a single reading above and five readings below the soybean canopy, between rows at two locations within each plot. Adeboye et al. (2016) utilized an AccuPAR LP 80 (Decagon Devices, Inc., WA, USA) to indirectly measure soybean LAI.
The loss of leaf area due to weather events, insect damage, and other herbivory events has been shown to have a small, if any, impact on yield (Conley et al., 2008; Conley et al., 2009;
Haile et al., 1998). This highlights that soybeans are overinvesting in leaves (Dermody et al.,
2006). Kumudini et al. (2008) found that infection of soybean rust (SBR) caused by the fungus
Phakopsora pachyrhizi inhibits the canopy from reaching maximum LAI when infection begins at R2 (Full flowering) (Fehr and Caviness, 1977). However, infection of SBR at R5 (beginning seed) (Fehr and Caviness, 1977) does not affect the canopy from reaching maximum LAI
(Kumudini et al., 2008).
Measuring the LAI by the previously discussed methods is insensitive to disease, especially foliar diseases such as frogeye leaf spot. Lesions on the leaves caused by the pathogens are not contributing to photosynthesis, however they would be measured by a leaf area meter. Boote et al. (1983) classified pests into seven groups; stand reducers, photosynthetic rate reducers, leaf senescence accelerators, light stealers, assimilate sappers, tissue consumers, and turgor reducers. Any pathogen that causes foliar lesions, falls into the ‘light stealer’ group (Boote et al., 1983). The lesions absorb photosynthetically active radiation that would otherwise be absorbed by functioning leaf area (Boote et al., 1983).
Accounting for the diseased area is important when determining the impact of the pathogen on the effective green LAI and use in subsequent calculations. However, accounting 13 for the foliar disease severity alone may be underestimating the impact the disease has on non- lesion green leaf area. Kumudini et al. (2008) found that SBR reduced the photosynthetic efficiency of lesion free leaf tissue. Similar results of disease infected leaves having reduced photosynthesis of non-lesion area have been previously reported (Bastiaans, 1991).
Determining fAPARgreen involves estimating the incoming PAR (PARinc), accounting for the radiation passing through the canopy (fPARtrans), radiation being reflected by the soil
(fPARsoil), and the radiation reflected by the canopy (fPARout) to determine the total fraction of absorbed photosynthetically active radiation (fPARtotal) (Gitelson et al., 2015; Lindquist et al.,
2005; Tewes and Schellberg, 2018; Viña and Gitelson 2005). The total is then multiplied by the fraction of green leaf tissue by dividing the green leaf area index by the total leaf area index
(Gitelson and Gamon, 2015).
Incoming radiation is commonly measured by weather stations. It is estimated that half of the incoming radiation is useable PARinc (Taiz and Zeiger, 2014). fPARtrans may be estimated using the Lambert-Beer law as demonstrated by Tewes and Schellberg (2018).
The extinction coefficient (k) of the Lambert-Beer law is a constant. This is defined by
Monteith (1975) as the area of shadow cast by the canopy and divided by the total leaf area of the canopy. It has been related to row spacing by Flénet et al. (1996) for soybean. Row spacing of 0.35-m for soybean was found to have a mean k value of 0.523 (Flénet et al., 1996). The k values for sunflower (Helianthus annuus L.), soybean, sorghum (Sorghum bicolor (L.) Moench), and corn show a linear decrease as the row spacing increases (Flénet et al. 1996).
The fPARsoil may be absorbed by the foliage above. The moisture level of the soil impacts the amount of radiation reflected (Lobell and Asner, 2002). The relationship between soil moisture and reflectance was explained when moisture was expressed on a volumetric basis 14 by an exponential model with wet soils reflecting less than dry (Lobell and Asner, 2002). fPARsoil has also been excluded from the equation due to the assumption that there is little influence from fPARsoil on APAR (Lindquist et al., 2005; Tollenaar and Aguilera, 1992).
Measuring the amount of PAR transmitted through the canopy is commonly done using sensors placed under the canopy. Muchow et al. (1993) placed a tube solarimeter diagonally across the inner rows of each plot. de Souza et al. (2009) utilized a line quantum sensor (LI-
191SA, LI-COR Biosciences, Lincoln, NE, USA) installed diagonally across rows. The line sensor was moved around a quantum sensor (LI-190SA, LI-COR Biosciences, Lincoln, NE,
USA) that was placed in the field to measure incident PAR (de Souza et al., 2009). Handheld sensors such as the SunScan Canopy Analysis System (Delta-T Devices, Cambridge, UK) have been used by Hatfield (2014) and Adeboye et al. (2016).
RUE of Soybeans
There are three photosynthetic pathways of terrestrial plants: C3, C4, and crassulacean acid metabolism (CAM). C3 and C4 plants are named for the first photosynthetic product that is produced (Ehleringer and Cerling, 2002). Both C3 and C4 utilize the ribulose bisphosphate
(RuBP) carboxylase-oxygenase (Rubisco) enzyme. However, C4 plants have unique physiological properties and biochemical modifications that increases the efficiency of Rubisco.
Soybean is a C3 plant whereas corn is a C4.
RUE values of C3 crops tend to be less than C4 (Sinclair and Muchow, 1999). Within C3 crops, the RUE of soybean generally tends to be less than sunflower, rice (Oryza sativa L.), barley (Hordeum vulgare L.), wheat (Triticum aestivum L.), and potato (Solanum tuberosum L.)
(Sinclair and Muchow, 1999). RUE values of soybean have ranged from 0.60 to 2.53 (g Mj-1)
(Muchow, 1985; Santos et al., 2003). Sinclair and Muchow (1999) reason that the lower values 15 of soybean RUE are due to the high energy content of the seed and the large cost of nitrogen fixation.
A general decline of RUE during pod-filling was observed by Littelton et al. (1979) and
Muchow (1985). However, Muchow et al. (1993) found that the slope of the relationship relating to RUE of soybean was linear. RUE was found to decline prior to maturity, but remains constant through growth (Muchow et al., 1993). The decrease in RUE was attributed to a combination of a loss of biomass due to leaf drop and a decline in specific leaf nitrogen (Muchow et al., 1993).
Hatfield (2014) associated the variation in RUE values across previous experiments with sufficient nutrients and water to alleviate stress and nonlimiting soil conditions.
UAVs in Agriculture
Unmanned aerial systems (UASs), also known as drones, and unmanned aerial vehicles
(UAVs) are increasingly being used in both agricultural operations and research (Barbedo and
Koenigkan, 2018; Beloev, 2016). In the agricultural setting, they are commonly used as platforms to house sensors for capturing imagery, providing the ability to capture the imagery at a high temporal and spatial scale. They have become appropriate tools to monitor and access water status, plant vigor, biomass, and diseases (Adão et al., 2017; Bendig et al., 2015; Calderón et al., 2015; Park et al., 2015; Primicerio et al., 2012).
By utilizing UAVs, farmers are able to obtain images of hundreds of acres in just one flight (Beloev, 2016). Through image processing software, multiple images may be joined together into a single map (Beloev, 2016). These whole field maps provide the farmer a view of their entire field to identify locations for scouting. This may help identify regions in the field that may be missed by traditional unassisted scouting and reduces the overall effort.
Agricultural applications performed by utilizing UAV technology include stress detection, biomass and canopy cover estimation, yield prediction, and vegetation classification 16
(Chianucci et al., 2016; Cruzan et al., 2016; Han et al., 2018; Ishida et al., 2018; Kefauver et al.,
2017; Matese and Di Gennaro, 2018; Shi et al., 2016; Yu et al., 2016; Zheng et al., 2018; Zhou et al., 2017). Due to the impact of early stress detection, this application has received a significant amount of attention (Barbedo, 2019). RGB and multispectral imagery are the most utilized forms to collect disease information (Altas et al., 2018; Calderón et al., 2014; Kerkech et al., 2018;
Sugiura et al., 2018; Tetila et al., 2017; Zhang et al., 2017).
Disease rating by human eye takes lots of time and poses inherent errors associated with intra and inter-rater repeatability. Disease ratings may also suffer at the end of the day due to raters becoming physically exhausted. Error may increase as the day progresses as well as other psychological and cognitive phenomena, leading to bias (Barbedo, 2019). However, UAV imagery contains its own issues that need to be addressed to adequately account for and minimize error (Barbedo, 2019). These issues include separating plant from background, illumination, covariate shift, disease symptom variations, and other stresses or disorders
(Barbedo, 2016; Barbedo, 2019).
Passive and Active Sensors
Sensors carried on UAVs and in general may be grouped into two types, passive or active. Passive sensors rely on an external source of illumination. This is usually the sun, but may also be another source that is not directly associated with the sensor. In contrast, active sensors do not rely on external illumination, they emit their own. Passive sensors are mostly used for image capture, which makes the illumination conditions an important aspect to consider.
Light detection and ranging (LiDAR) is an example of an active sensor. LiDAR utilizes laser pulses to measure distances to the earth, which are used to generate digital elevation models
(DEMs). Other active sensors may be handheld tools such as a SPAD-502 meter (Konica-
Minolta, Japan) and a GreenSeeker (Trimble, Sunnyvale, CA, USA). 17
The Soil Plant Analysis Development (SPAD) meter is commonly used to measure crop nitrogen status (Yuan et al., 2016). The SPAD-502 meter is an active sensor that provides a non- destructive method of measuring relative leaf chlorophyll levels. The sensor is composed of two light emitting diodes (LEDs) and a photodiode receptor. The LEDs emit wavelengths of light in the red (650nm) and the infrared (940nm) regions of the electromagnetic spectrum. This device measures the transmittance of these wavebands through the leaf. The resulting values normally fall between 0.0 and 50.0, which is proportional to the amount of chlorophyll in the leaf (Uddling et al., 2007). SPAD values may be converted into units of chlorophyll using a calibration curve
(Markwell et al., 1995). Calibration curves have been shown to vary among plant species
(Castelli et al., 1996; Uddling et al., 2007). Raw SPAD values have been used to determine the efficacy of nitrogen treatments (Fox et al., 1994; Kaakeh et al., 1992; Piekielek and Fox, 1992;
Turner and Jund, 1991; Wood et al., 1992).
The handheld GreenSeeker is another example of an active sensor. The GreenSeeker emits bursts of red and near-infrared (NIR) light. The reading frame emanates from the sensor in an oval patten, perpendicular to the direction of the handle. The diameter of the oval increases as the height above the canopy the sensor is held increases. The reflected light from the red and
NIR wavelengths are captured and used to determine the normalized difference vegetation index
(NDVI) value. This value is a direct indication of the plant’s health. Greater NDVI values indicate that there is more green leaf tissue than compared to low values.
Accurately separating the features of interest from their background is important for minimizing the error of the derived data from imagery. Methods of segmentation include supervised classifications, unsupervised classifications, and object-based image analysis. Within each method of classification, there are multiple techniques which vary on how each pixel is 18 assigned a class. Examples of techniques within the supervised classification method include maximum likelihood, principal components, and random forest. Each method and their associated techniques have advantages pending the specifics of the inputs and how the user would prefer the analysis to proceed.
Radiometric Calibration
Passive sensors rely on an external source of illumination, in the case of UAV imagery this is frequently the sun. Thus, the data captured by the sensor may be influenced by surface conditions, the suns geometry, and atmospheric and topographic effects (Berra et al., 2017;
Kelcey and Lucieer, 2012). Calibration of the digital numbers (DNs) accounts for differences between dates.
Digital numbers should be radiometrically calibrated into reflectance to have lasting value (Smith and Milton, 1999). To convert from DN to a physical value of reflectance, an empirical relationship is used. The empirical line method utilizes targets of known radiance values and compares those to their DNs. A line of best fit is used to develop a prediction equation. This equation may then be used to convert the DNs of the entire image to reflectance.
Each individual band will require a predictive equation to covert the DN to reflectance.
Targets may consist of areas or objects which are smooth and consistent, with basic
Lambertian properties. Snow and roads are examples of reasonable targets; however, both may contain impurities that result in varying levels of reflection. A single target may be used with assumptions that a linear relationship exists and that surfaces with zero reflectance produce zero radiance (Smith and Milton, 1999). This was shown to result in large errors of up to 20%
(Freemantle et al., 1992; McArdle et al., 1992). To reduce the error and eliminate the need of a linear assumption, Farrand et al. (1994) and Price et al. (1995) used four calibration targets. The errors were reduced to a few percent when using the additional calibration targets (Smith and 19
Milton 1999). Although the reflectance of Lambertian surfaces used for calibration are relatively constant, the DNs of the captured image may fluctuate. DNs of highly reflective features may increase under full illumination and be less under cloudy conditions, just as they may change due to the suns position in the sky (Wang and Myint, 2015)
The assumption of a linear relationship may cause problems and has also been found to be untrue by Wang and Myint (2015) when calibrating UAV imagery. Wang and Myint (2015) found that pixel DN increased more slowly than reflectance, resulting with an exponential equation being the best fit for calibration. By using a logarithm transformation, the exponential equation may be converted to a linear equation. Wang and Myint (2015) showed that the raw image DN has a linear relationship with the log-transformed surface reflectance value.
Indices
An index is a combination of reflections at different frequencies. Vegetation Indices
(VIs), such as the normalized difference vegetation index (NDVI), have been used in research as key parameters to discern crop types, separate living plant material from non-plant, monitor health, and estimate yield (Bolton and Friedl, 2013; Devadas et al., 2009; Sonobe et al., 2018;
Woebbecke et al., 1995; Yoem et al., 2019). The whole field orthomosaics may be transformed to display the specific index.
Vegetation indices derived from only the red, green, and blue bands has been shown to accurately delineate green vegetation from other material such as soil and residue (Meyer and
Neto, 2008). These indices include the normalized difference index (NDI), green-red vegetation index (GRVI), excess green (ExG), excess red (ExR), and the normalized green red difference index (NGRDI) (Hunt et al., 2005; Meyer et al., 1998; Motohka et al., 2010; Perez et al., 2000;
Woebbecke et al., 2005). The ExG index is widely used and has been shown to accurately separate green leaf area from the background (Gitelson et al., 2002; Lamm et al., 2002; Mao et 20 al., 2003; Ponti, 2013; Zhou et al., 2017). Woebbecke et al. (2005) found that the Exg index resulted in a binary separation between vegetation and soil background once a threshold value was identified. This binary image separating foliage and background may then be used as a mask layer to delineate the regions of the initial image to be extracted for further utilization (Camargo
Neto et al., 2006).
Literature Cited
Adão, T., Hruška, J., Pádua, L., Bessa, J., Peres, E., Morais, R., and Sousa, J.J. 2017. Hyperspectral imaging: A review on UAV-based sensors, data processing and applications for agriculture and forestry. Remote Sens. 9:1110.
Adeboye, O.B., Schultz, B., Adekalu, K.O., and Prasad, K. 2016. Impact of water stress on radiation interception and radiation use efficiency of soybeans (Glycine max L. Merr.) in Nigeria. Braz. J. Sci. Tech. 3:1–21.
Ali, N. 2010. Soybean processing and utilization. In G. Singh (Ed.), The soybean: 345–374. CABI.
Akem, C.N., Dashiell, K.E., and Uwala, A.C. 1992. Prevalence of frogeye leaf spot of soybean in Nigeria. Int. J. Trop. Plant Dis. 10:181–183.
Allen, T.W., Bradley, C.A., Sisson, A.J., Byamukama, E., Chilvers, M.I., Coker, C.M., et al. 2017. Soybean yield loss estimates due to diseases in the United States and Ontario, Canada from 2010 to 2014. Plant Health Prog. 18:19–27.
Altas, Z., Ozguven, M.M., and Yanar, Y. 2018. Determination of sugar beet leaf spot disease level (Cercospora Beticola Sacc.) with image processing technique by using drone. Curr. Investig. Agric. Curr. Res. 5:621–631.
Amaro A.C.E., Baron D., Ono E.O., and Rodrigues J.D. 2019. Physiological effects of strobilurin and carboxamides on plants: an overview. Acta Physiol. Plant. 42:4 10.1007/s11738- 019-2991-x
Athow, K.L., and Probst, A.H. 1952. The inheritance of resistance to frogeye leaf spot of soybeans. Phytopathology. 42:660–662.
Avenot, H.F., and Michailides, T.J. 2010. Progress in understanding molecular mechanisms and evolution of resistance to succinate dehydrogenase inhibiting (SDHI) fungicides in phytopathogenic fungi. Crop Prot. 29:643-651.
Balba, H. 2007. Review of strobilurin fungicide chemicals. J. Environ. Sci. 42:441-451. 21
Bandara, A.Y., Weerasooriya, D.K., Bradley, C.A., Allen, T.W., and Esker, P.D. 2020. Dissecting the economic impact of soybean diseases in the United States over two decades. PLoS ONE. 15:e0231141.
Barbedo, J.G.A. 2016. A review on the main challenges in automatic plant disease identification based on visible range images. Biosyst. Eng. 144:52–60.
Barbedo, J.G.A. 2019. A review on the use of unmanned aerial vehicles and imaging sensors for monitoring and assessing plant stresses. Drones 3:40.
Barbedo, J.G.A., and Koenigkan, L.V. 2018. Perspectives on the use of unmanned aerial systems to monitor cattle. Outlook Agric. 47:214–222.
Bartlett, D.W., Clough, J.M., Godwin, J.R., Hall, A.A., Hamer, M., and Parr-Dobrzanski, B. 2002. The strobilurin fungicides. Pest Manag. Sci. 58:649-662.
Bastiaans, L. 1991. Ratio between virtual and visual lesion size as a measure to describe reduction in leaf photosynthesis of rice due to leaf blast. Phytopathology 81:611–615.
Beloev, I.H. 2016. A review on current and emerging application possibilities for unmanned aerial vehicles. Acta Technol. Agric. 19:70–76.
Bendig, J., Yu, K., Aasen, H., Bolten, A., Bennertz, S., Broscheit, J., Gnyp, M.L., and Bareth, G. 2015. Combining UAV-based plant height from crop surface models, visible, and near infrared vegetation indices for biomass monitoring in barley. Int. J. Appl. Earth Obs. Geoinf. 39:79–87.
Bernaux, P. 1979. Identification of some soybean diseases in Cameroon. Agronomy Trop. 34:301-304.
Berra, E.F., Gaulton, R., and Barr, S. 2017. Commercial off-the-shelf digital cameras on unmanned aerial vehicles for multitemporal monitoring of vegetation reflectance and ndvi. IEEE Trans. Geosci. Remote Sens. 55:4878–4886.
Bisht, V.S., and Sinclair, J.B. 1985. Effect of Cercospora sojina and Phomopsis sojae alone or in combination on seed quality and yield of soybeans. Plant Dis. 69:436–439.
Bolton, D.K., and Friedl, M.A. 2013. Forecasting crop yield using remotely sensed vegetation indices and crop phenology metrics. Agricultural and Forest Meteorology. 173:74–84.
Boote, K.J., Jones, J.W., Mishoe, J.W., and Berger, R.D. 1983. Coupling pests to crop growth simulators to predict yield reductions. Phytoputholoy 73:1581-7.
Bradley, K.W., and Sweets, L.E. 2008. Influence of glyphosate and fungicide coapplications on weed control, spray penetration, soybean response, and yield in glyphosate-resistant soybean. Agronomy. J. 100:1360-1365.
Bréda, N. 2003. Ground-based measurements of leaf area index: a review of methods, instruments and current controversies. Journal of Experimental Botany 54:2403–2417. 22
Bryson, R.J., Paveley, N.D., Clark, W.S., Sylvester-Bradley, R., and Scott, R.K. 1997. Use of in- field measurements of green leaf area and incident radiation to estimate the effects of yellow rust epidemics on the yield of winter wheat. Eur. J. Agronomy. 7:53–62.
Calderón, R., Montes-Borrego, M., Landa, B.B., Navas-Cortés, J.A., and Zarco-Tejada, P.J. 2014. Detection of downy mildew of opium poppy using high-resolution multi-spectral and thermal imagery acquired with an unmanned aerial vehicle. Precis. Agric. 15:639–661.
Calderón, R., Navas-Cortés, J.A., and Zarco-Tejada, P.J. 2015. Early detection and quantification of verticillium wilt in olive using hyperspectral and thermal imagery over large areas. Remote Sens. 7:5584–5610.
Camargo Neto, Meyer, G.E., Jones, D.D., and Samal, A.K. 2006. Plant species identification using elliptic Fourier analysis. Computers and Electronics in Agriculture 50:121–134.
Campbell, C.L., and Madden, L.V. 1990. Introduction to plant disease epidemiology. John Wiley, New York.
Carretero, R., Serrago, R.A., Bancal, M.O., Perelló, A.E., and Miralles, D.J. 2010. Absorbed radiation and radiation use efficiency as affected by foliar diseases in relation to their vertical position into the canopy in wheat. Field Crops Res. 116:184–195.
Castelli, F., Contillo, R., and Miiceli, F. 1996. Non-destructive determination of leaf chlorophyll content in four crop species. J. Agronomy Crop Sci. 177:275–283.
Ceotto, E., and Castelli, F. 2002. Radiation use efficiency in flue-cured tobacco (Nicotiana tabacum L.): response to nitrogen supply, climate variability and sink limitations. Field Crops Res. 74:117–130.
Chianucci, F., Disperati, L., Guzzi, D., Bianchini, D., Nardino, V., Lastri, C., Rindinella, A., and Corona, P. 2016. Estimation of canopy attributes in beech forests using true colour digital images from a small fixed-wing uav. Int. J. Appl. Earth Obs. Geoinf. 47:60–68.
Clemente, T.E., and Cahoon, E.B. 2009. Soybean oil: genetic approaches for modification of functionality and total content. Plant Phys. 151:1030–1040.
Conley, S.P., Abendroth, L., Elmore, R., Christmas, E.P., and Zarnstorff, M. 2008. Soybean seed yield and composition response to stand reduction at vegetative and reproductive stages. Agronomy J. 100:1666–1669.
Conley, S.P., Pedersen, P., and Christmas, E.P. 2009. Main-stem node removal effect on soybean seed yield and composition. Agronomy J. 101:120–123.
Cruz, C.D., and Dorrance, A.E. 2009. Characterization and survival of Cercospora sojina in Ohio. Online. Plant Health Progress. doi:10.1094/PHP-2009-0512-03-RS. 23
Cruzan, M.B., Weinstein, B.G., Grasty, M.R., Kohrn, B.F., Hendrickson, E.C., Arredondo, T.M., and Thompson, P.G. 2016. Small unmanned aerial vehicles (micro-uavs, drones) in plant ecology. Appl. Plant Sci. 4.
Dashiell, K.E., and Akem, C.N. 1991. Yield losses in soybeans from frogeye leaf spot caused by Cercospora sojina. Crop Prot. 10:465–468. de Souza, P.J.D.O.P., Ribeiro, A., da Rocha, E.J.P., et al. 2009. Solar radiation use efficiency by soybean under field conditions in the Amazon region. Pesquisa Agropecuaria Brasileira. 44:1211–1218.
Dermody, O., Long, S.P., and DeLucia, E.H. 2006. How does elevated CO2 or ozone affect the leaf-area index of soybean when applied independently? New Phytologist 169:145–155.
Devadas, R., Lamb, D.W., Simpfendorfer, S., Backhouse, D. 2009. Evaluating ten spectral vegetation indices for identifying rust infection in individual wheat leaves. Precision Agric. 10:459–470.
Dorrance, A.E., Cruz, C., Mills, D., Bender, R., Koenig, M., LaBarge, G., Leeds, R., Mangione, D., McCluer, G., Ruhl, S., Siegrist, H., Sundermeier, A., Sonnenberg, D., Yost, J., Watters, H., and Wilson, G. 2010. Effects of foliar fungicide and insecticide applications on soybean in Ohio. Plant Health Progress doi:10.1094/PHP-2010-0122-01-RS
Ehleringer, J.R., and Cerling, T.E. 2002. in Encyclopedia of Global Environmental Change. (ed. Munn, T.) 2:186–190.
Farrand, W.H., Singer, R.B., and Merenyi, E. 1994. Retrieval of apparent surface reflectance from AVIRIS data-a comparison of empirical line, radiative-transfer and spectral mixture methods. Remote Sensing of Environment. 47:311-321.
Fehr, W.R., and Caviness, C.E. 1977. Stages of soybean development. Iowa State University Cooperative Extension Service, Special Report 80.
Filho, A.B., Carneiro, S.M.T.P.G., Godoy, C.V., Amorim, L., Berger, R.D., and Hau, B. 1997. Angular leaf spot of Phaseolus beans: relationships between disease, healthy leaf area, and yield. Phytopathology 87:506–15.
Flénet, F., Kiniry, J., Board, J., Westgate, M., and Reicosky, D.C. 1996. Row spacing effects on light extinction coefficients of corn, sorghum, soybean and sunflower. Agronomy. J. 88:185-190.
Fox, R.H., Piekielek, W.P., and Macneal, K.M. 1994. Using a chlorophyll meter to predict nitrogen fertilizer needs of winter wheat. Commun. Soil Sci. Plant Anal. 25:171-181.
Freemantle, J.R., Pu, R., and Miller, J.R. 1992. Calibration of imaging spectrometer data to reflectance using pseudo-invariant features. Proceedings of the 15th Canadian Symposium on Remote Sensing (Toronto: Canadian Remote Sensing Society and Canadian Aeronautics and Space Institute). 452-455. 24
Giltelson, A.A., Kaufman, Y.J., Stark, R., and Rundquist, D. 2002. Novel algorithm for remote estimation of vegetation fraction. Remote Sensing of Environment 80:76–87.
Gitelson, A.A., and Gamon, J.A. 2015. The need for a common basis for defining light-use efficiency: Implications for productivity estimation. Remote Sens. Environ. 156:196–201.
Gitelson, A.A., Peng, Y., Arkebauer, T.J., and Suyker, A.E. 2015. Productivity, absorbed photosynthetically active radiation, and light use efficiency in crops: Implications for remote sensing of crop primary production. J. Plant Physiol. 177:100–109.
Gossen, B.D., Carisse, O., Kawchuk, L., Van Der Heyden, H., and McDonald, M.R. 2014. Recent changes in fungicide use and fungicide insensitivity ofcrop pathogens in Canada, 2014. Can. J. Plant Pathol. 36:327-340.
Grau, C.R., Dorrance, A.E., Bond, J., and Russin, J.S. 2004. Fungal diseases. In H.R. Boerma and J.E. Specht (ed.) Soybeans: Improvement, production, and uses. 3rd ed. Agronomy Monogr. 16:679–763. ASA, CSSA, and SSSA, Madison, WI.
Guan, J., and Nutter, F.W.Jr. 2001. Factors affecting the quality and quantity of sunlight reflected from alfalfa canopies. Plant Dis. 85:865–874.
Guan, J., and Nutter, F.W.Jr. 2004. Comparison of single-point alfalfa yield models based on visual disease intensity and remote sensing assessments. Canadian Journal of Plant Pathol. 26:314–324.
Haile, F.J., Higley, L.G., Specht, J.E., and Spomer, S.M. 1998. Soybean leaf morphology and defoliation tolerance. Agronomy J. 90:353–362.
Han, L., Yang, G., Yang, H., Xu, B., Li, Z., and Yang, X. 2018. Clustering field-based maize phenotyping of plant-height growth and canopy spectral dynamics using a uav remote-sensing approach. Front. Plant Sci. 9:1638.
Hatfield, J.L., Boote, K.J., Kimball, B.A., Sizka, L.H., Izaurralde, R.C., Ort, D., Thomson, A.M., and Wolfe, D. 2011. Climate impacts on agriculture: implications for crop production. Agronomy J. 103: 351–370.
Hatfield, J.L. 2014. Radiation use efficiency: evaluation of cropping and management systems. Agronomy. J. 106:1820–1827. doi:10.2134/agronj2013.0310.
Harbach, C.J., Allen, T.W., Bowen, C.R., Davis, J.A., Hill, C. B., Leitman, M., Leonard, B.R., Mueller, D.S., Padgett, G.B., Phillips, X.A., Schneider, R.W., Sikora, E.J., Singh, A.K., and Hartman, G.L. 2016. Delayed scenescence in soybean: Terminology, research update, and survey results from growers. Plant Health Prog. 17:76-83.
Hartman, G.L., West, E.D., and Herman, T.K. 2011. Crops that feed the World 2. Soybean worldwide production, use, and constraints caused by pathogens and pests. Food Secur. 3:5–17. 25
Hunt, E.R.J., Cavigelli, M., Daughtry, C.S.T., McMurtrey, J.I., Walthall, C.L. 2005. Evaluation of digital photography from model aircraft for remote sensing of crop biomass and nitrogen status. Precision Agric. 6:359–378.
Ishida, T., Kurihara, J., Viray, F.A., Namuco, S.B., Paringit, E.C., Perez, G.J., Takahashi, Y., and Marciano, J.J. 2018. A novel approach for vegetation classification using uav-based hyperspectral imaging. Comput. Electron. Agric. 144:80–85.
Jones, H.G. 1992. Plant and microclimate, 2nd edn. Cambridge, UK: Cambridge University Press.
Joshi, J., Sharma, S., and Guruprasad, K.N. 2014. Foliar application of pyraclostrobin fungicide enhances the growth, rhizobial-nodule formation and nitrogenase activity in soybean (var. JS- 335). Pestic. Biochem. Physiol. 114:61–66.
Kaakeh, W., Pfeiffer, D.G., and Marini, R.P. 1992. Combined effects of spirea aphid (Homoptera: Aphididae) and nitrogen fertilization on net photosynthesis, total chlorophyll content, and greenness of apple leaves. J. Econ. Entomo. 185:939-946.
Kandel, Y.R., Mueller, D.S., Bestor, N.R.C., Bradley, C.A., Giesler, L.J., Wise, K.A. 2016. Analyses of yield and net economic response from foliar fungicide and insecticide applications to soybean in the North Central United States. Plant health Prog. 17:232-238 doi: 10.1094 / PHP- RS-16-0038.
Karl, T.R., Melillo, J.M., and Peterson, T.C. 2009. Global climate change impacts in the United States. Cambridge University Press, New York,NY.
Kefauver, S.C., Vicente, R., Vergara-Díaz, O., Fernandez-Gallego, J.A., Kerfal, S., Lopez, A., Melichar, J.P.E., Serret Molins, M.D., and Araus, J.L. 2017. Comparative uav and field phenotyping to assess yield and nitrogen use efficiency in hybrid and conventional barley. Front. Plant Sci. 8:1733.
Kerkech, M., Hafiane, A., and Canals, R. 2018. Deep leaning approach with colorimetric spaces and vegetation indices for vine diseases detection in uav images. Comput. Electron. Agric. 155:237–243.
Kelcey, J., and Lucieer, A. 2012. Sensor correction of a 6-band multispectral imaging sensor for uav remote sensing. Remote Sens. 4:1462-1493.
Klittich, C.J. 2008. Milestones in fungicide discovery: Chemistry that changed agriculture. Online. Plant Health Progress. doi:10.1094/PHP-2008-0418-01-RV.
Koch, K. A., Potter, B.D., and Ragsdale, D.W. 2010. Non-target impacts of soybean rust fungicides on the fungal entomopathogens of soybean aphid. J. Invertebrate Pathol. 103:156-164.
Kuck, K.H., Scheinpflug, H., and Pontzen, R. 1987. DMI fungicides. In: Modern selective fungicides, 2nd Edn. H. Lyr, ed. Gustav Fisher Verlag, Jena, Germany. 26
Kumudini, S., Hume, D.J., and Chu, G. 2001. Genetic improvement in short season soybeans: I. dry matter accumulation, partitioning, and leaf area duration. Crop Sci. 41:391-398.
Kumudini, S., Godoy, C.V., Board, J.E., Omielan, J., and Tollenaar, M. 2008. Mechanisms involved in soybean rust-induced yield reduction. Crop Sci. 48:2334–2342.
Kyveryga, P.M., Blackmer, T.M., and Mueller, D.S. 2013. When do foliar pyraclostrobin fungicide applications produce profitrable soybean yield responses? Online. Plant Health Prog. doi: 10.1094/PHP_2013-0928-01-RS.
Lamm, R.D., Slaughter, D.C., and Giles, D.K. 2002. Precision weed control for cotton. Transactions of the ASAE 45:231–238.
Lehman, S.G. 1928. Frog-eye leaf spot of soybean caused by Cercospora diazu Miura. J. Agric. Res. 35:811–833.
Leite, R.M.B.V.C., Amorim, L., and Filho, A.B. 2006. Relationships of disease and leaf area variables with yield in the Alternaria helianthi-sunflower pathosystem. Plant Pathol. 55:73–81.
Leroux, P. 1996. Recent developments in the mode of action of fungicides. Pestic. Sci. 47:191- 197.
Lindquist, J.L., Arkebauer, T.J., Walters, D.T., Cassman, K.G., and Dobermann, A. 2005. Maize radiation use efficiency under optimal growth conditions. Agronomy. J. 97:72–78.
Littleton, E.J., Dennett, M.D., Monteith, J.L., and Elston, J.1979. The growth and development of cowpeas (Vigna unguiculata) under tropical field conditions. 2. Accumulation and partition of dry weight. J. Agric. Sci. Camb. 93:309-320.
Lobell, D.B., and Asner, G.P. 2002. Moisture effects on soil reflectance. Soil Sci. Soc. Am. J. 66:722–727.
Ma, G.Z. 1994. Review and forecast of study on frogeye leaf spot. Soybean J. 1:6–7.
Madden, L.V., and Nutter, F.W. 1995. Modeling crop losses at the field scale. Can. J. Plant Pathol. 17:124–35.
Madden, L.V., Hughes, G., and van den Bosch, F., 2007. The Study of Plant Disease Epidemics. APS Press, St. Paul.
Maitree, L., and Toyota, M. 2017. A high seed yield and associated attributes of dry matter production achieved by recent Japanese soybean cultivars. Plant Production Science. doi: 10.1080/1343943X.2017.1294463
Mao, W., Wang, Y., and Wang, Y. 2003. Real-time detection of between-row weeds using machine vision. ASAE paper number 031004. The Society for Agricultural, Food, and Biological Systems, St. Joseph, MI. 27
Markwell, J., Osterman, J.C., and Mitchell, J.L. 1995. Calibration of the Minolta SPAD-502 leaf chlorophyll meter. Photosynthesis Research. 46:467-472.
Matese A., and Di Gennaro, S.F. 2018. Practical applications of a multisensor UAV platform based on multispectral, thermal and rgb high resolution images in precision viticulture. Agriculture. 8:116.
Mathew, F.M., Byamukama, E., Neves, D.L., and Bradley, C.A. 2019. Resistance to quinone outside inhibitor fungicides conferred by the G143A mutation in Cercospora sojina (causal agent of frogeye leaf spot) isolates from South Dakota soybean fields. Plant Health Prog. 20:104–105.
McArdle, S.S., Miller, J.R., and Freemantle, J.R. 1992. Airborne image acquisition under cloud: preliminary comparisons with clear-sky scene radiance and reflectance imagery. Proceedings of the 15th Canadian Symposium on Remote Sensing (Toronto:Canadian Remote Sensing Society and Canadian Aeronautics and Space Institute). 446-449.
McKay, A.H., Hagerty, G.C., Follas, G.B., Moore, M.S., Christie, M.S., and Beresford, R.M. 2011. Succinate dehydrogenase inhibitor (SDHI) fungicide resistance prevention strategy. N. Z. Plant Protec. 64:119–124.
Mengistu, A., Kurtzweil, N.C., and Grau, C.R. 2002. First report of Frogeye Leaf Spot (Cercospora sojina) in Wisconsin. Plant Dis. 11:1272.
Mengistu, A., Kelly, H.M., Bellaloui, N., Arelli, P.R., Reddy, K.N., and Wrather, A.J. 2014. Tillage, fungicide, and cultivar effects on frogeye leaf spot severity and yield in soybean. Plant Dis. 98:1476-1484.
Meyer, G.E., Hindman, T.W., and Lakshmi, K. 1998. Machine vision detection parameters for plant species identification. In: Meyer, G.E., DeShazer, J.A. (Eds.), Precision Agriculture and Biological Quality, Proceedings of SPIE. 3543:327–335.
Meyer, G.E., Neto, J.C. 2008. Verification of color vegetation indices for automated crop imaging applications. Comput. Electron. Agric. 63:282–293.
Mian, M.A.R., Phillips, D.V., Boerma, H.R., Missaoui, A.M., and Walker, D.R. 2008. Frogeye leaf spot of soybean: A review and proposed race designations for isolates of Cercospora sojina Hara. Crop Sci. 48:14–24.
Miles, L.A., Miles, T.D., Kirk, W.W., and Schilder, A.M.C. 2012. Strobilurin (QoI) resistance in populations of Erysiphe necator on grapes in Michigan. Plant Dis. 96:1621-1628.
Mofijur, M., Atabani, A.E., Masjuki, H.H., Kalam, M.A., and Masum, B.M. 2013. A study on the effects of promising edible and non-edible biodiesel feedstocks on engine performance and emissions production: a comparative evaluation. Renew. Sustain. Energy Reviews. 23: 391–404.
Monteith, J.L. 1975. Principles of Environmental Physics. Edward Arnold, London. 28
Monteith, J.L. 1977. Climate and the efficiency of crop production in Britian. Philos. Trans. R. Soc. London B. 281:277-294.
Monton, V., and Staub, T. 2008. A short history of fungicides. Online, APSnet Features. doi: 10.1094/APSnetFeature-2008-0308.
Morrison, M.J., Voldeng, H.D., and Cober, E.R. 1999. Physiological changes from 58 years of genetic improvement of short-season soybean cultivars in Canada. Agronomy. J. 91:685-689.
Motohka, T., Nasahara, K.N., Oguma, H., et al. 2010. Applicability of green-red vegetation index for remote sensing of vegetation phenology. Remote Sens. 2:2369-2387.
Muchow, R.C. 1985. Canopy development in grain legumes grown under different soil water regimes in a semi-arid tropical environment. Field Crop. Res. 11:99–109.
Muchow, R.C., Robertson, M.J., and Pengelly, B.C. 1993. Radiation-use efficiency of soybean, mugbean and cowpea under different environmental conditions. Field Crops Research 32:1-16.
Mueller, D.S., Wise, K.A., Dufault, N.S., Bradley, C.A., and Chilvers, M.I. 2013. Fungicides for field crops. American Phytopathological Society. ISBN:9780890544204
Nutter, F.W.Jr. 1997. Quantifying the temporal dynamics of plant virus epidemics: a review. Crop Protection 16:603–618.
Nutter, F.W.Jr. 1999. Understanding the interrelationship between botanical, human, and veterinary epidemiology: The Ys and Rs of it all. Ecosystem Health 5:131–140.
Nutter, F.W.Jr., Teng, P.S. and Shokes, R.M. 1991. Disease assessment terms and concepts. Plant Disease 75:1187–1188.
Nutter, F.W.Jr. and Gaunt, R.E. 1996. Recent developments in methods for assessing disease losses in forage/pasture crops. In Pasture and Forage Crop Pathology. 93–118. Madison, WI: ASA, CSSA, and SSSA.
Nutter, F.W.Jr., Esker, P.D. and Coelho, Netto, R.A. 2006. Disease assessment concepts in plant pathology. Euro. J. Plant Path. 115: 95–103.
Park, S., Nolan, A., Ryu, D., Fuentes, S., Hernandez, E., Chung, H., and O’Connell, M. 2015. Estimation of crop water stress in a nectarine orchard using high-resolution imagery from unmanned aerial vehicle (UAV). In Proceedings of the 21st International Congress on Modelling and Simulation, Gold Coast, Australia. 1413–1419.
Perez, A.J., Lopez, F., Benlloch, J.V., Christensen, S. 2000. Colour and shape analysis techniques for weed detection in cereal fields. Comput. Electron. Agric. 25:197–212.
Phillips, D.V. 1999. Frogeye leaf spot. In G.L. Hartman, Sinclair, J.B., and Rupe, J.C. (ed.) Compendium of soybean diseases. Reproduced from Crop Science. Published by Crop Science Society of America. 48:20–21. 2008 4th ed. American Phytopathological Soc., St. Paul, MN. 29
Phillips, X.A., Singh, A.K., Kandel, Y.R., and Mueller, D.S. 2017. Effect of pod removal, foliar fungicides, and cultivar on green stem disorder of soybean. Agronomy J. 109:2680-2688.
Piekielek, W.P., and Fox, R.H. 1992. Use of a chlorophyll meter to predict sidedress nitrogen requirements for maize. Agronomy. J. 84:59-65.
Ponti, M.P. 2013. Segmentation of low-cost remote sensing images combining vegetation indices and mean shift. In IEEE Geoscience and Remote Sensing Letters. 10:67-70.
Prein, A. F., Liu, C., Ikeda, K., Trier, S.B., Rasmussen, R.M., Holland, G.J., and Clark, M.P. 2017. Increased rainfall volume from future convective storms in the US. Nat. Clim. Change 7: 880–884.
Price, R., Anger, C.D., and Mah, S. 1995. Preliminary evaluation of casi preprocessing techniques. Proceedings of the 17th Canadian Symposium on Remote Sensing (Saskatoon, Canada: Canadian Remote Sensing Society and Canadian Aeronautics and Space Institute). 694- 697.
Primicerio, J., Di Gennaro, S.F., Fiorillo, E., Genesio, L., Lugato, E., Matese, A., and Vaccari, F.P. 2012. A flexible unmanned aerial vehicle for precision agriculture. Precis. Agric. 13:517– 523.
Raghuvansh, R., and Bisht, K. 2010. Uses of soybean: products and preperation. In G. Singh (Ed.), The soybean :404–426. CABI.
Ross, J. 1981. The radiation regime and architecture of plant stands. The Hague, The Netherlands: Dr Junk W.
Roth, M.G., Webster, R.W., Mueller, D.S., Chilvers, M.I., Faske, T.R., Mathew, F.M., Bradley, C.A., Damicone, J.P., Kabbage, M., and Smith, D.L. 2020. Integrated management of important soybean pathogens of the united states in changing climate. Journal of Integrated Pest Management. 11:17.
Santos, J.B., Procopio, S.O., Silva, A.A., and Costa, L.C. 2003. Capture and utilization of solar radiation by the soybean and common bean crops and by weeds. Bragantia. 62:147-153.
Savary, S., Willocquet, L., Pethybridge, S.J., Esker P.D., McRoberts, N., and Nelson, A. 2019. The global burden of pathogens and pests on major food crops. Nat. Ecol. Evol. 3:430–439.
Schneider, R.W., Hollier, C.A., Whitam, H.K., Palm, M.E., McKemy, J.M., Hernández, J.R., Levy, L., and DeVries-Paterson, R. 2005. First report of soybean rust caused by Phakopsora pachyrhizi in the Continental United States. Plant Dis. 89:774-774.
Serrago, R.A., Carretero, R., Bancal, M.O., and Miralles, D.J. 2009. Foliar diseases affect the ecophysiological attributes linked with yield and biomass in wheat (Triticuma estivum L.). Eur. J. Agronomy. 31:195–203. 30
Setiyono, T.D., Weiss, A., Specht, J.E., Cassman, K.G., and Dobermann, A. 2008. Leaf area index simulation in soybean grown under near-optimal conditions. Field Crop Res. 108:82–92.
Shaner, G., and Finney, R.E. 1977. The effect of nitrogen fertilization on the expression of slow- mildewing resistance in Knox wheat. Phytopathology 67:1051–1056.
Sherwin, H.S., and Kreitlow, K.W. 1952. Discoloration of soybean seeds by the frogeye fungus, Cercospora sojina. Phytopathology 42:568–572.
Shi, Y., Thomasson, J.A., Murray, S.C., Pugh, N.A., Rooney, W.L., Shafian, S., Rajan, N., Rouze, G., Morgan, C.L.S., Neely, H.L., et al. 2016. Unmanned aerial vehicles for high- throughput phenotyping and agronomic research. PLoS ONE. 11:e0159781.
Sinclair, J.B., and Backman, P.A. (ed.). 1989. Frogeye leaf spot. p.19–21. In Compendium of soybean diseases. 3rd ed. American Phytopathological Soc., St Paul, MN.
Sinclair, T.R., and Muchow, R.C. 1999. Radiation use efficiency. Adv. Agronomy. 65:215–265.
Smith, G.M., and Milton, E.J. 1999. The use of the empirical line method to calibrate remotely sensed data to reflectance. Int. J. Remote Sens. 20:2653–2662.
Sonobe, R., Yamaya, Y., Tani, H., Wang, X., Kobayashi, N., and Mochizuki, K.I. 2018. Crop classification from Sentinel-2-derived vegetation indices using ensemble learning. J. Appl. Remote Sens. 12:026019.
Stöckle C.O., and Kemanian A.R. 2009. Crop radiation capture and use efficiency: a framework for crop growth analysis In ‘Crop physiology: Applications for genetic improvement and agronomy’. eds. Sadras V.O., and Calderini D.F. (San Diego, CA, USA: Academic Press; ). 145– 170.
Sugiura, R., Tsuda, S., Tsuji, H., and Murakami, N. 2018. Virus-infected plant detection in potato seed production field by uav imagery. In Proceedings of the 2018 ASABE Annual International Meeting, Detroit, MI, USA. 1800594.
Tagliapietra, E.L., Streck, N.A., Rocha, T.S.M., Richter, G.L., da Silva, M.R., Cera, J.C., et al. 2018. Optimum leaf area index to reach soybean yield potential in subtropical environment. Agronomy. J. 110:932–938.
Taiz, L., Zeiger, E., Møller, I.M., and Murphy, A. 2014. Plant physiology and development. 2015. 6th ed.; Sinauer Assotiates: Sunderland, MA, USA. ISBN:1605352551.
Tebaldi, C., Hayhoe, K., Arblaster, J.M., and Meehl, G.E. 2006. Climate change, going to the extremes, an intercomparison of model simulated historical and future changes in extreme events. Clim. Change. 79:185–211.
Tetila, E.C., Machado, B.B., Belete, N.A.S., Guimarães, D.A., and Pistori, H. 2017. Identification of soybean foliar diseases using unmanned aerial vehicle images. IEEE Geosci. Remote Sens. Lett. 14:2190–2194. 31
Tewes, A., and Schellberg, J. 2018. Towards remote estimation of radiation use efficiency in maize using UAV-based low-cost camera imagery. Agronomy. 8:e16.
Tollenaar, M., and Aguilera, A. 1992. Radiation use efficiency of an old and a new maize hybrid. Agronomy J. 84:536–541.
Turner, F.T., and Jund, M.F. 1991. Chlorophyll meter to predict nitrogen topdress requirement for semidwarf rice. Agronomy J. 83:926-928.
Uddling, J., Gelang-Alfredsson, J., Piikki, K., and Pleijel, H. 2007. Evaluating the relationship between leaf chlorophyll concentration and SPAD-502 chlorophyll meter readings. Photosynth. Res. 91:37–46.
Viña, A., and Gitelson, A.A. 2005. New developments in the remote estimation of the fraction of absorbed photosynthetically active radiation in crops. Geophys. Res. Lett. 32.
Waggoner, P.E., and R.D. Berger. 1987. Defoliation, disease, and growth. Phytopathology 77:393 398.
Walker, J.T., Phillips, D.V., Melin, J., and Spradlin, D. 1994. Effect of intermittent acidic irrigation on soybean yields and frogeye leaf spot. Environ. Exp. Bot. 34:311-318.
Wang, C., and Myint, S.W. 2015. A simplified empirical line method of radiometric calibration for small unmanned aircraft systems-based remote sensing. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 8:1876–1885.
Warren-Wilson, J.W. 1967. Ecological data on dry-matter production by plants and plant communities. In: E.F. Bradley and O.T. Denmead, editors, The collection and processing of field data. Interscience Publ. New York. p.77–123.
Watson, D.J. 1947. Comparative physiological studies in the growth of feld crops. I. Variation in net assimilation rate and leaf area between species and varieties, and within and between years. Annals of Botany 11:41-76.
Westphal, A., Abney, T.S., and Shaner, G. 2006. Diseases of soybean. Frogeye leaf spot. Purdue University Department of Botany and Plant Pathology and USDA-ARS. http://www.ces.purdue.edu/extmedia/BP/BP-131-W.pdf
Wise, K., and Mueller, D. 2011. Are fungicides no longer just for fungi? An analysis of foliar fungicide use in corn. APSnet Features. doi:10.1094/APSnetFeature-2011-0531.
Woebbecke, D.M., Meyer, G.E., Von Bargen, K., and Mortensen, D.A. 1995. Color indices for weed identification under various soil residue and lighting conditions. Transactions of the ASAE. 38:259–269.
Wood, C.W., Reeves, D.W., Duffield, R.R., and Edmisten, K.L. 1992. Field chlorophyll measurements for evaluation of corn nitrogen status. J. Plant Nutri. 15:487-500. 32
Wrather, J.A., and Koenning, S.R. 2006. Estimates of disease effects on soybean yields in the United States 2003-2005. J. Nematol. 38:173-180.
Yang, X.B., Uphoff, M.D., and Sanogo, S. 2001. Outbreaks of soybean frogeye leaf spot in Iowa. Plant Dis. 85:443.
Yeom, J., Jung, J., Chang, A., Ashapure, A., Maeda, M., Maeda, A., and Landivar, J. 2019. Comparison of vegetation indices derived from uav data for differentiation of tillage effects in agriculture. Remote Sens. 11:1548.
Yu, N., Li, L., Schmitz, N., Tian, L.F., Greenberg, J.A., and Diers, B.W. 2016. Development of methods to improve soybean yield estimation and predict plant maturity with an unmanned aerial vehicle based platform. Remote Sens. Environ. 187:91–101.
Yuan, Z., Cao, Q., Zhang, K., Ata-Ul-Karim, S.T., Tian, Y., Zhu, Y., Cao, W., and Liu, X. 2016. Optimal leaf positions for spad meter measurement in rice. Front. Plant Sci. 7:719.
Zadoks, J.C. and Schein, R.D. 1979. Epidemiology and plant disease management. Oxford University Press, New York.
Zhang, G.R., Newman, M.A., and Bradley, C.A. 2012. First report of the soybean Frogeye leaf spot fungus (Cercospora sojina) resistant to quinone outside inhibitor fungicides in North America. Plant Dis. 96:767.
Zhang, D., Zhou, X., Zhang, J., Huang, L., and Zhao, J. 2017. Developing a small uav platform to detect sheath blight of rice. In Proceedings of the 2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Fort Worth, TX, USA. 3190–3193.
Zhang, G.R., Allen, T.W., Bond, J.P., Fakhoury, A.M., Dorrance, A.E., Weber, L., Faske, T.R., Giesler, L.J., Hershamn, D.E., Kennedy, B.S., et al. 2018. Widespread occurrence of quinone outside inhibitor fungicide-resistant isolates of Cercospora sojina, causal agent of frogeye leaf spot of soybean, in the United States. Plant Health Prog. 19:295–302.
Zheng, H., Cheng, T., Li, D., Zhou, X., Yao, X., Tian, Y., Cao, W., and Zhu, Y. 2018. Evaluation of RGB, color-infrared and multispectral images acquired from unmanned aerial systems for the estimation of nitrogen accumulation in rice. Remote Sens. 10:824.
Zhou, X., Zheng, H., Xu, X., He, J., Ge, X., Yao, X., Cheng, T., Zhu, Y., Cao, W., and Tian, Y. 2017. Predicting grain yield in rice using multi-temporal vegetation indices from uav-based multispectral and digital imagery. ISPRS J. Photogramm. Remote Sens. 130:246–255. 33
CHAPTER 2. ESTIMATING SOYBEAN RADIATION USE EFFICIENCY USING A UAV IN IOWA
Xavier A. Phillips 1, Yuba R. Kandel 1, Mark A. Licht 2 and Daren S. Mueller 1 1 Department of Plant Pathology and Microbiology, Iowa State University, Ames, IA 50011.
2 Department of Agronomy, Iowa State University, Ames, IA 50011.
Modified from a manuscript published in Agronomy.
Abstract
Radiation use efficiency (RUE) is difficult to estimate and unreasonable to perform on a small plot scale using traditional techniques. However, the increased availability of Unmanned
Aerial Vehicles (UAVs) provides the ability to collect spatial and temporal data at high resolution and frequency, which has made a potential workaround. An experiment was completed in Iowa to (i) demonstrate RUE estimation of soybean [Glycine max (L.) Merr.] from reflectance data derived from consumer-grade UAV imagery and (ii) investigate the impact of foliar fungicides on RUE in Iowa. Some fungicides are promoted to have plant health benefits beyond disease protection, and changes in RUE may capture their effect. Frogeye leaf spot severity did not exceed 2%. RUE values ranged from 0.98 to 1.07 and 0.96 to 1.12 across the entire season and the period post-fungicide application, respectively, and fell within the range of previously published soybean RUE values. Plots treated with fluxapyroxad + pyraclostrobin had more canopy cover (P = 0.078) compared to the non-treated control 133 days after planting
(DAP), but yields did not differ. A “greening effect” was detected at the end of the sample collection. RUE estimation using UAV imagery can be considered a viable option for the evaluation of management techniques on a small plot scale. Since it is directly related to yield,
RUE could be an appropriate parameter to elucidate the impact of plant diseases and other stresses on yield. 34
Introduction
Radiation use efficiency (RUE), also known as light use efficiency (LUE), is defined as the plant’s ability to convert photosynthetically active radiation (PAR) into biomass on a per unit basis (Monteith, 1977). This measurement of photosynthetic performance is important for crop growth modeling (Muchow et al., 1993). RUE varies between species and even among cultivars, but generally, cultivar-dependent RUE values are unavailable (Morel et al., 2014).
There is no uniform procedure for RUE estimation. Of the three commonly used methods to estimate RUE (i) from incoming radiation (RUEinc), (ii) total absorbed light (RUEtotal), and
(iii) radiation absorbed by photosynthetically active vegetation (RUEgreen) (Gitelson and
Gamon 2015), Tewes and Schellberg (2018) made the case to have RUEgreen (Equation (1)) be the standard method. By only considering the photosynthetically active vegetation to determine the fraction of absorbed photosynthetically active radiation (fAPARgreen) (Equation (2)), RUE estimation becomes sensitive to the changes of the photosynthetically active radiation (PAR) absorption through the reproductive stages, senescence, and potentially to disease.
퐷푀 푅푈퐸푔푟푒푒푛 = × 푃퐴푅푖푛푐 (1) 푓퐴푃퐴푅푔푟푒푒푛
퐺퐿퐴퐼 푓퐴푃퐴푅 = 푓퐴푃퐴푅 × ( ) (2) 푔푟푒푒푛 푇퐿퐴퐼
RUE estimation uses in-field sensors and destructive sampling to acquire the needed components consisting of incoming radiation, reflectance data, total leaf area index (TLAI), green leaf area index (GLAI), and dry matter (DM). Sensors are placed below the canopy facing the soil for PARsoil, above the canopy facing down to gather reflectance (PARout), above the 35 canopy facing up for PARinc, and beneath the canopy facing up to determine the PAR passing through the canopy (PARtrans). This cumbersome method makes RUE estimation on a small plot scale with various treatments unreasonable. However, applications involving Unmanned
Aerial Vehicles (UAV) provides a potential workaround. Tewes and Schellberg (2018) demonstrated the derivation of reflectance (PARout) for estimation of RUE in corn (Zea mays,
L.) using a UAV and a consumer-grade camera. Their RUE estimations were similar to previously published values for maize (Tewes and Schellberg, 2018).
Soybean (Glycine max (L.) Merr.) RUE values range from 0.60 to 2.53 g Mj−1 (Muchow,
1985; Santos et al., 2003) (Table 1). RUE values are assumed to be constant. However, Gitelson et al. (2014) describe facultative and constitutive changes that result in diurnal and seasonal fluctuations of RUEgreen (Gamon and Berry, 2012). For example, soybean plants experience stress from diseases or disorders during the growing season, which may negatively affect RUE values.
Foliar disease would be considered a constitutive property since disease development may be relatively slow and irreversible. Foliar diseases such as soybean rust (SBR) and frogeye leafspot (FLS) caused by the pathogens Phakopsora pachyrhizi Syd. and Cercospora sojina
Hara, respectively, reduce green leaf area. The severity of the foliar disease often increases over the growing season. Specifically, Kumudini et al. (2008) showed biomass reductions were correlated to the reduction of photosynthetically active radiation (APAR) and the RUE of non- lesion green LAI caused by SBR lesions. With respect to disease severity, the lesion area of the foliar pathogen SBR reduced APAR between 5% and 20% (Kumudini et al., 2008). 36
The use of foliar fungicides has increased since 2005 in soybean to manage foliar diseases (Wise and Mueller, 2011). In addition, the quinone outside inhibitor (QoI) containing fungicides have been promoted to have plant health benefits in the absence of disease, called the
“greening effect” (Wise and Mueller, 2011; Balba, 2007; Bartlett et al., 2002). This phenomenon is described as improved photosynthetic efficiency by maintaining green leaf area to allow for additional dry matter accumulation, resulting in increased yields (Morrison et al., 1999;
Kumudini et al., 2001; Kyverga et al., 2013). However, yield benefits have been inconsistent
(Bradley and Sweets 2008; Kyverga et al., 2013; Kandel et al., 2016). Conversely, Phillips et al.
(2017) cautioned that prophylactic applications of QoI fungicides increased the likelihood of developing green stem disorder (GSD).
With new technology, soybean RUE may be estimated using remote sensing. The efficient estimation may encourage the incorporation of RUE into soybean small plot research, improvements of yield models, and to evaluate the usefulness of in-season management practices. The objectives of this study were to: (i) demonstrate RUE estimation of soybean from reflectance data derived from consumer-grade UAV imagery and (ii) investigate the impact of foliar fungicides on RUE in Iowa.
Materials and Methods
Soybean cultivar NK S29-K3X was planted near Kanawha, IA on 16 May 2019 at a rate of 395,362 seeds ha−1 in eight row plots that were 7.8-m long and spaced 76-cm apart (Table 2).
Location soil type was Nicollet clay loam with a slope of 1 to 3 percent. Three fungicide treatments were organized in a randomized complete block design (RCBD) with six replications.
The three treatments were (1) non-treated control, (2) fluxapyroxad + pyraclostrobin (Priaxor©,
0.05-L ha−1, BASF, Research Triangle Park, NC), and (3) flutriafol + fluoxastrobin (Preemptor©, 37
0.06-L ha−1, FMC Agricultural Solutions, Philadelphia, PA, USA). Pyraclostrobin and fluxastrobin are QoI (FRAC code 11) fungicides, while fluxapyroxad is a succinate dehydrogenase inhibitor (SDHI; FRAC code 7) and flutriafol is a demethylation inhibitor (DMI;
FRAC code 3) fungicide. Fungicides were applied 67 days after planting at the R3 growth stage, beginning pod (Fehr and Caviness, 1977), with a self-propelled research sprayer powered by a
CO2 tank that delivered fungicides using XR 11002 nozzles at 142 to 189 L ha–1 at 241 kpa. The middle two rows of each plot were harvested on 29 October 2019 to assess yield using a 2009
Almaco SPC20 research plot combine (ALMACO, Nevada, IA, USA). Yield calculations were adjusted to 13% seed moisture.
Field data were collected every 6 to 11 days from the VC through R7.5 growth stage (22–
130 days after planting) (Fehr and Caviness, 1977) for a total of 14 sampling dates. At each sampling date (SD), a single 0.5-m section from one of the three outermost rows was examined.
To avoid edge effects, this sampling section was kept 0.5-m from row ends and previously sampled sections (Probst, 1943). From this portion of the row, total plant count, disease severity, leaf area, and dry mass data were obtained. FLS severity was determined from the upper and lower canopies of all plants and averaged for each experimental unit, a single plot. Leaf area and
DM measurements were derived from a maximum of 10 plants cut at the soil line and bundled together. After cutting, plant bundles were stored in labeled plastic bags at 4 °C until further processing 2 to 10 days later.
The leaf area of the measured section of the row was determined using an LI-3100C (LI-
COR, Inc., Lincoln, NE, USA). All leaves were stripped from the collected plants and separated into two groups, green and non-green (>50% non-green for any reason). A single leaf area value, consisting of the sum area of all leaves, was assigned to each green and non-green group. This 38 partitioning allowed us to collect GreenLAI (GLAI) and YellowLAI (YLAI) values to be used in the APAR equation.
Leaf area was converted to m2 and used to estimate leaf area index (LAI; m2 m−2) based on the number of plants sampled and the number of plants in the sampled area (Maitree and
Toyota, 2017). All above-ground plant parts from each plot were dried for seven days in brown paper bags placed in drying bins with fans at 60 °C (Phillips et al., 2017). Dried samples were weighed in bags and adjusted to exclude bag weight to obtain dry matter of the entire plant including seeds (DM; Taylor Precision Products, Oak Brook, IL, Glass Digital Food Scale, model: 3842BL9).
True color aerial imagery was obtained during each collection date using a Phantom 4
ProV2.0 (DJI, Shenzhen, China), a UAV, with an onboard 2.54-cm CMOS 20 MP camera.
Individual plot images were taken for the first four sampling dates at an above-ground altitude
(AGL) of 9-m. Mission planning and image stitching of the remaining 10 sampling dates were completed through service acquired from DroneDeploy (DroneDeploy, San Francisco, CA,
USA). The 14 min flight plan was consistent across sampling dates 9-m above ground altitude
(AGL), with 75% overlap, and a flight speed of 3-m s−1. Images (~670) were uploaded to and stitched together by DroneDeploy. The resulting orthomosaic had a resolution of 0.69-cm/px and was exported as a geotiff for image processing in ArcGIS (ArcGIS Pro 2.4.2, Esri, Redlands,
CA, USA).
Three 1.2-m portable fabric targets (Group 8 Technology, Inc., Provo, UT, USA) with reflectance values of approximately 3%, 13%, and 56% were placed at the field edge during each flight for calibration from digital number (DN) to reflectance (Table 3). The DN’s of each target 39 of all three bands were extracted and averaged for each sampling date. The empirical line method
(Smith and Milton, 1999) was used for calibration. However, since the raw target DN vs. reflectance plot showed an exponential relationship, a natural logarithm transformation of the reflectance values for the calibration targets was performed to obtain a linear relationship (Wang and Myint, 2015). A line of best fit was added to the scatter plot and the resulting linear regression equation was used to convert the DNs of each band to -ln(reflectance) (Table A1).
This was then converted to reflectance using the exponential term (Wang and Myint, 2015).
In ArcGIS (ArcGIS Pro 2.4.2), the three-band orthomosaics were converted into a single raster layer using the raster calculator and the excess green (ExG) index (Woebbecke et al.,
2005). The resulting raster was examined manually to identify a threshold value to separate foliage from the background (Meyer and Neto, 2008). Using the manually identified threshold value, the raster was binarized into “foliage” and “other”. The “foliage” class contained both yellow and green soybean foliage. The “other” class contained bare soil, undecipherable shadowed area, and residue. Only the inner two rows used for yield were used in the image analysis. The binary raster was used as a mask to extract only the DNs of the “foliage” class from the three collected bands (red, green, and blue) of the respective images and orthomosaics.
The fraction of foliar cover (fCover) was determined by extracting the sum of the areas of each class and determining the percent of the “foliar” class with respect to the combined total area.
Plot summaries containing average DNs of each band were exported to Microsoft Excel after image processing (Figure 1).
The fAPARgreen and fAPARtotal values were calculated using Equations (2) and (3), as described by Tewes and Schellberg (2018) with the exception of excluding PARsoil from the calculations (Lindquist et al., 2005; Tollenaar and Aguilera, 1992). fPARtrans was estimated using 40 the Lambert–Beer law equation using an extinction coefficient (k) of 0.523 (Flènet et al., 1996) based on crop and row spacing. PARinc was determined by taking the total incoming solar radiation collected from a weather station near the experiment location and multiplying it by 0.5
(Taiz and Zeiger, 2014). It was assumed that the RGB camera collected bands within the range of photosynthetically active radiation (PAR; 400-700-nm). The definite integral of the running total of PAR during the growing season vs. fAPARgreen was then plotted against the DM (Tewes and Schellberg, 2018). RUE was determined from the slope value obtained by the regression equation from the plot of the sum of DM vs. the sum of APARgreen.
푓퐴푃퐴푅푡표푡푎푙 = (100% − 푓푃퐴푅표푢푡 − 푓푃퐴푅푡푟푎푛푠) × 푓퐶표푣푒푟 (3)
Disease severity ratings taken during each destructive sample date were removed from the GLAI by converting the severity rating to a 0-1 scale and multiplying this by the GLAI value
(Equation (4)). This reduced only the GLAI by the observed percent of disease severity to result in GLAIDx. A handheld GreenSeeker (Trimble, Sunnyvale, CA, USA) was used to collect
Normalized Difference Vegetation Index (NDVI) values. Values were collected on the last four sampling dates to avoid soil background noise early in the season and value saturation during full canopy. A yield row was walked with the GreenSeeker and NDVI values were recorded beginning and ending approximately 0.5m from the ends of the plot to avoid edge-effects
(Probst, 1943).
퐺퐿퐴퐼퐷푥 = 퐺퐿퐴퐼 × (1 − 푥) (4)
41
Statistical analyses were performed in SAS 9.4 (SAS Institute, Cary, NC, USA). All measured variables were subjected to an analysis of variance using the Glimmix procedure. For testing the fungicide effect, fungicide treatment was set as a fixed factor while replication and days after planting (DAP) were considered random factors. Replication was nested within DAP.
Least squared means (LS-means) were estimated using the Lsmeans statement in PROC
GLIMMIX. Means were separated by using “PDIFF lines” option in the LSmeans statement at an α level of 0.10. Since fungicide treatments were applied after 63 DAP, a separate data set only using sampling periods after the fungicide treatment was used to determine treatment effects on collected variables using the same statistical methods. The Wyffeos slope comparison was used to determine the significance of the RUE values. The effect of fungicides was determined using orthogonal contrasts of the RUE values.
Results
RUE values of soybean from VC to R7.5 (22 to 130 DAP) were estimated to be 0.98,
1.07, and 1.07 g MJ-1 for the non-treated control, fluxapyroxad + pyraclostrobin, and the flutriafol + fluoxastrobin fungicide treatments, respectively (Figure 2). R2 values of the line of best fit were 0.98, 0.98, and 0.99 for the non-treated control, fluxapyroxad + pyraclostrobin, and flutriafol + fluoxastrobin treatments, respectively. RUE values from R3 to R7.5 (70 to 130 DAP) were estimated to be 0.96, 1.12, and 1.06 g MJ−1 for the non-treated control, the fluxapyroxad + pyraclostrobin, and the flutriafol + fluoxastrobin treatments, respectively (Figure 2). R2 values of the line of best fit were 0.98, 0.98, and 0.99 for the non-treated control, fluxapyroxad + pyraclostrobin, and flutriafol + fluoxastrobin treatments, respectively, from R3 to R7.5 (70-130
DAP). Orthogonal contrasts showed the estimated RUE values did not differ (P > 0.10) among the treatments within either timeframe (Table 4). 42
Canopy cover increased rapidly following 29 DAP, plateaued at 90 DAP, and reached a maximum coverage of nearly 100% then began to decline 120 DAP (Figure 3). An interaction (P
= 0.025) was detected between treatment and DAP for the percent of foliage cover (Table 5). At
130 DAP (SD 14), the last sampling date, plots treated with fluxapyroxad + pyraclostrobin had approximately 10% more foliage cover than the non-treated control (P < 0.1; Table 5). Foliar cover of the flutriafol + fluoxastrobin treatment was approximately 5% greater (P > 0.1) than the non-treated control (Table 5).
DAP effect was significant (P ≤ 0.001) for TLAI, GLAIDx, DM, and NDVI values. The
TLAI reached a maximum value of 5.45 at 90 DAP (SD 10; Figure 4A). At 82, 90, and 98 DAP
(SD 9, 10, and 11), TLAI did not differ (Figure 4A). The TLAI decreased after 90 DAP. TLAI at
108 DAP and later were all significantly less than that at 98 DAP (Figure 4A). GLAIDx reached a maximum value of 5.30 at 90 DAP (SD 10; Figure 4B). Samples collected after 90 DAP showed significantly diminishing GLAIDx (Figure 4B). Average DM was not significantly different between the first four sampling dates and was relatively low. It sharply increased beginning 49 DAP and continued to increase up to the last sampling date (Figure 4C). The greatest NDVI values were measured at 0.91, at 90 DAP (SD 10; Figure 4D). The value was significantly less than 130 DAP (SD 14) than the previous four sampling dates, with a value of
0.51.
A treatment by date interaction (P = 0.025) was seen for FLS severity. FLS was not observed until 108 DAP (SD 12). Severity was low, not exceeding 2% during the entire experiment. However, the non-treated control had significantly more FLS on the SD 13 than either of the fungicide treatments on the same date (Table 6). FLS severity of the non-treated control was numerically, however not significantly (P = 0.358), less on the last sampling date 43
(130 DAP) than either fungicide treatment (Table 6). Treatment had a significant (P < 0.001) effect on yield (Table 6). The non-treated control and the fluxapyroxad + pyraclostrobin treatment had greater yields than the flutriafol + fluoxastrobin treatment by approximately 134.5 kg ha-1 (Table 6). P-values of date, treatment, and the date*treatment interaction for all variables can be viewed in Table A2.
Discussion
This is the first study to estimate soybean RUE using consumer-grade UAV imagery. Our robust data set, albeit a single year, was built by using six replications collected every 6–11 days, as Tewes and Schellberg (2018) did with corn RUE estimation. Using this technology, we could estimate the impact of certain QoI-containing fungicides on RUE at low levels of FLS severity in
Iowa.
Littleton et al. (1979) and Muchow (1985) reported RUE declined during pod fill.
Muchow et al. (1993) postulated that the decline of RUE was due to leaf drop and a reduction in specific leaf nitrogen. Here, the RUE values of the non-treated control and the flutriafol + fluoxastrobin treatment were slightly less for the post-fungicide application timeframe beginning at R3 (beginning pod), while the RUE values for fluxapyroxad + pyraclostrobin treatment were numerically greater for the post-fungicide application timeframe beginning at R3 (beginning pod) compared to the season long estimated RUE values.
The RUE values of the non-treated control of both the season-long and the post-fungicide sampling periods were numerically, but not significantly, less than the fungicide treatments for the same sampling periods. The beneficial effect of these fungicides was not captured with the estimated RUE values. Our sampling dates may not have captured the main benefit of the 44
“greening effect”, which extends the period of photosynthetic activity by the leaves (Bertelsen et al., 2001).
At the last sampling date (130 DAP), canopy cover of the fluxapyroxad + pyraclostrobin treatment was significantly greater than the control. This significant retention of canopy cover in the fluxapyroxad + pyraclostrobin treatment did not contribute enough to impact the TLAI or
DM. However, we did not collect fallen leaves to be included with the DM and LAI measurements, and we speculate this may have altered the other components of RUE enough to show an effect.
LAI values followed the typical logistic curve until senescence began (Setiyono et al.,
2008). The average maximum value found in this experiment was slightly less than the range of
6 to 7 identified by Dermody et al. (2006) and Tagliapietra et al. (2018). The number of plants collected for LAI determination match or exceed previous studies (Muchow et al., 1993; Maitree and Toyota, 2017). TLAI and GLAIDx plots and values were similar. This is reasonable since the severity of FLS was low. Because soybeans overinvest in leaves, this small amount of leaf area loss is unlikely to impact yield (Dermody et al., 2006; Haile et al., 1998; Conley et al., 2008;
Conley et al., 2009).
Fungicide application did not have a significant effect on DM, so the entire sampling period (22–130 DAP) was analyzed similarly to the TLAI and GLAIDx. The average DM of the last two sampling dates (120 and 130 DAP) was the greatest of all the dates and not significantly different from one another. However, during these same sampling periods, TLAI, GLAIDx, and
NDVI values were at their lowest (Figure 4). At the last sampling date (130 DAP), TLAI was reduced to approximately 28% of the previous date (P < 0.001; 120 DAP). In addition, the 45
GLAIDx was approximately one-tenth of the prior date, and NDVI values were also significantly less than the prior date (P < 0.001; Figure 4). Despite these large and significant reductions in
LAI, they were not enough to show a change in DM.
During June-August 2019, total precipitation was 20% less and average temperature (F°) was 0.3% below the 30-year average (http://mesonet.agron.iastate.edu). Thus, environmental conditions for FLS were not highly conducive, and severity was low. However, there was a significant effect with both fungicide treatments, showing less disease. This low level of severity did not correspond to the yield data and was not captured by the estimated RUE. Yield of the fluxapyroxad + pyraclostrobin treatment did not differ from the non-treated control, despite showing greater canopy cover at the last sampling date (P = 0.078; 130 DAP). Treatment did not have an effect on the DM or either LAI values. Perhaps the sink was reduced and resulted in an imbalance that favored the source. Thus, additional photosynthates may have been stored in the stem, which the DM measurements failed to capture. In this study, the components of the plant were not separated into individual plant parts (pods, stems, and leaves). Doing so may have provided insight into these contradictions.
The “greening effect” of the canopy cover as described by Balba (2007) and Bartlett et al.
(2002) may have been beginning, but NDVI values collected by a handheld GreenSeeker did not detect a treatment effect. Phillips et al. (2017) found that the application of pyraclostrobin + fluxapyroxad applied at R3 increased the dry mass of stems at harvest. Perhaps the potential benefit of these QoI chemistries would have developed if sampling would have been extended to include samples closer to plant maturity. 46
Although we do not know the exact sensitivity of the consumer-grade unmodified camera that was used in our study, our RUE estimates fell within previously estimated values. Berra et al. (2015) found the consumer-grade unmodified cameras they tested covered the 400-700 nm window of PAR. Burggraaff et al. (2019) found that the stock camera of the Phantom 4 Pro covered the window of PAR but with a near-infrared (NIR) cut-off of approximately 670 nm.
The UAV used in this study is stated to have the same camera as the one tested by Burggraaff et al. (2019). We assumed, as Tewes and Schellberg (2018) did, that our sensor captured the PAR region.
Conclusion
RUE estimation with the help of UAV imagery may be considered a viable option for the evaluation of management techniques on a small plot scale. Since it is directly related to yield,
RUE is a more appropriate parameter to evaluate stress and management practices. This could be especially beneficial to phenotyping experiments, though these would require non-destructive methods of determining LAI and DM due to their small plot size and large plot number.
Consumer-grade UAVs and imagery were intentionally used in hopes that their cost and availability would lend to the widespread adoption of RUE as a standard evaluation parameter.
The application of these fungicides did not provide a yield benefit. However, the “greening effect” may have occurred as we detected significantly more canopy cover in one of the fungicide treatments while the other fungicide had numerically greater canopy cover compared to the non-treated control. If FLS would have occurred earlier or with greater severity, yield may have been significantly impacted, and the fungicide treatments may have significantly affected the RUE. 47
Tables and Figures
Table 1. Radiation use efficiency (RUE) of soybean estimated in previous reports. Publication RUE g M−1 Muchow, 1985 0.60 Muchow et al., 1993 0.88 Adeboye et al., 2016 1.07 Nakaseko and Gotoh, 1983 1.20 Schöffel and Volpe, 2001 1.23 Singer et al., 2011 1.46 Confalone and Dujmovich, 1999 1.92 de Souza et al., 2009 1.99 Santos et al., 2003 2.53
48
Table 2. Details of the 2019 field experiment in Kanawha, Iowa, evaluating the use of an unmanned aerial vehicle with a consumer-grade camera for the estimation of the radiation use efficiency of soybean. Detail Specifics Latitude and Longitudez 42°54′36.6′′ N and 93°47′33.8′′ W Cultivar NK S29-K3X Planting Population 395,362 per hectare Plot Length 7.8-m Planting Date 16 May Emergence 30 May Fungicide application date (R3) b 22 July Harvest Date 29 Oct z Latitude and Longitude format in degrees, minutes, and seconds (DMS). b R3 = beginning of pod formation (Fehr and Caviness, 1977).
49
Table 3. Target mean reflectance values used for image calibration during the field experiment in Kanawha, Iowa, during 2019. Average Reflectancez Targety 420–700 nm 420–1050 nm 900–1700 nm Black (3%) 2.7 2.8 3.1 Gray (13%) 13.4 12.7 11.2 White (56%) 55.6 56.1 55.5 y Type 822–1.2-m grayscale calibration panels (Group 8 Technology, Inc., Provo, UT, USA) with average reflectance value. z The average measured reflectance values (%) of the calibration targets using a Perkin- Elmer Lambda 1050 Spectrophotometer with a 150 mm diameter integrating sphere. Data collected by Group 8 Technology (Group 8 Technology, Inc, Provo, UT, USA).
50
Table 4. Single degree of freedom contrasts of radiation use efficiency (RUE) derived from the linear line of best fit added to the plot of the sum of absorbed photosynthetically active radiation (APAR) and cumulative dry mass across the entire growing season and after the application of fungicidey treatments in 2019 in Iowa. Comparisons of P Timeframez Treatments RUE (g Mj−1) Value NTC vs. fluxapyroxad + 0.98 vs. 1.07 0.132 pyraclostrobin NTC vs. flutriafol + Season long 0.98 vs. 1.07 0.124 fluoxastrobin fluxapyroxad + pyraclostrobin 1.07 vs. 1.07 0.963 vs. flutriafol + fluoxastrobin NTC vs. fluxapyroxad + 0.96 vs. 1.12 0.325 pyraclostrobin NTC vs. flutriafol + Post-fungicide application 0.96 vs. 1.06 0.513 fluoxastrobin fluxapyroxad + pyraclostrobin 1.12 vs. 1.06 0.733 vs. flutriafol + fluoxastrobin y Treatments = Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions). z There were a total of 14 sampling dates during the growing season that ranged from 22 to 130 DAP. The season long comparisons encompassed all 14 dates while the post-fungicide application comparisons consisted of the last eight sampling dates (70–130 DAP). Fungicides were applied at 67 days after planting.
51
Table 5. Canopy cover (%) of sampling dates after fungicide application at growth stage R3w in a field trial conducted in Iowa during 2019. Canopy Coverxy Days After Planting Fungicidez 70 75 82 90 98 108 120 130 NTC 87.5 92.8 96.6 99.8 96.2 99.9 96.5 80.5 b Fluxapyroxad + 87.5 92.9 96.8 99.9 96.3 99.9 96.4 89.9 a pyraclostrobin 85.0 Flutriafol + fluoxastrobin 86.5 92.8 96.3 99.9 96.2 99.9 96.9 ab P value 0.864 0.992 0.903 0.680 0.980 0.426 0.793 0.078 w Growth stage R3, beginning pod (Fehr and Caviness, 1977). x Canopy cover was determined by performing raster classification using ArcGIS (ArcGIS Pro 2.4.2) and extracting the sum of the areas of each class and determining the percent of the “foliar” class with respect to the combined total area of the center two yield rows. y Least-squares means were separated by Fishers’ protected least significant difference at = 0.10. Numbers followed by the same letter within a column are not significantly different at α 0.10 level. z Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions).
52
Table 6. Frogeye leaf spot (FLS) severity and yield by fungicides applied at growth stage R3w and date from the field experiment in 2019 in Iowa. Yield (kg FLSxy ha−1) Days After
Planting Fungicide z 120 130 NTC 1.7 a 0.7 a 3551 a Fluxapyroxad + 1.0 b 1.3 a 3544 a pyraclostrobin Flutriafol + fluoxastrobin 1.0 b 1.2 a 3390 b P value 0.091 0.358 < 0.001 w Growth stage R3, beginning pod (Fehr and Caviness, 1977). x FLS severity was visually estimated from 0-100% of the 0.5-m sampling area of each plot. y Least-squares means were separated by Fishers’ protected least significant difference at = 0.10. Numbers followed by the same letter within a column are not significantly different at α 0.10 level. z Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions).
53
Figure 1. Radiation use efficiency (RUE) estimation workflow illustrating the components and their source, needed to derive RUEgreen. fPARout = fraction of photosynthetically active radiation reflected; fCover = percent of canopy cover; fPARtrans= fraction of photosynthetically active radiation transmitted through the canopy; GLAI = green leaf area index; TLAI = total leaf area index; DM = dry matter; fAPARtotal = the fraction of absorbed photosynthetically active radiation of all leaf tissue; fAPARgreen = the fraction of absorbed photosynthetically active radiation by only the green leaf tissue; PARinc = incoming photosynthetically active radiation; APARgreen = absorbed photosynthetically active radiation by the green leaf tissue; RUEgreen = radiation use efficiency of only the green leaf tissue.
54
Non-treated control Fluxapyroxad + pyraclostrobin Flutriafol + fluoxastrobin
900
800 y = 1.07x ± 0.05 y = 1.07x ± 0.04 700 y = 0.98x ± 0.04 2 2 2 600 R = 0.98 R = 0.98 R = 0.99
500
400
300
200
100 )
-2 0
0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800
800 Dry matter m Dry (g 700 y = 0.96x ± 0.10 y = 1.12x ± 0.13 y = 1.06x ± 0.09 2 2 2 600 R = 0.94 R = 0.93 R = 0.96
500
400
300
200 200 300 400 500 600 700 800 200 300 400 500 600 700 800 200 300 400 500 600 700 800
APAR (Mj m-2) green Figure 2. The function of accumulated dry matter and absorbed photosynthetically active radiation by the green leaf area (APARgreen) excluding frogeye leaf spot disease severity. The top row represents sampling dates from 22 to 130 days after planting (DAP) and the bottom row represents sampling dates after fungicide application, 70 to 130 DAP. Radiation use efficiency (RUE) is the slope of the regression equation. The intercept of the equation is not shown. 55
Figure 3. The development of foliar cover across the collected sampling dates from 22 to 130 days after planting (DAP) with error bars showing standard deviation.
56
Figure 4. (A) Total Leaf Area Index (TLAI), (B) Green Leaf Area Index with disease severity removed (GLAIDx) development during the entire sampling collection time frame (22 to 130 days after planting, DAP), (C) dry mass, and (D) Normalized Difference Vegetation Index (NDVI) values collected with a handheld GreenSeeker with error bars showing standard deviation.
57
Literature Cited
Adeboye, O.B.; Schultz, B.; Adekalu, K.O.; Prasad, K. Impact of water stress on radiation interception and radiation use efficiency of soybeans (Glycine max L. Merr.) in Nigeria. Braz. J. Sci. Tech. 2016, 3, 1–21, doi:10.1186/s40552-016-0028-1.
Balba, H. Review of strobilurin fungicide chemicals. J. Environ. Sci. (China) 2007, 42, 441– 451, doi:10.1080/03601230701316465.
Bartlett, D.W.; Clough, J.M.; Godwin, J.R.; Hall, A.A.; Hamer, M.; Parr‐Dobrzanski, B. The strobilurin fungicides. Pest Manag. Sci. 2002, 58, 649–662, doi:10.1002/ps.520.
Berra, E.; Gibson-Poole, S.; MacArthur, A.; Gaulton, R.; Hamilton, A. Estimation of the spectral sensitivity functions of un-modified and modified commercial off-the-shelf digital cameras to enable their use as a multispectral imaging system for UAVs. ISPRS-Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2015, XL-1/W4, 207–214, doi:10.5194/isprsarchives-XL-1-W4-207-2015.
Bertelsen, J.R.; De Neergaard, E.; Smedegaard‐Petersen, V. Fungicidal effects of azoxystrobin and epoxiconazole on phyllosphere fungi, senescence and yield of winter wheat. Plant Pathol. 2001, 50, 190–205, doi:10.1046/j.1365-3059.2001.00545.x.
Bradley, K.W.; Sweets, L.E. Influence of glyphosate and fungicide coapplications on weed control, spray penetration, soybean response, and yield in glyphosate-resistant soybean. Agron. J. 2008, 100, 1360–1365, doi:10.2134/agronj2007.0329.
Burggraaff, O.; Schmidt, N.; Zamorano, J.; Pauly, K.; Pascual, S.; Tapia, C.; Spyrakos, E.; Snik, F. Standardized spectral and radiometric calibration of consumer camers. Opt. Express 2019, 27, 19075–19101, doi:10.1364/OE.27.019075.
Confalone, A.E.; Dujmovich, M.N. Influence of deficit water on the efficiency of solar radiation on soybean. Brazilian J. Agrociencia. 1999, 5, 195–198.
Conley, S.P.; Abendroth, L.; Elmore, R.; Christmas, E.P.; Zarnstorff, M. Soybean seed yield and composition response to stand reduction at vegetative and reproductive stages. Agron. J. 2008, 100, 1666–1669, doi:10.2134/agronj2008.0082.
Conley, S.P.; Pedersen, P.; Christmas, E.P. Main-stem node removal effect on soybean seed yield and composition. Agron. J. 2009, 101, 120–123, doi:10.2134/agronj2008.0123.
Dermody, O.; Long, S.P.; DeLucia, E.H. How does elevated CO2 or ozone affect the leaf- area index of soybean when applied independently? New Phytologist 2006, 169, 145–155, doi:10.1111/j.1469-8137.2005.01565.x. 58 de Souza, P.J.D.O.P.; Ribeiro, A.; da Rocha, E.J.P.; et al. Solar radiation use efficiency by soybean under field conditions in the Amazon region. Pesquisa Agropecuaria Brasileira. 2009, 44, 1211–1218, doi:10.1590/S0100-204X2009001000001
Fehr, W.R.; Caviness, C.E. Stages of Soybean Development; Special Report 80; Iowa State University Coop. Ext. Serv.: Ames, IA, USA, 1977.
Flénet, F.; Kiniry, J.R.; Board, J.E.; Westgate, M.E.; Reicosky, D.C. Row spacing effects on light extinction coefficients of corn, sorghum, soybean and sunflower. Agron. J. 1996, 88, 185–190, doi:10.2134/agronj1996.00021962008800020011x.
Gamon, J.A.; Berry, J.A. Facultative and constitutive pigment effects on the Photochemical Reflectance Index (PRI) in sun and shade conifer needles. Isr. J. Plant Sci. 2012, 60, 85–95, doi:10.1560/IJPS.60.1-2.85.
Gitelson, A.A.; Peng, Y.; Arkebauer, T.J.; Schepers, J. Relationships between gross primary production, green LAI, and canopy Chlorophyll content in maize: Implications for remote sensing of primary production. Remote Sens. Environ. 2014, 144, 66–72, doi:10.1016/j.rse.2014.01.004.
Gitelson, A.A.; Gamon, J.A. The need for a common basis for defining light-use efficiency: Implications for productivity estimation. Remote Sens. Environ. 2015, 156, 196–201, doi:10.1016/j.rse.2014.09.017.
Haile, F.J.; Higley, L.G.; Specht, J.E.; Spomer, S.M. Soybean leaf morphology and defoliation tolerance. Agron. J. 1998, 90, 353–362, doi:10.2134/agronj1998.00021962009000030007x.
Kandel, Y.R.; Mueller, D.S.; Hart, C.E.; Bestor, N.R.; Bradley, C.A.; Ames, K.A.; Giesler, L.J.; Wise, K.A. Analyses of yield and net economic response from foliar fungicide and insecticide applications to soybean in the North Central United States. Plant Health Prog. 2016, 17, 232–238, doi:10.1094/PHP-RS-16-0038.
Kumudini, S.; Hume, D.J.; Chu, G. Genetic improvement in short season soybeans: I. Dry matter accumulation, partitioning, and leaf area duration. Crop Sci. 2001, 41, 391–398, doi:10.2135/cropsci2001.412391x.
Kumudini, S.; Godoy, C.V.; Board, J.E.; Omielan, J.; Tollenaar, M. Mechanisms involved in soybean rust-induced yield reduction. Crop Sci. 2008, 48, 2334–2342, doi:10.2135/cropsci2008.01.0009.
Kyveryga, P.M.; Blackmer, T.M.; Mueller, D.S. When do foliar pyraclostrobin fungicide applications produce profitable soybean yield responses? Plant Health Prog. 2013, 14, 6, doi:10.1094/PHP-2013-0928-01-RS. 59
Lindquist, J.L.; Arkebauer, T.J.; Walters, D.T.; Cassman, K.G.; Dobermann, A. Maize radiation use efficiency under optimal growth conditions. Agron. J. 2005, 97, 72–78, doi:10.2134/agronj2005.0072.
Littleton, E.J.; Dennett, M.D.; Monteith, J.L.; Elstonff, J. The growth and development of cowpeas (Vigna unguiculata) under tropical field conditions. 2. Accumulation and partition of dry weight. J. Agric. Sci. Camb. 1979, 93, 309–320, doi:10.1017/S0021859600037989.
Madden, L.V.; Nutter, F.W., Jr. Modeling crop losses at the field scale. Can. J. Plant Pathol. 1995, 17, 124–135, doi:10.1080/07060669509500703.
Maitree, L.; Toyota, M. A high seed yield and associated attributes of dry matter production achieved by recent Japanese soybean cultivars. Plant Prod. Sci. 2017, 20, 193–204, doi:10.1080/1343943X.2017.1294463.
Meyer, G.E.; Neto, J.C. Verification of color vegetation indices for automated crop imaging applications. Comput. Electron. Agric. 2008, 63, 282–293, doi:10.1016/j.compag.2008.03.009.
Monteith, J.L. Climate and the efficiency of crop production in Britain. Philos. Trans. R. Coc. Lond. B 1977, 281, 277–294, doi:10.1098/rstb.1977.0140.
Morel, J.; Bégué, A.; Todoroff, P.; Martiné, J.F.; Lebourgeois, V.; Petit, M. Coupling a sugarcane crop model with the remotely sensed time series of fIPAR to optimize the yield estimation. Eur. J. Agron. 2014, 61, 60–68, doi:10.1016/0378-4290(93)90017-H.
Morrison, M.J.; Voldeng, H.D.; Cober, E.R. Physiological changes from 58 years of genetic improvement of short-season soybean cultivars in Canada. Agron. J. 1999, 91, 685–689, doi:10.2134/agronj1999.914685x.
Muchow, R.C. An analysis of the effects of water deficits on grain legumes grown in a semi- arid tropical environment in terms of radiation interception and its efficiency of use. Field Crops Res. 1985, 11, 309–323, doi:10.1016/0378-4290(85)90111-X.
Muchow, R.C.; Robertson, M.J.; Pengelly, B.C. Radiation-use efficiency of soybean, mugbean and cowpea under different environmental conditions. Field Crops Res. 1993, 32, 1–16, doi:10.1016/0378-4290(93)90017-H.
Nakaseko, K.; Gotoh, K. Comparative studies on dry matter production, plant type and productivity in soybean, azuki bean and kidney bean. Jpn. J. Crop Sci. 1983, 52, 49–58, doi:10.1626/jcs.52.49.
Phillips, X.A.; Singh, A.K.; Kandel, Y.R.; Mueller, D.S. Effect of pod removal, foliar fungicides, and cultivar on green stem disorder of soybean. Agron. J. 2017, 109, 2680–2688, doi:10.2134/agronj2017.04.0203. 60
Probst, A.H. Border effects in soybean nursery plots. J. Am. Soc. Agron. 1943, 35, 662–666, doi:10.2134/agronj1943.00021962003500080002x.
Santos, J.B.; Procópio, S.D.; Silva, A.A.; Costa, L.C. Capture and utilization of solar radiation by the soybean and common bean crops and by weeds. Bragantia 2003, 62, 147– 153, doi:10.1590/S0006-87052003000100018.
Schöffel, E.R.; Volpe, C.A. Conversion efficiency of photosynthetically active adiation intercepted by soybean for the production of biomass. Braz. J. of Agrome. 2001, 9, 241–249.
Setiyono, T.D.; Weiss, A.; Specht, J.E.; Cassman, K.G.; Dobermann, A. Leaf area index simulation in soybean grown under near-optimal conditions. Field Crop Res. 2008, 108, 82– 92, doi:10.1016/j.fcr.2008.03.005.
Singer, J.W.; Meek, D.W.; Sauer, T.J.; Hatfield, J.L. Variability of light interception and radiation use efficiency in maize and soybean. Field Crops Res. 2011, 121, 147–152, doi:10.1016/j.fcr.2010.12.007.
Smith, G.M.; Milton, E.J. The use of the empirical line method to calibrate remotely sensed data to reflectance. Int. J. Remote Sens. 1999, 20, 2653–2662, doi:10.1080/014311699211994.
Tagliapietra, E.L.; Streck, N.A.; da Rocha, T.S.; Richter, G.L.; da Silva, M.R.; Cera, J.C.; Guedes, J.V.; Zanon, A.J. Optimum leaf area index to reach soybean yield potential in subtropical environment. Agron. J. 2018, 110, 932–938, doi:10.2134/agronj2017.09.0523.
Taiz, L.; Zeiger, E.; Møller, I.M.; Murphy, A. Plant Physiology and Development, 6th ed.; Sinauer Assotiates: Sunderland, MA, USA, 2015; ISBN 1605352551.
Tewes, A.; Schellberg, J. Towards Remote Estimation of Radiation Use Efficiency in Maize Using UAV-Based Low-Cost Camera Imagery. Agronomy 2018, 8, 16, doi:10.3390/agronomy8020016.
Tollenaar, M.; Aguilera, A. Radiation use efficiency of an old and a new maize hybrid. Agron. J. 1992, 84, 536–541, doi:10.2134/agronj1992.00021962008400030033x.
Wang, C.; Myint, S.W. A simplified empirical line method of radiometric calibration for small unmanned aircraft systems-based remote sensing. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2015, 8, 1876–1885, doi:10.1109/JSTARS.2015.2422716.
Wise, K.; Mueller, D. Are fungicides no longer just for fungi? An analysis of foliar fungicide use in corn. APSnet Features 2011, 10, doi:10.1094/APSnetFeature-2011-0531. 61
Woebbecke, D.M.; Meyer, G.E.; Von Bargen, K.; Mortensen, D.A. Color indices for weed identification under various soil, residue and lighting conditions. Trans. Am. Soc. Agric. Eng. 1995, 38, 259–269, doi:10.13031/2013.2783 62
Appendix. Tables
Table A1. Calibration equations for each date of sample collection in 2019 and band for digital number conversion to ln reflectance. Day after planting Band 1z Band 2z Band 3z (DAP) y = -0.0163x + y = -0.0166x + y = -0.0172x + 06-07 (22) 4.9117 5.0490 5.2591 y = -0.0169x + y = -0.0165x + y = -0.0163x + 06-14 (29) 4.1094 4.2062 4.3436 y = -0.0169x + y = -0.0173x + y = -0.0181x + 06-25 (40) 4.7826 5.0625 5.3184 y = -0.0151x + y = -0.0152x + y = -0.0156x + 07-02 (47) 4.4368 4.6308 4.7909 y = -0.0161x + y = -0.0162x + y = -0.0167x + 07-11 (56) 4.7722 4.8563 5.0842 y = -0.0188x + y = -0.0187x + y = -0.0191x + 07-18 (63) 5.8016 5.7713 5.9523 y = -0.0180x + y = -0.0181x + y = -0.0183x + 07-25 (70) 5.4610 5.5604 5.6615 y = -0.0179x + y = -0.0178x + y = -0.0180x + 07-31 (75) 5.4756 5.4708 5.5429 y = -0.0205x + y = -0.0200x + y = -0.0204x + 08-06 (82) 6.2682 6.1831 6.3366 y = -0.0196x + y = -0.0196x + y = -0.0192x + 08-14 (90) 5.9732 6.0580 6.0268 y = -0.0200x + y = -0.0197x + y = -0.0197x + 08-22 (98) 6.1518 6.0740 6.1333 y = -0.0194x + y = -0.0192x + y = -0.0195x + 09-02 (108) 5.9197 5.9224 6.0587 y = -0.0158x + y = -0.0158x + y = -0.0160x + 09-13 (120) 4.1944 4.2840 4.4268 y = -0.0161x + y = -0.0163x + y = -0.0170x + 09-23 (130) 4.9565 5.0610 5.3023 z Calibration equation derived from the linear line of best fit of the negative natural log transformed measured reflectance values and collected average digital number of the calibration targets for the specific band of the collected imagery.
63
Table A2. ANOVA results of all collected variables and the effect of day after planting (DAP), treatment, and their combination from the field experiment in 2019 in Iowa. Variablez DAP Treatment DAP*Treatment FLS <0.001 NS 0.045 DM <0.001 NS NS TLAI <0.001 NS NS
GLAIDx <0.001 NS NS Canopy cover <0.001 0.110 0.025 NDVI <0.001 NS NS Yield --- <0.001 --- z FLS = frogeye leaf spot, DM = dry matter, TLAI = total leaf area index, GLAIDx = green leaf area index with frogeye leaf spot severity removed, NDVI = Normalized Difference Vegetation Index.
64
CHAPTER 3. A COMPARISON BETWEEN DISEASE SEVERITY OF FROGEYE LEAF SPOT AND RADIATION USE EFFICIENCY OF SOYBEAN ACROSS FUNGICIDE TREATMENTS
Abstract
Frogeye leaf spot, caused by Cercospora sojina K. Hara., is a major soybean (Glycine max L. Merr.) disease that has become more prevalent in the upper Midwest. Management techniques include application of foliar fungicides although resistance to certain fungicides has been identified in the pathogen population. Incorporating disease severity into a parameter directly related to yield may better relay the impact of disease on yield and yield components than severity alone. During the 2018 and 2019 growing seasons in fields located in north central and southwestern Iowa, experiments involving fungicides took place to (i) determine how foliar fungicides affect frogeye leaf spot, remotely sensed plant health indicators, and soybean yield and (ii) compare the relationship and impact of foliar fungicides and frogeye leaf spot on radiation use efficiency (RUE) estimated using UAV reflectance data and a reduced sampling approach across foliar fungicide treatments. In two of the three site-years (Kanawha 2018 and
2019), frogeye leaf spot severity was less than 5 and 2%, respectively. Fungicides did not affect frogeye severity or yield and RUE values were statistically similar across treatments at these locations. The most frogeye leaf spot was recorded in Lewis 2018 (14.9%) and yield with an application of flutriafol + fluoxastrobin was 19% greater (P = 0.080) than the non-treated control. In addition, frogeye leaf spot severity of the flutriafol + fluoxastrobin treatment was less than the severity of the fluxapyroxad + pyraclostrobin treatment and non-treated control (P =
0.010). Applications of foliar fungicides increased canopy cover compared to the non-treated control (P = 0.012), but NDVI, SPAD values, and RUE values did not differ between fungicide treatments at all three locations. Estimated soybean RUE values (1.05 to 1.66 g Mj-1) were within 65 the range of known values. Pathogen specific impacts on soybean may confound the variables needed to estimate RUE. Using RUE to estimate the impact of disease on yield remotely may be a valuable resource, however, confounding factors will require additional work to use RUE within certain pathosystems.
Introduction
Frogeye leaf spot of soybean, caused by the pathogen Cercospora sojina K. Hara, is found worldwide with the potential to cause yield loss of up to 60% (Akem et al., 1992; Athow and Probst, 1952; Bernaux, 1979; Dashiell and Akem, 1991; Ma, 1994). C. sojina is an ascomycete that favors warm and humid conditions (Mian et al., 2008). The primary inoculum arises from diseased crop residue and infected seed (Sherwin and Kreitlow, 1952). Frogeye leaf spot is common in the southern United States due to the favorable environmental conditions and mild winters, but northern states such as Iowa, Ohio, and Wisconsin have seen a rise in prevalence (Cruz and Dorrance, 2009; Mengistu et al., 2002; Yang et al., 2001). It has been postulated that this increase in frogeye leaf spot may be attributed to a combination of no-till farming practices and warming global temperatures (Cruz and Dorrance, 2009; Dorrance et al.,
2010; Roth et al., 2020; Wrather and Koenning, 2006; Yang et al., 2001).
Frogeye leaf spot symptoms include characteristic foliar lesions that, when mature, are light brown or tan spots surrounded by a dark reddish-brown margin (Mian et al., 2008; Figure
1A). Lesions can also occur on the stems and pods, but appear different from the distinctive foliar lesions (Westphal et al., 2006; Figure 1B). Infection can occur at any growth stage of the soybean plant. Across the United States from 1996 to 2019, frogeye leaf spot reduced soybean yields an approximately 288,000 metric tons, which equated to an annual loss of $98 million
(Crop Protection Network, 2020). In Iowa, frogeye leaf spot caused losses of over 43,500 metric tons and $14.5 million annually from 1996 to 2019. In Iowa during 2018, frogeye leaf spot 66 caused an estimated loss of over 360,000 metric tons and over 140,000 metric tons in 2019, which equates to over $119.4 million ($12.01 a-1) in 2018 and $46.5 million ($5.06 a-1) in 2019
(Crop Protection Network, 2020).
Management strategies for frogeye leaf spot include planting resistant cultivars and applying foliar fungicides (Cruz and Dorrance, 2009; Dorrance et al., 2010; Mengistu et al.,
2014). However, fungicide resistant strains of C. sojina have been identified, first in Tennessee in 2010 (Zhang et al., 2012). Since then, it has been reported in isolates collected from Alabama,
Arkansas, Delaware, Illinois, Indiana, Iowa, Kentucky, Louisiana, Mississippi, Missouri, North
Carolina, Ohio, Tennessee, Virginia, and South Dakota (Mathew et al., 2019; Zhang et al.,
2018).
In the absence of disease, certain fungicides have been promoted to increase yield and improvements in yield have been shown, though inconsistently (Bradley and Sweets, 2008;
Kandel et al. 2016; Kyveryga et al., 2013; Wise and Mueller, 2011). Yield increases may be attributed to a phenomenon called the “greening effect”, which is described to extend the duration of green leaf area, maintaining photosynthetic efficiency, and allowing continuation of dry matter accumulation (Balba, 2007; Bartlett et al., 2002; Kumudini et al., 2001; Morrison et al., 1999). Joshi et al. (2014) reported application of pyraclostrobin benefited the formation of root nodules; nitrogen fixation was enhanced, growth improved and yields were greater. Amaro et al. (2019) concluded that the physiological effects associated with an application of strobilurin fungicide are due to an increase in net photosynthesis. In addition, applications of pyraclostrobin delay senescence by impacting hormonal levels (Amaro et al., 2019).
Yield loss caused by frogeye leaf spot is primarily attributed to the reduction of green leaf area and premature defoliation that occurs when severity is greater than 30% (Dashiell and 67
Akem, 1991). Reducing the photosynthetic leaf area can result in a reduced seed weight and consequently less yield (Dashiell and Akem, 1991). Historically, disease severity and the area under the disease progress curve (AUDPC) have been used to measure the impact of disease severity on yield (Filho et al., 1997; Kumudini et al., 2008; Madden and Nutter, 1995). However, disease severity is not directly related to yield and the resulting relationship between AUDPC and yield can be inconsistent (Carretero et al., 2010; Serrago et al., 2009; Waggoner and Berger,
1987). Instead, yield has more consistently been predicted by healthy leaf area duration and the absorption of solar radiation (Bryson et al., 1997; Leite et al., 2006; Waggoner and Berger,
1987).
Yield and green biomass have been correlated with various indices using remote sensing
(Genovese et al., 2001). Handheld active sensors such as a GreenSeeker (Trimble, Sunnyvale,
CA, USA) and a Soil Plant Analysis Development (SPAD) meters (Konica-Minolta, Japan) are commonly used to measure plant health and crop nitrogen status (Yuan et al., 2016). These sensors measure transmittance and reflectance of specific regions of the magnetic spectrum. Raw
SPAD values are proportional to the chlorophyll in the leaf and have been used to determine the efficacy of nitrogen treatments (Fox et al., 1994; Kaakeh et al., 1992; Piekielek and Fox, 1992;
Turner and Jund, 1991; Uddling et al., 2007; Wood et al., 1992;). The normalized difference vegetation index (NDVI) is a ratio between the difference and sum of red and near infrared frequencies. It has been used to monitor health and estimate yield (Bolton and Friedl, 2013;
Devadas et al., 2009).
A parameter that incorporates yield related measurements in addition to disease severity may better characterize the impact of the disease. Radiation use efficiency (RUE) is a measure of the relationship between accumulated biomass and intercepted photosynthetically active 68 radiation (PAR) (Ceotto and Castelli, 2002; Monteith, 1977; Muchow et al., 1993; Warren-
Wilson, 1967). Madden and Nutter (1995) suggested that using the healthy leaf area and RUE could quantify the effects of disease on yield. Foliar disease may be accounted for by removing the severity from the green leaf area index (GLAI) by converting the severity rating to a 0 - 1 scale and multiplying it by the GLAI (Madden and Nutter, 1995). There is no standard method of estimating RUE. Gitelson and Gamon (2015) proposed that the standard method should be the radiation absorbed by photosynthetically active tissue (RUEgreen). However, by only considering the green leaf area in the RUE estimation, it becomes sensitive to inevitable pigment changes during senescence (Gitelson and Gamon, 2015). Tewes and Schellberg (2018) used the RUEgreen method to estimate RUE of corn. This method was first used to estimate the RUE of soybean using reflectance data from unmodified consumer grade imagery captured by an unmanned aerial vehicle (UAV) (Philips et al. 2020).
UAVs are increasingly being used in the agricultural setting to monitor water status, plant vigor, biomass, and diseases (Adão et al., 2017; Bendig et al., 2015; Calderón et al., 2015; Park et al., 2015; Primicerio et al., 2012). UAVs can be used to capture imagery at high spatial and temporal scales across large areas in a single flight (Beloev, 2016). RGB and multispectral imagery are the most utilized forms to collect disease information (Altas et al., 2018; Calderón et al., 2014; Kerkech et al., 2018; Sugiura et al., 2018; Tetila et al., 2017; Zhang et al., 2017).
Soybean RUE determined on a small plot scale with high sampling frequency under low disease conditions by Phillips et al. (2020) was within the range of previously published RUE values (Muchow, 1985; Santos et al., 2003). Since RUE is directly associated with aboveground biomass (Monteith, 1977), we propose RUE may be a better parameter to monitor the impact of diseases and their management. The objectives of this study were to (i) determine how foliar 69 fungicides affect frogeye leaf spot, remotely sensed plant health indicators, and soybean yield and (ii) compare the relationship and impact of foliar fungicides and frogeye leaf spot on RUE estimated using UAV reflectance data with a reduced sampling approach across foliar fungicide treatments.
Materials and Methods
Field experiments were completed near Kanawha and Lewis, IA in 2018 and Kanawha in 2019. Experiments were laid out in a randomized complete block design with three treatments and four replications. Plots were 7.8-m long, and four rows wide spaced 76-cm apart. The three treatments were 1) non-treated control, 2) fluxapyroxad + pyraclostrobin (Priaxor©, FRAC group 7 + 11, 0.05-L ha-1, BASF, Research Triangle Park, NC) and 3) flutriafol + fluoxastrobin
(Preemptor©, FRAC group 3 + 11, 0.06-L ha-1, FMC Agricultural Solutions, Philadelphia, PA).
Cultivars and other experimental details are given in Table 1. Fungicides were applied at the R3 growth stage (beginning pod; Fehr and Caviness, 1977) at a rate of 0.05 and 0.06-L ha-1 active ingredient for the fluxapyroxad + pyraclostrobin and flutriafol + fluoxastrobin treatments, respectively. All four rows were sprayed with a self-propelled research sprayer that delivered
-1 fungicides using XR 11002 nozzles with 142 to 189 L ha at 241 kpa powered by a CO2 tank.
Field data were collected every 6-16 days beginning at R3 growth stage, beginning pod
(Fehr and Caviness, 1977). At both locations in 2018 there were four sampling dates (SDs) and six SDs in Kanawha in 2019. Each SD was paired with a growing degree day (GDD) sum (Table
1). The GDD was calculated with a base of 10°C (http://mesonet.agron.iastate.edu). GDD values were summed beginning at the planting date of the specific site year, through the calendar date of the field sampling. Sampling dates across locations were grouped based on the GDD of the respective SD and is referred to as the growing degree day group (GDDG) (Table 1). Six
GDDGs (GDDG-1 to GDDG-6) were created. The range between GDDs within GDDG-1 was 70
1280 to 1298, 1542 to 1560 for GDDG-2, 1808 to 1858 for GDDG-3, 2140 to 2243 for GDDG-
4, 2344 to 2711 for GDDG-5, and 2832 for GDDG-6.
During each sampling, a 0.5-m section from one of the non-yield rows was marked for data collection. This section was kept 0.5-m away from row ends and previously sampled areas to avoid edge effects (Probst, 1943). From this section of row, frogeye leaf spot severity data,
NDVI values, SPAD readings, plant count, and destructive plant samples were collected.
Frogeye leaf spot severity (0-100%) was determined by examining all plants in the 0.5-m sampling section. Both the upper and lower canopies were rated and averaged for a single plot estimation because the leaf area was determined from all of the plants leaves. The severity recorded on the last SD respective to each site year was used in the analysis.
SPAD readings were collected with a SPAD-502 meter (Konica-Minolta, Japan) in 2018 and a CCM-200 (Opti-Sciences, Hudson, NH, USA) in 2019. The SPAD-502 meter uses the
940nm and 650nm wavelengths while CCM-200 uses 940nm and the 660nm wavelengths to measure the Ratio Vegetative Index (RVI). The reading frame of the two devices differ. The area of reading is 0.06 and 0.71-cm2 for the SPAD-502 and CCM-200, respectively. This was done by taking 10 readings on the center leaf of a mature, fully expanded trifoliate from the upper canopy. We avoided taking readings directly on frogeye leaf spot lesions. Readings were averaged to a single value. A handheld GreenSeeker (Trimble, Sunnyvale, CA, USA) was used to measure NDVI of the section of row. The 0.5-m section was measured three times and averaged to a single NDVI rating per SD. Five plants were then destructively sampled by cutting every other plant at the soil line. The five plants were bundled in a plastic bag, and stored at 4°C until further processing, 2 to 10 days later. 71
Leaves were grouped as, “green” and “non-green” to obtain separate green and yellow leaf areas. If a leaf was over 50% yellow for any reason, it was placed into the “non-green” group. All of the leaves from the bundle of five plants were processed to measure a “green” and a “non-green” leaf area. A LI-3100C (LI-COR, Inc., Lincoln, NE, USA) was used to measure the leaf area. Leaf area was converted to m2 and used to estimate leaf area index (LAI; m2 m-2) by using the number of plants sampled and the total number of plants in the sampled area (Maitree and Toyota, 2017). The sum of the green leaf area index (GLAI) and the yellow leaf area index
(YLAI) from the “non-green” group, was the total leaf area index (TLAI). The frogeye leaf spot severity rating of each respective date was used to reduce the GLAI and determine the green leaf area index with frogeye leaf spot severity removed (GLAIDx). This was used for all date specific needs of GLAI.
All leaves, stems, and pods from each plot were placed in brown paper bags and dried for seven days in custom drying bins with fans held at a constant temperature of 60°C. Sample dry matter (DM) was measured for each bag and adjusted to exclude bag weight (DM; Taylor
Precision Products, Oak Brook, IL, Glass Digital Food Scale, model: 3842BL9). Thus, the DM determination at all SDs consisted of the entire plant including seeds.
Images in 2019 were collected with a Phantom 4 Pro (DJI, Shenzhen, China) in 2018 and a Phantom 4 ProV2.0 (DJI, Shenzhen, China), both with an onboard 2.54-cm CMOS 20 MP camera. Whole field images were captured in 2018 at an aboveground altitude of 120-m, which provided images with an estimated ground sampling distance (GSD) of 3.3-cm per pixel. Images in 2019 were captured at 9-m AGL and stitched together using a flight plan and stitching service from DroneDeploy (DroneDeploy, San Francisco, CA) with an estimated GSD of 0.69-cm. The 72 images and orthomosaics were processed in ArcGIS (ArcGIS Pro 2.4.2, Esri, Redlands, CA,
USA).
Three 1.2-m portable fabric targets (Group 8 Technology, Inc, Provo, UT) with reflectance values of approximately 3, 13, and 56% were placed at the field edge during each flight for calibration from digital number (DN) to reflectance. The DNs of each target of all three bands were extracted and averaged for each sampling date. The raw target DN vs. reflectance plot was an exponential relationship, similar to Wang and Myint (2015), and opposed to the linear relationship described by Smith and Milton (1999). A linear relationship was obtained using a natural logarithm transformation of the reflectance values for the calibration targets
(Wang and Myint, 2015). A line of best fit was added to the scatter plot and the resulting linear regression equation was used to convert the DNs of each band to -ln(reflectance). This was converted to reflectance using the exponential term (Wang and Myint, 2015).
In ArcGIS (ArcGIS Pro 2.4.2) the three band orthomosaics were converted into a single raster layer using the raster calculator and the excess green (ExG) index (Woebbecke et al.,
2005). The resulting raster was examined manually to identify a threshold used to binarize the image and separate foliage from background (Meyer and Neto, 2008).
A threshold DN value ≥ 31 was identified to accurately separate foliage from background. The pixels of the image were classified as either “foliage” (1) or “other” (0). The
“foliage” class contained both yellow and green soybean foliage. The “other” class contained bare soil, indecipherable shadowed area, and unknown pixels. A 100-point accuracy assessment was completed on each date of image collection. The kappa values were between 0.85 and 1.00.
The relatively low kappa values came from the sampling dates later in the season when it was 73 difficult to decipher between yellow stems, fallen leaves, and yellow leaves still attached to the soybean plant in the reference image.
All four rows of the plot were included in the image analysis. The binary raster was used as a mask to extract only the DNs of the “foliage” class from the three collected bands (red, green, and blue) of the respective images and orthomosaics. The fraction of foliar cover (fCover) was determined by extracting the sum of the areas of each class and determining the percent of the “foliar” class with respect to the combined total area. Two tools were created in ArcGIS
ModelBuilder (ArcGIS Pro2.4.2) to streamline and partially automate the iterations of image analysis.
The first model, RUE2019, was comprised of 12 individual tools, used repeatedly for a total of 32 tools and four input features. The model was comprised of four general steps in the image analysis workflow. The first step was titled COMBINE BANDS (Figure A1) where the three individual bands were copied and converted to a tif format, updated with an attribute table, and combined into a single raster with data for each band in the raster’s table.
The next step was titled RASTER CALCULATOR (EXG) (Figure A1). In this step, the three, individual updated rasters were used to create a new raster using the ExG index. The raster calculator tool was again used to binarize the ExG raster, separating green pixels from non- green, background pixels. This binary raster and the combined bands raster from the previous model fed into the next group titled EXTRACT TO EXCEL (Figure A1). In this step, the shape file delineating the area of interest was used to clip the rasters and as input for tools downstream that extracted the DNs and pixels from the desired features. This included the extract by attributes tool to pull the DNs or the area of grouped classes for each plot into a table. 74
All of the individual tables were sequentially joined and exported to Excel in the final step EXCEL (Figure A1). Parameters were set in the model to create a functioning, reusable geoprocessing tool (Figure A2). The parameters were designated with a “P” and can be seen throughout the whole model (Figure A1). Designating a variable as a parameter allows it to be included in the model tool dialogue box (Figure A2). The input variables set to parameters for this model were bands 1, 2, 3, and the plots shape file (Figure A2). The binary ExG raster and the Excel file were set as parameters to maintain a check and direct the location where the Excel file should be saved, respectively (Figure A1). All other processes were considered intermediary and their outputs were not maintained.
The second tool, Targets, was created to extract the DNs from the specified reflectance targets present in the image and export them to Excel for calibration. It was comprised of eight unique tools and six input variables, while the whole model had 37 tools. This model had four steps, the first being the same as the RUE2019, COMBINE BANDS. In this step the three individual bands were converted to a tif format, created and joined to an attribute table, (Figure
A3). The next three steps were iterations corresponding to the specific targets (Black, Gray, and
White). The workflow for the whole model can be seen in Figure A3. After the bands were combined, the area of the target was clipped and the DNs of each respective band were joined and exported to Excel (Figure A3). In the model, nine inputs and outputs were set as parameters to show up in the dialogue box of the corresponding tool. These parameters were the three bands of the raster, the three calibration targets, and the resulting output Excel files (Figure A4).
RUEgreen (Eq. 1) was derived by estimating the total fraction of absorbed photosynthetically active radiation (fAPARtotal) using Eq. 2 and the fAPAR of the green leaf area
(fAPARgreen) using Eq. 3. The fAPAR of the soil was omitted from Eq. 2 (Lindquist et al., 2005; 75
Phillips et al., 2020; Tollenaar and Aguilera, 1992). The total incoming solar radiation was captured by a weather station near the location of the experiment. A running total of the incoming radiation began the first date of data collection for each location. The incoming photosynthetically active radiation (PARinc) was determined by multiplying the total radiation by
0.5 (Taiz and Zeiger, 2014). The fraction of photosynthetically active radiation passing through the canopy (fPARtrans) was estimated using the Lambert-Beer law equation and an extinction coefficient (k) of 0.523 (Flènet et al., 1996; Phillips et al., 2020) based on crop and row spacing.
In addition, the frogeye leaf spot severity was removed from the GLAI on each SD (Tewes and
Schellberg, 2018; Phillips et al., 2020). The inner two rows of each plot were harvested for yield using a 2009 Almaco SPC20 research plot combine (ALMACO, Nevada, IA). Yield calculations were adjusted to 13% seed moisture.
Weather data for site years was obtained from the Iowa Environmental Mesonet
(http://mesonet.agron.iastate.edu). The summer month selection of June–August was compared to the 30-year average for total precipitation and average temperature. In addition, September was included for total precipitation and average temperature and also compared to the 30-year average.
퐷푀 Equation 1. 푅푈퐸푔푟푒푒푛 = × 푃퐴푅푖푛푐 푓퐴푃퐴푅푔푟푒푒푛
Equation 2. 푓퐴푃퐴푅푡표푡푎푙 = (100% − 푓푃퐴푅표푢푡 − 푓푃퐴푅푡푟푎푛푠) × 푓퐶표푣푒푟 퐺퐿퐴퐼 Equation 3. 푓퐴푃퐴푅 = 푓퐴푃퐴푅 × ( ) 푔푟푒푒푛 푇퐿퐴퐼 Statistical analyses were performed in SAS 9.4 (SAS Institute, Cary NC). Analysis of variance was performed for all variables. A mixed model procedure was fitted using Proc
Glimmix. Treatment was set as a fixed factor while replication, GDDG, and site year were considered random factors. Replication was nested within date during analysis within site years for variables with multiple sampling times. When analysis combined site years, replication was 76 then nested within a GDDG. Least square means (LS-means) were estimated using the LSmeans statement. Means were separated using “lines” options in the LSmeans statement at 10% level of significance (α = 0.10). Proc corr procedure was used to correlate frogeye leaf spot, RUE and yield. The Lewis 2018 site year showed significant errors across all treatments throughout the first replication, and thus was removed from the location for analysis.
Results
Yield varied across all site years and treatments at Lewis 2018 (Table 2; Table 3). Across site years, yield was greatest in Lewis 2018, approximately 22% more than the 3,174.2 kg ha-1 yield in Kanawha 2018 (Table 3). The flutriafol + fluoxastrobin treatment had the greatest yield, which was 9% more than the non-treated control. Yields of the Fluxapyroxad + pyraclostrobin treatment and non-treated control were statistically similar. Treatment × site-year interaction was significant for yield (P = 0.08) thus we analyzed yield by site-year. Treatment difference was significant (P = 0.059) in Lewis 2018 only, although similar trends were observed in other locations. In Lewis 2018, soybeans treated with flutriafol + fluoxastrobin produced 19% more yield than the non-treated control (Table 4).
Frogeye leaf spot severity differed across the site-years. Severity of 15% at Lewis 2018 was approximately three and ten times more than severity at Kanawha 2018 and 2019, respectively (Table 3). Both treatment (P = 0.005) and the treatment × site year interaction (P =
0.010) were significant for frogeye leaf spot severity (Table 2). The non-treated control had the most frogeye leaf spot, around 9% severity averaged, while the fluxapyroxad + pyraclostrobin and flutriafol + fluoxastrobin treatments averaged around 7 and 5%, respectively, over all site- years (Table 3). Overall, the frogeye leaf spot severity of the flutriafol + fluoxastrobin treatment was significantly less than the severity of the fluxapyroxad + pyraclostrobin treatment (Table 3.)
In Lewis 2018, the flutriafol + fluoxastrobin treatment reduced frogeye leaf spot severity by over 77
53% compared to the non-treated control (Table 4). However, in Kanawha 2018 and 2019, frogeye leaf spot severity across treatments did not differ, although there was numerically less disease in the fungicide treatment (Table 4).
Mean RUE values were 1.27 to 1.38 -g Mj-1 across the site years (Table 3) and did not differ across site year or treatment (Table 2). RUE values for the treatments were 1.19, 1.43, and
1.33-g Mj-1 for the non-treated control, fluxapyroxad + pyraclostrobin treatment, and flutriafol + fluoxastrobin treatment, respectively (Table 3). Although not significantly different, RUE values were numerically greater for the fungicide treatments.
The effect of GDDG and the GDDG × site year interaction was significant for plant count
(Table 2). Plant count of the measured section of row was greatest at GDDG-1 (GDD ranged from 1280 to 1298) compared to all subsequent groups (Table 2). The plant count across all groups ranged from an average of 11.0 to 13.2 plants (Table 5). Treatment did not affect the plant count in any GDDGs.
The effect of the GDDG was significant for dry matter, but treatment had no effect (Table
2). The dry matter within the 0.5-m section of row was 98.6 g in GDDG-1, increased with growing degree days (Table 5) and peaked on GDDG-6 (GDD 2832) at 365.0 g (Table 5). Lewis
2018 was the only site year to have significantly different yield and frogeye leaf spot severity across treatments. In Lewis 2018, dry matter was significantly less on the initial date of data collection, GDDG-3 (P = 0.016; Table 6). However, the following three sampling periods were not significantly different with regards to dry matter and decreased numerically after GDDG-4
(Table 6).
Canopy cover varied across treatments and GDDGs (Table 2). Both fungicide treatments resulted in significantly greater (P = 0.012) canopy cover compared to the non-treated control, 78 which had 83% CC, by approximately 5 and 8% for the fluxapyroxad + pyraclostrobin and flutriafol + fluoxastrobin treatments, respectively (Table 5). In general, the canopy cover peaked on GDDG-4, at 96.7% and decreased in following GDDGs (Table 5). The canopy cover for
Kanawha 2018 was significantly less than that of Kanawha 2019 in GDDGs-1 and 2 (Table 7).
Treatment did not have a significant effect on either TLAI or GLAIDx (Table 2). Among the treatments, TLAI ranged from 2.8 to 3.0 while GLAIDx ranged from 2.5-2.6. Effect of GDDG was significant for both TLAI and GLAIDx. TLAI and GLAIDx both plateaued on GDDG-3 at 4.0 and 3.8, respectively (Table 5). The leaf area begins to decline following GDDG-3. The overall values of TLAI and GLAIDx for Kanawha 2018 did not surpass 2.5 and 2.2, respectively (Table
8). The Kanawha 2019 location had the highest TLAI value on GDDG-2 and GDDG-3 (around
5.2), approximately 16% more than the next highest value from Lewis 2018 on GDDG-3 (Table
8).
Fungicide treatment did not significantly affect NDVI or SPAD (Table 2). Overall, the non-treated control combined across all site years had 0.78 NDVI and 32.7 SPAD. GDDG significantly impacted both variables. NDVI was 0.87 and SPAD was 40.0 in the first sampling.
Both numerically plateaued on GDDG-2 with NDVI 0.89 and SPAD 40.8 (Table 5). NDVI values began to significantly decline after GDDG-3 (NDVI value 0.88), whereas SPAD values began to diminish after GDDG-4 (SPAD reading 34.6) (Table 5).
Correlations among yield, frogeye leaf spot severity, and RUE were not significant
(Table 9). Correlations coefficients ranged from -0.17 to 0.15.
During June to August in Kanawha 2018 and Lewis 2018, total precipitation was approximately 39 and 47% greater than the 30-year average, respectively
(http://mesonet.agron.iastate.edu; Table 10). However, precipitation in Kanawha 2019 was 79 approximately 20% less than the 30-year average. During the month of September, total precipitation was 3 and 2 times greater than the 30-year average for Kanawha and Lewis in 2018, respectively (Table 10). The average temperature of all site years from June-August did not vary by more than ±2% from the 30-year average (Table 10). It was slightly warmer in 2018 site years and roughly 2% warmer in Lewis compared to Kanawha (Table 10).
Discussion
RUE of soybean was estimated on a small plot scale as described previously (Phillips et al. 2020). This study was done to determine if RUE could be used to monitor the impact of diseases and their management, specifically frogeye leaf spot and foliar fungicide applications on soybean yield. All estimated RUE values in this study fell within the range of previously reported values of 0.60 to 2.53 (g Mj-1) (Muchow, 1985; Phillips et al. 2020; Santos et al., 2003).
In our studies, RUE did not differ between field trial locations or among treatments within locations, even though frogeye leaf spot and yield did.
Across site years, Lewis 2018 had the greatest numerical value of RUE, the greatest average yield, and the most frogeye leaf spot severity. However, RUE values did not differ significantly among treatments, although the fungicide treatments trended to have greater RUE than the non-treated control.
The severity levels of frogeye leaf spot were removed from the GLAI on each date. We do not know the level of frogeye leaf spot severity that starts to affect yield. Previous studies have shown that fluxapyroxad + pyraclostrobin provided a yield benefit of 6% at frogeye leaf spot severity levels of approximately 11%, while flutriafol reduced severity by 81% and provided a yield benefit of 10% compared to the non-treated control that had approximately 20% severity (Mengistu et al., 2018). By removing the frogeye leaf spot severity, the GLAI and subsequently the APAR were reduced during early sampling dates. But, if the level of severity 80 was not impacting DM, this could bolster early sampling dates on the regression plot and potentially wash out effects when severity reaches yield reducing levels at later dates. Srinvasan et al. (2017) showed reducing LAI, increased RUE, if DM was maintained. In Lewis 2018, it was expected to have difference in RUE across the treatments because of high disease severity and the influence of treatments on disease severity and yield, however the compounding effect of experimental error and reduced number of sampling units to derive RUE may have contributed to the lack of difference between treatments. Although, DM and GLAI, the components needed for
RUE derivation, collected from the field did not progress across collection dates. It is possible that the time sampling duration and the interval between sampling dates may have contributed to the lack of progress measured in DM and GLAI. At this location, the first date of sampling occurred seven days after the treatment application, at R4 growth stage, full pod (Fehr and
Caviness, 1977), while at the other locations, sampling began prior to R4. The reduced component sampling resulted in RUE values within the range of previously published values.
Additional collection dates before and after R4 and a larger number of plants should be sampled to better account for error.
In this study, there were significant differences across site years with regards to frogeye leaf spot severity and yield, which may have been related to weather. Both locations in 2018 received substantially more precipitation during the summer months than the 30-year average. In contrast, the 2019 location received less precipitation compared to 2018 locations. Temperatures were slightly warmer in Lewis 2018. The ideal environmental conditions for frogeye leaf spot development is 27°C with 72 hours of leaf wetness (Camera et al., 2016). We did not evaluate initial pathogen levels, but it is reasonable to assume that the warmer and humid weather in 81
Lewis location was more favorable for frogeye infection and development than the other site- years.
While both fluxapyroxad + pycraclostrobin and flutriafol + fluoxastrobin have been shown to reduce frogeye leaf spot (Mengistu et al., 2018; Mueller et al., 2019), there were differences in their effectiveness in these studies. Severity of frogeye leaf spot when treated with flutriafol + fluoxastrobin was on average 50% less than that of the fluxapyroxad + pyraclostrobin treatment. Similarly, Mengistu et al. (2018) reported the active ingredient flutriafol (Topguard©,
FRAC group 3) controlled frogeye leaf spot better than fluxapyroxad + pyraclostrobin and difenconazole + azozystrobin (Quadris TOP SBX©, FRAC group 3 + 11) in some years. QoI fungicide resistance in isolates of C. sojina collected from Iowa has been reported (Zhang et al.
2018). Mengistu et al. (2018) cautioned additional pressure on the pathogen population to develop DMI resistance exists, when resistance to QoI is likely already present.
At Lewis 2018, yield was protected by application of flutriafol + fluoxastrobin applied at
R3. The predominant disease recorded in these soybeans was frogeye leaf spot, however,
Septoria brown spot was also present. The yield and frogeye leaf spot severity of the fluxapyroxad + pyraclostrobin treatment did not differ from the non-treated control.
QoI containing fungicides promoted to have plant health benefits provided varying levels of protection to frogeye leaf spot, while failing to impact yield at low disease levels (Bradley and
Sweets, 2008; Kandel et al. 2016; Kyveryga et al., 2013; Mengistu et al., 2018; Wise and
Mueller, 2011). In agreement with Swoboda and Pedersen (2009), the fungicide application did not benefit yield at low levels of disease. Swoboda and Pedersen (2009) did see that the total biomass was increased by 10% with a pyraclostrobin treatment, mainly due to the increased stem 82 weight. Phillips et al. (2017) also found that pyraclostrobin + fluxapyroxad applied at R3 increased the DM of stems at harvest.
The promoted yield benefit in soybean by prophylactic QoI fungicide application is inconsistent (Bradley and Sweets, 2008; Kandel et al. 2016; Kyveryga et al., 2013; Phillips et al.,
2020; Wise and Mueller, 2011). There are various explanations for how a QoI fungicide application might improve yield. Amaro et al. (2019) concluded that the physiological benefits to healthy plants of QoI fungicide applications are due to an increase in net photosynthesis. In our study, canopy cover of both fungicide treatments was greater compared to the non-treated control. However, the increase in net photosynthesis that may have occurred due to the additional canopy cover did not result in yield increases under the low disease conditions in our study.
Additionally, in our studies canopy cover was approximately at 85% at R3 in 2018 and 2019 in
Kanawha. Since full canopy cover by or soon after R1 is needed for maximum yield (Ball et al.,
2000; Board, 2000; Purcell et al., 2002), our research suggests that if full canopy cover is not attained by flowering, a yield response associated with a fungicide application may not occur.
The application of QoI fungicides can increase leaf greenness, chlorophyll content, and delayed senescence (Bryson et al., 2000; Grossmann et al., 1999). No effect of treatment for
NDVI values across treatments was detected in any site year under low frogeye leaf spot conditions. Even at the Lewis 2018 location, where frogeye leaf spot severity was significantly different across treatments, no treatment difference was observed for NDVI. Similarly, no differences in SPAD readings among treatments in all site years were detected. Since the exact reading frame cannot be seen when taking the SPAD value, and SPAD meters used here cover a very small area (0.06-0.71 cm2) of the leaf, we cannot be certain of the condition of the measured tissue. The SPAD-502 meter readings are known to be reliable indicators of leaf photosynthetic 83 rate (Ma et al., 1995). With no differences between treatments and the readings from the active sensors, perhaps the impact of frogeye leaf spot on the leaf is local to the lesion, as opposed to the negative impact the soybean rust fungus, Phakopsora pachyrhizi, has on non-lesion green leaf area (Kumudini et al. 2008).
The final sampling date across locations showed reduced LAI and similar dry matter compared to previous dates. Muchow et al. (1993) describes a decline in RUE prior to maturity that was attributed to a combination of a loss of biomass with leaf drop and a decline in specific leaf nitrogen. The declining parameters seen on the final sampling dates in this study could have been the start of RUE decline.
Yield at Kanawha 2018 was 18% less yield than the yield at Lewis 2018, despite the same cultivar, planting population, and significantly less frogeye leaf spot severity. This was evidenced by the fact that Kanawha 2018 showed nearly half TLAI and GLAIDx of Lewis 2018, across multiple GDDGs. Differences in LAI values also will affect RUE values. Srinivasan et al.
(2017) modeled that the optimal LAI of a modern U.S. Midwest soybean cultivar under current
CO2 was 4.2. TLAI values in Lewis 2018 and Kanawha 2019 were above the optimal value identified by Srinivasan et al. (2017), but below the optimal value by approximately 43% in
Kanawha 2018. It was shown that decreasing their observed LAI of 6.8, to the optimum model values, would result in an increase of yield (Srinivasan et al., 2017). Srinvasan et al. (2017) also describes an increase in RUE to be attributed to the maintenance of DM and diminishing APAR caused by reduced LAI. In addition, Srinivasan et al. (2017) found the RUE was 9% greater at the optimal modeled LAI than their observed value, despite the modeled LAI value being approximately 40% less than the observed. 84
Conclusion
RUE is capable of being calculated on a small plot scale with the help of UAV reflectance data. The foliar fungicides used in this study resulted in increased canopy cover compared to the non-treated control. However, this additional canopy cover did not result in a significant yield benefit in two of the three locations. The yield benefit of a fungicide application was seen at one location that had greater levels of frogeye leaf spot severity. No other physiological advantage was detected by applications of foliar fungicides at low frogeye leaf spot disease pressure. RUE values, NDVI, and SPAD values, however, failed to capture the protection that fungicides provided yield against frogeye leaf spot. The level of severity that frogeye leaf spot begins to impact yield is unknown. Since soybean produces an overabundance of leaves, non-limiting levels of frogeye leaf spot could result in greater RUE estimations if the dry matter produced remains the same. This highlights the issues of understanding how diseases affect soybean yield. Accounting for disease severity when it is not impacting yield may result in an inflated RUE value. Using RUE to estimate the impact of disease on yield remotely will be a valuable resource, however, confounding factors (accounting for disease, component collection method) will require additional work to use RUE within certain pathosystems.
Tables and Figures
Table 1. Location, cultivar, and other details of the field experiments evaluating the impact of frogeye leaf spot and fungicide treatments on radiation use efficiency. Site year Soil type/ Cultivar Planting Planting Fungicide Sampling Dates Harvest Slope population date applicationx (GDD;GDDG)yz date Kanawha 2018 Nicollet S28-N6 370,000 29 May 30 July 25 Jul (1280;1), 9 Aug 23 Oct clay loam / (1560;2), 21 Aug 1 to 3 (1808;3), 12 Sep (2204;4) Lewis 2018 Marshall S28-N6 370,000 9 May 23 July 30 Jul (1858;3), 15 24 Oct silty clay Aug (2243;4), 6 Sep loam / 2 to 5 (2711;5), 13 Sep 85
(2832;6) Kanawha 2019 Nicollet S29-K3X 395,200 16 May 22 July 25 Jul (1298;1), 6 Aug 26 Oct clay loam / (1542;2), 22 Aug 1 to 3 (1824;3), 2 Sep (1975), 13 Sep (2140;4), 23 Sep (2344;5) x Fungicides were applied at (R3), beginning pod development (Fehr and Caviness, 1977). y Sum of growing degree days (GDD) with a base of 10° Celsius beginning at the planting date. z GDDG = growing degree day group.
Table 2. Analysis of variance for different variablesz to test the effect of site year, treatment, growing degree day group (GDDG), and their combinations in field experiments conducted in Kanawha and Lewis county Iowa in 2018 and 2019.
Effect Yield Frogeye RUE PC DM CC TLAI GLAIDx NDVI SPAD leaf spot Site year 0.035 <0.001 0.773 0.212 <0.001 0.185 <0.001 <0.001 <0.001 <0.001 Treatment 0.005 0.005 0.342 0.594 0.908 0.012 0.764 0.689 0.431 0.773 Site year × treatment 0.080 0.010 0.462 0.824 0.552 0.987 0.378 0.477 0.952 0.860 GDDG 0.039 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
86
Site year × GDDG <0.001 0.429 <0.001 <0.001 <0.001 <0.001 <0.001 Treatment × GDDG 0.908 0.959 0.024 0.824 0.885 0.644 0.297 Site year × treatment × GDDG 0.968 0.991 0.984 0.982 0.945 0.999 0.878 z RUE = radiation use efficiency; PC = plant count; DM = dry matter; CC = canopy cover; TLAI = total leaf area index; GLAIDx = green leaf area index with disease severity removed; NDVI = normalized difference vegetation index; SPAD = soil plant analysis development. 87
Table 3. Least square meansy of yield, frogeye leaf spot severity, and radiation use efficiency (RUE) recorded in field experiments in Kanawha and Lewis county Iowa in 2018 and 2019. Site year Treatmentz Yield Frogeye leaf spot RUE (kg ha-1) (%) (g Mj-1) Kanawha 2018 3,174.2 b 4.5 b 1.29 Lewis 2018 3,880.3 a 14.9 a 1.38 Kanawha 2019 3,840.0 a 1.5 c 1.27 NTC 3,503.7 b 8.9 a 1.19 Fluxapyroxad + pyraclostrobin 3,571.0 b 7.0 b 1.43 Flutriafol + fluoxastrobin 3,826.5 a 4.9 c 1.33 y Means followed by the same letter for site years and by the same letters for treatments were not statistically different at α = 0.1. Means were separated using “Lines” option in the LSMEANS statement. z Treatment = Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF, Triangle Park, NC), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions, Philadelphia, PA)
88
Table 4. Least square meansx of yield, frogeye leaf spot severity, and radiation use efficiency (RUE) for different fungicide treatments in three locations in Iowa during 2018 and 2019. Site year Treatmenty Yieldz Frogeye leaf RUE (g Mj-1) (kg ha-1) spot (%) Kanawha 2018 NTC 3,086.8 4.8 1.37 Fluxapyroxad + 3,214.6 3.8 1.25 pyraclostrobin Flutriafol + fluoxastrobin 3,221.3 4.6 1.26 P value 0.797 0.191 0.890 Lewis 2018 NTC 3,631.5 b 19.6 a 1.05 Fluxapyroxad + 3,725.7 b 15.8 a 1.66 pyraclostrobin Flutriafol + fluoxastrobin 4,310.7 a 9.3 b 1.43 P value 0.008 0.059 0.234 Kanawha 2019 NTC 3,806.4 2.5 1.15 Fluxapyroxad + 3,779.5 1.5 1.37 pyraclostrobin Flutriafol + fluoxastrobin 3,947.6 0.5 1.29 P value 0.259 0.375 0.659 x Means followed by the same letter for site years and by the same letters for treatments were not statistically different at α = 0.1. Means were separated using “Lines” option in the LSMEANS statement. y Treatments = Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF, Triangle Park, NC), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions, Philadelphia, PA). z Yield was adjusted to 13% seed moisture.
Table 5. Least square meansx of the plant count (PC), dry matter (DM), canopy cover (CC), total leaf area index (TLAI), green leaf area index with disease severity removed, normalized difference vegetation index (NDVI), and soil plant analysis development (SPAD) readings recorded in field experiments conducted in in Kanawha and Lewis county Iowa in 2018 and 2019.
y z GDDG Treatment PC DM (g) CC (%) TLAI GLAIDx NDVI SPAD GDDG-1 13.2 a 98.6 e 84.3 c 2.5 c 2.0 c 0.87 a 40.0 a GDDG-2 11.9 b 154.0 d 92.9 ab 3.9 ab 3.7 ab 0.89 a 40.8 a GDDG-3 11.0 b 201.8 c 96.0 a 4.0 a 3.8 a 0.88 a 37.7 a GDDG-4 11.8 b 317.8 b 96.6 a 3.3 b 3.1 b 0.82 b 34.6 a GDDG-5 11.2 b 335.3 ab 90.7 b 2.0 c 1.2 d 0.63 c 23.7 b GDDG-6 11.9 ab 365.0 a 58.3 d 0.9 d 0.5 d 0.50 d 17.3 b
NTC 11.6 233.1 83.0 b 3.0 2.6 0.78 32.7 89
Fluxapyroxad + 11.7 227.1 87.0 a 2.9 2.5 0.78 33.5 pyraclostrobin Flutriafol + 12.0 233.6 89.4 a 2.8 2.5 0.77 32.6 fluoxastrobin x Means followed by the same letter for site years and by the same letters for treatments were not statistically different at α = 0.1. Means were separated using “Lines” option in the LSMEANS statement. y GDDG = Growing degree day group. z Treatments = Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF, Triangle Park, NC), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions, Philadelphia, PA).
Table 6. Least square meansy of the plant count (PC), dry matter (DM), canopy cover (CC), total leaf area index (TLAI), green leaf area index with disease severity removed, normalized difference vegetation index (NDVI), and soil plant analysis development (SPAD) readings for different growing degree date groups (GDDG) and treatments recorded in a field experiment conducted in Lewis 2018.
a z GDDG Treatment PC DM (g) CC (%) TLAI GLAIDx NDVI SPAD GDDG-3 11.3 254.9 b 90.7 a 4.4 a 4.3 a 0.88 a 45.9 a GDDG-4 11.9 403.8 a 99.1 a 4.3 a 4.2 a 0.89 a 48.0 a GDDG-5 11.4 396.2 a 97.8 a 2.2 b 1.9 b 0.76 b 45.0 a GDDG-6 11.9 371.1 a 58.3 b 0.9 c 0.5 c 0.50 c 17.3 b P value 0.906 0.016 <0.001 <0.001 <0.001 <0.001 <0.001 NTC 11.3 354.6 81.5 3.1 2.9 0.76 37.6 Fluxapyroxad + 11.5 346.1 87.3 3.0 2.5 0.76 40.6 pyraclostrobin Flutriafol + fluoxastrobin 12.1 368.8 90.5 2.8 2.7 0.75 39.0 P value 0.653 0.781 0.120 0.414 0.356 0.597 0.529 y Means followed by the same letter for site years and by the same letters for treatments were not statistically different at α = 0.1. 90
Means were separated using “Lines” option in the LSMEANS statement. z Treatments = Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF, Triangle Park, NC), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions, Philadelphia, PA).
Table 7. Least square meansx of the canopy cover (CC) measured from experiments in Iowa during 2018 and 2019. CCy Site year Treatmentz GDDG-1 GDDG-2 GDDG-3 GDDG-4 GDDG-5 GDDG-6 P value Kanawha 81.8 b, C 90.2 b, B 98.7 a, A 93.2 b, AB 0.012 2018 Lewis 2018 90.7 c, A 99.1 a, A 97.8 a, A 58.3 B < 0.001 Kanawha 87.0 a, B 95.6 a, A 97.4 b, A 98.4 ab, A 85.3 b, B < 0.001 2019 P value 0.046 0.003 < 0.001 0.136 < 0.001 NTC 86.5 a, A 92.5 a, A 95.4 a, A 94.2 a, A 88.6 a, A 40.7 a, B < 0.001 Fluxapyroxad + 82.4 a, C 93.6 a, AB 96.3 a, A 97.6 a, A 91.6 a, B 60.6 a, D < 0.001 pyraclostrobin Flutriafol + 84.2 a, C 92.5 a, B 96.3 a, AB 98.1 a, A 91.8 a, B 73.4 a, D < 0.001 fluoxastrobin P value 0.481 0.863 0.814 0.512 0.776 0.246 x Means within a column followed by the same lower-case letter and means within a row followed by the same upper-case letter do 91 not differ significantly at α = 0.1. Means were separated using “Lines” option in LSMEANS statement y GDDG = Growing degree day group. The range between GDDs within GDDG-1 was 1280 to 1298, 1542 to 1560 for GDDG-2, 1808 to 1858 for GDDG-3, 2140 to 2243 for GDDG-4, 2344 to 2711 for GDDG-5, and 2832 for GDDG-6. z Treatments = Non-treated control (NTC) did not receive a fungicide application, fluxapyroxad + pyraclostrobin (Priaxor, BASF, Triangle Park, NC), flutriafol + fluoxastrobin (Preemptor, FMC Agricultural Solutions, Philadelphia, PA). 92
Table 8. Least square meansy of the total leaf area index (TLAI) and the green leaf area index with frogeye leaf spot severity removed (GLAIDx) measured from experiments in Iowa during 2018 and 2019z. TLAI Site year GDDG- GDDG- GDDG- GDDG- GDDG- GDDG-6 P value 1 2 3 4 5 Kanawha 2018 2.4 a, A 2.5 b, A 2.4 c, A 1.7 b, B 0.083 Lewis 2018 4.4 b, A 4.3 a, A 2.2 a, B 0.9 C < 0.001 Kanawha 2019 2.5 a, C 5.2 a, A 5.2 a, A 4.2 a, B 1.8 a, D < 0.001 P value 0.746 < 0.001 <0.001 < 0.001 0.287
GLAIDx Kanawha 2018 1.6 b, B 2.2 b, A 2.2 c, A 1.4 b, B 0.068 Lewis 2018 4.3 b, A 4.2 a, A 1.9 a, B 0.5 C < 0.001 Kanawha 2019 2.5 a, C 5.2 a, A 5.0 a, A 4.0 a, B 0.8 b, D < 0.001 P value 0.003 < 0.001 < 0.001 < 0.001 0.004 y Means within a column followed by the same lower-case letter and means within a row followed by the same upper-case letter do not differ significantly at α = 0.1. Means were separated using “Lines” option in LSMEANS statement. z GDDG = Growing degree day group. The range between GDDs within GDDG-1 was 1280 to 1298, 1542 to 1560 for GDDG-2, 1808 to 1858 for GDDG-3, 2140 to 2243 for GDDG-4, 2344 to 2711 for GDDG-5, and 2832 for GDDG-6.
93
Table 9. Pearson’s rank correlations between yield, frogeye leaf spot, and radiation use efficiency (RUE) combined across years and locations. Variable Yield Frogeye RUE leaf spot Yield - 0.15 -0.17 (0.420) (0.357) FLS 0.15 - 0.16 (0.420) (0.404) RUE -0.17 0.16 - (0.357) (0.404)
94
Table 10. Precipitation and average temperature during the summer period of June-August and September and the 30-year average for site years in Iowa. Precipitation (cm) 30-year average Average 30-year average precipitation temperature (°C) temperature Site year Jun-Aug Sep Jun-Aug Sep Jun-Aug Sep Jun-Aug Sep Lewis 53.8 16.8 36.7 8.4 22.9 19.5 22.5 17.9 2018 Kanawha 49.7 23.4 35.7 7.8 22.0 18.1 21.4 16.9 2018 Kanawha 29.4 14.3 36.6 8.2 21.2 19.0 21.3 16.9 2019
95
Figure 1. (A) Symptoms of frogeye leaf spot on soybean leaves. Image: X. Phillips. Taken 09- 06-2018 in Lewis, IA. (B) Symptoms of frogeye leaf spot on a soybean pod. Image: X. Phillips. Taken 09-13-2018 in Lewis, IA 96
Literature Cited
Adão, T., Hruška, J., Pádua, L., Bessa, J., Peres, E., Morais, R., and Sousa, J.J. 2017. Hyperspectral imaging: A review on UAV-based sensors, data processing and applications for agriculture and forestry. Remote Sens. 9:1110.
Akem, C.N., Dashiell, K.E., and Uwala, A.C. 1992. Prevalence of frogeye leaf spot of soybean in Nigeria. Int. J. Trop. Plant Dis. 10:181–183.
Altas, Z., Ozguven, M.M., and Yanar, Y. 2018. Determination of sugar beet leaf spot disease level (Cercospora Beticola Sacc.) with image processing technique by using drone. Curr. Investig. Agric. Curr. Res. 5:621–631.
Amaro, A.C.E., Baron D., Ono E.O., and Rodrigues J.D. 2019. Physiological effects of strobilurin and carboxamides on plants: an overview. Acta Physiol. Plant. 42:4 10.1007/s11738-019-2991-x
Athow, K.L., and Probst, A.H. 1952. The inheritance of resistance to frogeye leaf spot of soybeans. Phytopathology. 42:660–662.
Balba, H. 2007. Review of strobilurin fungicide chemicals. J. Environ. Sci. 42:441-451.
Ball, R.A., Purcell, L.C., and Vories, E.D. 2000. Optimizing soybean plant population for a short-season production system in the Southern USA. Crop Science. 40:757–764.
Bartlett, D.W., Clough, J.M., Godwin, J.R., Hall, A.A., Hamer, M., and Parr-Dobrzanski, B. 2002. The strobilurin fungicides. Pest Manag. Sci. 58:649-662.
Beloev, I.H. 2016. A review on current and emerging application possibilities for unmanned aerial vehicles. Acta Technol. Agric. 19:70–76.
Bendig, J., Yu, K., Aasen, H., Bolten, A., Bennertz, S., Broscheit, J., Gnyp, M.L., and Bareth, G. 2015. Combining UAV-based plant height from crop surface models, visible, and near infrared vegetation indices for biomass monitoring in barley. Int. J. Appl. Earth Obs. Geoinf. 39:79–87.
Bernaux, P. 1979. Identification of some soybean diseases in Cameroon. Agronomy Trop. 34:301-304.
Board, J. 2000. Light interception efficiency and light quality affect yield compensation of soybean at low plant populations. Crop Science. 40: 1285–1294.
Bolton, D.K., and Friedl, M.A. 2013. Forecasting crop yield using remotely sensed vegetation indices and crop phenology metrics. Agricultural and Forest Meteorology. 173:74–84. 97
Bradley, K.W., and Sweets, L.E. 2008. Influence of glyphosate and fungicide coapplications on weed control, spray penetration, soybean response, and yield in glyphosate-resistant soybean. Agronomy. J. 100:1360-1365.
Bryson, R.J., Paveley, N.D., Clark, W.S., Sylvester-Bradley, R., and Scott, R.K. 1997. Use of in-field measurements of green leaf area and incident radiation to estimate the effects of yellow rust epidemics on the yield of winter wheat. Eur. J. Agronomy. 7:53–62.
Bryson, R.J., Leandro, L., and Jones, D.R. 2000. The physiological effects of kresoxim- methyl on wheat leaf greenness and the implications for crop yield. 739–749. In The Proc. of the BCPC Conf., Brighton, UK. 13–16 Nov. 2000. British Crop Production Council, Hampshire, UK.
Calderón, R., Montes-Borrego, M., Landa, B.B., Navas-Cortés, J.A., and Zarco-Tejada, P.J. 2014. Detection of downy mildew of opium poppy using high-resolution multi-spectral and thermal imagery acquired with an unmanned aerial vehicle. Precis. Agric. 15:639–661.
Calderón, R., Navas-Cortés, J.A., and Zarco-Tejada, P.J. 2015. Early detection and quantification of verticillium wilt in olive using hyperspectral and thermal imagery over large areas. Remote Sens. 7:5584–5610.
Camera, J.N., Ghissi, V.C., Reis, E.M., and Deuner, C.C. 2016. The combined effects of temperature and leaf wetness periods on soybean frogeye leaf spot intensity. Semin. Cienc. Agrar. 37:77–84.
Carretero, R., Serrago, R.A., Bancal, M.O., Perelló, A.E., and Miralles, D.J. 2010. Absorbed radiation and radiation use efficiency as affected by foliar diseases in relation to their vertical position into the canopy in wheat. Field Crops Res. 116:184–195.
Ceotto, E., and Castelli, F. 2002. Radiation use efficiency in flue-cured tobacco (Nicotiana tabacum L.): response to nitrogen supply, climate variability and sink limitations. Field Crops Res. 74:117–130.
Cruz, C.D., and Dorrance, A.E. 2009. Characterization and survival of Cercospora sojina in Ohio. Online. Plant Health Progress. doi:10.1094/PHP-2009-0512-03-RS.
Dashiell, K.E., and Akem, C.N. 1991. Yield losses in soybeans from frogeye leaf spot caused by Cercospora sojina. Crop Prot. 10:465–468.
Devadas, R., Lamb, D.W., Simpfendorfer, S., Backhouse, D. 2009. Evaluating ten spectral vegetation indices for identifying rust infection in individual wheat leaves. Precision Agric. 10:459–470.
Dorrance, A.E., Cruz, C., Mills, D., Bender, R., Koenig, M., LaBarge, G., Leeds, R., Mangione, D., McCluer, G., Ruhl, S., Siegrist, H., Sundermeier, A., Sonnenberg, D., Yost, J., Watters, H., and Wilson, G. 2010. Effects of foliar fungicide and insecticide applications on soybean in Ohio. Plant Health Progress doi:10.1094/PHP-2010-0122-01-RS 98
Fehr, W.R., and Caviness, C.E. 1977. Stages of soybean development. Iowa State University Cooperative Extension Service, Special Report 80.
Filho, A.B., Carneiro, S.M.T.P.G., Godoy, C.V., Amorim, L., Berger, R.D., and Hau, B. 1997. Angular leaf spot of Phaseolus beans: relationships between disease, healthy leaf area, and yield. Phytopathology 87:506–15.
Flénet, F., Kiniry, J., Board, J., Westgate, M., and Reicosky, D.C. 1996. Row spacing effects on light extinction coefficients of corn, sorghum, soybean and sunflower. Agronomy. J. 88:185-190.
Fox, R.H., Piekielek, W.P., and Macneal, K.M. 1994. Using a chlorophyll meter to predict nitrogen fertilizer needs of winter wheat. Commun. Soil Sci. Plant Anal. 25:171-181.
Genovese, G., Vignolles, C., Nègre, T., and Passera, G. 2001. A methodology for a combined use of normalised difference vegetation index and CORINE land cover data for crop yield monitoring and forecasting. A case study on Spain. Agronomie 21:91–111.
Gitelson, A.A., and Gamon, J.A. 2015. The need for a common basis for defining light-use efficiency: Implications for productivity estimation. Remote Sens. Environ. 156:196–201.
Grossmann, K., Kwaltowski, J., and Caspar, G. 1999. Regulation of phytohormone levels, leaf senescence and transpiration by the strobilurin kresoxim-methyl in wheat (Triticum aestivum). J. Plant Physiol. 154:805–808.
Joshi, J., Sharma, S., and Guruprasad, K.N. 2014. Foliar application of pyraclostrobin fungicide enhances the growth, rhizobial-nodule formation and nitrogenase activity in soybean (var. JS-335). Pestic. Biochem. Physiol. 114:61–66.
Kaakeh, W., Pfeiffer, D.G., and Marini, R.P. 1992. Combined effects of spirea aphid (Homoptera: Aphididae) and nitrogen fertilization on net photosynthesis, total chlorophyll content, and greenness of apple leaves. J. Econ. Entomo. 185:939-946.
Kandel, Y.R., Mueller, D.S., Bestor, N.R.C., Bradley, C.A., Giesler, L.J., Wise, K.A. 2016. Analyses of yield and net economic response from foliar fungicide and insecticide applications to soybean in the North Central United States. Plant health Prog. 17:232-238 doi:10.1094 / PHP-RS-16-0038.
Kerkech, M., Hafiane, A., and Canals, R. 2018. Deep leaning approach with colorimetric spaces and vegetation indices for vine diseases detection in uav images. Comput. Electron. Agric. 155:237–243.
Kumudini, S., Godoy, C.V., Board, J.E., Omielan, J., and Tollenaar, M. 2008. Mechanisms involved in soybean rust-induced yield reduction. Crop Sci. 48:2334–2342.
Kyveryga, P.M., Blackmer, T.M., and Mueller, D.S. 2013. When do foliar pyraclostrobin fungicide applications produce profitable soybean yield responses? Online. Plant Health Prog. doi:10.1094/PHP_2013-0928-01-RS. 99
Leite, R.M.B.V.C., Amorim, L., and Filho, A.B. 2006. Relationships of disease and leaf area variables with yield in the Alternaria helianthi-sunflower pathosystem. Plant Pathol. 55:73– 81.
Lindquist, J.L., Arkebauer, T.J., Walters, D.T., Cassman, K.G., and Dobermann, A. 2005. Maize radiation use efficiency under optimal growth conditions. Agronomy. J. 97:72–78.
Ma, G.Z. 1994. Review and forecast of study on frogeye leaf spot. Soybean J. 1:6–7.
Madden, L.V., and Nutter, F.W. 1995. Modeling crop losses at the field scale. Can. J. Plant Pathol. 17:124–35.
Maitree, L., and Toyota, M. 2017. A high seed yield and associated attributes of dry matter production achieved by recent Japanese soybean cultivars. Plant Production Science. doi: 10.1080/1343943X.2017.1294463
Mathew, F.M., Byamukama, E., Neves, D.L., and Bradley, C.A. 2019. Resistance to quinone outside inhibitor fungicides conferred by the G143A mutation in Cercospora sojina (causal agent of frogeye leaf spot) isolates from South Dakota soybean fields. Plant Health Prog. 20:104–105.
Mengistu, A., Kurtzweil, N.C., and Grau, C.R. 2002. First report of frogeye leaf spot (Cercospora sojina) in Wisconsin. Plant Dis. 11:1272.
Mengistu, A., Kelly, H.M., Bellaloui, N., Arelli, P.R., Reddy, K.N., and Wrather, A.J. 2014. Tillage, fungicide, and cultivar effects on frogeye leaf spot severity and yield in soybean. Plant Dis. 98:1476-1484.
Mengistu, A., Kelly, H.M., Arelli, P.R., Bellaloui, N., and Lin, B. 2018. Quantifying the effects of fungicides and tillage on Cercospora sojina severity and yield of soybean. Plant Health Prog. 19:226-232.
Meyer, G.E., Neto, J.C. 2008. Verification of color vegetation indices for automated crop imaging applications. Comput. Electron. Agric. 63:282–293.
Mian, M.A.R., Phillips, D.V., Boerma, H.R., Missaoui, A.M., and Walker, D.R. 2008. Frogeye leaf spot of soybean: A review and proposed race designations for isolates of Cercospora sojina Hara. Crop Sci. 48:14–24.
Monteith, J.L. 1977. Climate and the efficiency of crop production in Britian. Philos. Trans. R. Soc. London B. 281:277-294.
Morrison, M.J., Voldeng, H.D., and Cober, E.R. 1999. Physiological changes from 58 years of genetic improvement of short-season soybean cultivars in Canada. Agronomy. J. 91:685- 689. 100
Muchow, R.C., Robertson, M.J., and Pengelly, B.C. 1993. Radiation-use efficiency of soybean, mugbean and cowpea under different environmental conditions. Field Crops Research 32:1-16.
Mueller, D.S., Kandel, Y.R., and Wiggs, S.N. 2019. Evaluation of Foliar Fungicides on Soybeans in 2018. Iowa State University Integrated Crop Management Newsletter. https://crops.extension.iastate.edu/cropnews/2019/02/evaluation-foliar-fungicides-soybeans- 2018.
Park, S., Nolan, A., Ryu, D., Fuentes, S., Hernandez, E., Chung, H., and O’Connell, M. 2015. Estimation of crop water stress in a nectarine orchard using high-resolution imagery from unmanned aerial vehicle (UAV). In Proceedings of the 21st International Congress on Modelling and Simulation, Gold Coast, Australia. 1413–1419.
Phillips, X.A., Singh, A.K., Kandel, Y.R., and Mueller, D.S. 2017. Effect of pod removal, foliar fungicides, and cultivar on green stem disorder of soybean. Agronomy J. 109:2680- 2688.
Phillips, X.A., Kandel, Y.R., Licht, M.A., and Mueller, D.S. 2020. Estimating soybean radiation use efficiency using a UAV in Iowa. Agronomy. doi.org/10.3390/agronomy10122002
Piekielek, W.P., and Fox, R.H. 1992. Use of a chlorophyll meter to predict sidedress nitrogen requirements for maize. Agronomy. J. 84:59-65.
Primicerio, J., Di Gennaro, S.F., Fiorillo, E., Genesio, L., Lugato, E., Matese, A., and Vaccari, F.P. 2012. A flexible unmanned aerial vehicle for precision agriculture. Precis. Agric. 13:517–523.
Probst, A.H. Border effects in soybean nursery plots. 1943. J. Am. Soc. Agronomy 35:662- 666.
Purcell, L.C., Ball, R.A., Reaper, J.D., and Vories, E.D. 2002. Radiation use efficiency and biomass production in soybean at different plant population densities. Crop Science 42:172– 177.
Roth, M.G., Webster, R.W., Mueller, D.S., Chilvers, M.I., Faske, T.R., Mathew, F.M., Bradley, C.A., Damicone, J.P., Kabbage, M., and Smith, D.L. 2020. Integrated management of important soybean pathogens of the united states in changing climate. Journal of Integrated Pest Management. 11:17.
Santos, J.B., Procopio, S.O., Silva, A.A., and Costa, L.C. 2003. Capture and utilization of solar radiation by the soybean and common bean crops and by weeds. Bragantia. 62:147-153.
Serrago, R.A., Carretero, R., Bancal, M.O., and Miralles, D.J. 2009. Foliar diseases affect the ecophysiological attributes linked with yield and biomass in wheat (Triticuma estivum L.). Eur. J. Agronomy. 31:195–203. 101
Sherwin, H.S., and Kreitlow, K.W. 1952. Discoloration of soybean seeds by the frogeye fungus, Cercospora sojina. Phytopathology 42:568–572.
Smith, G.M., and Milton, E.J. 1999. The use of the empirical line method to calibrate remotely sensed data to reflectance. Int. J. Remote Sens. 20:2653–2662.
Srinivasan, V., Kumar, P., and Long, S. P. 2016. Decreasing, not increasing, leaf area will raise crop yields under global atmospheric change. Glob. Chang. Biol. 23:1626–1635.
Sugiura, R., Tsuda, S., Tsuji, H., and Murakami, N. 2018. Virus-infected plant detection in potato seed production field by uav imagery. In Proceedings of the 2018 ASABE Annual International Meeting, Detroit, MI, USA. 1800594.
Swoboda, C., and Pedersen, P. 2009. Effect of fungicide on soybean growth and yield. Agron. J. 101:352–356.
Taiz, L., Zeiger, E., Møller, I.M., and Murphy, A. 2014. Plant physiology and development. 2015. 6th ed.; Sinauer Assotiates: Sunderland, MA, USA. ISBN:1605352551.
Tetila, E.C., Machado, B.B., Belete, N.A.S., Guimarães, D.A., and Pistori, H. 2017. Identification of soybean foliar diseases using unmanned aerial vehicle images. IEEE Geosci. Remote Sens. Lett. 14:2190–2194.
Tewes, A., and Schellberg, J. 2018. Towards remote estimation of radiation use efficiency in maize using UAV-based low-cost camera imagery. Agronomy. 8:e16.
Tollenaar, M., and Aguilera, A. 1992. Radiation use efficiency of an old and a new maize hybrid. Agronomy J. 84:536–541.
Turner, F.T., and Jund, M.F. 1991. Chlorophyll meter to predict nitrogen topdress requirement for semidwarf rice. Agronomy J. 83:926-928.
Uddling, J., Gelang-Alfredsson, J., Piikki, K., and Pleijel, H. 2007. Evaluating the relationship between leaf chlorophyll concentration and SPAD-502 chlorophyll meter readings. Photosynth. Res. 91:37–46.
Waggoner, P.E., and R.D. Berger. 1987. Defoliation, disease, and growth. Phytopathology 77:393 398.
Wang, C., and Myint, S.W. 2015. A simplified empirical line method of radiometric calibration for small unmanned aircraft systems-based remote sensing. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 8:1876–1885.
Warren-Wilson, J.W. 1967. Ecological data on dry-matter production by plants and plant communities. In: E.F. Bradley and O.T. Denmead, editors, The collection and processing of field data. Interscience Publ. New York. p.77–123. 102
Westphal, A., Abney, T.S., and Shaner, G. 2006. Diseases of soybean. Frogeye leaf spot. Purdue University Department of Botany and Plant Pathology and USDA-ARS. http://www.ces.purdue.edu/extmedia/BP/BP-131-W.pdf
Wise, K., and Mueller, D. 2011. Are fungicides no longer just for fungi? An analysis of foliar fungicide use in corn. APSnet Features. doi:10.1094/APSnetFeature-2011-0531.
Woebbecke, D.M., Meyer, G.E., Von Bargen, K., and Mortensen, D.A. 1995. Color indices for weed identification under various soil residue and lighting conditions. Transactions of the ASAE. 38:259–269.
Wood, C.W., Reeves, D.W., Duffield, R.R., and Edmisten, K.L. 1992. Field chlorophyll measurements for evaluation of corn nitrogen status. J. Plant Nutri. 15:487-500.
Wrather, J.A., and Koenning, S.R. 2006. Estimates of disease effects on soybean yields in the United States 2003-2005. J. Nematol. 38:173-180.
Yang, X.B., Uphoff, M.D., and Sanogo, S. 2001. Outbreaks of soybean frogeye leaf spot in Iowa. Plant Dis. 85:443.
Yuan, Z., Cao, Q., Zhang, K., Ata-Ul-Karim, S.T., Tian, Y., Zhu, Y., Cao, W., and Liu, X. 2016. Optimal leaf positions for spad meter measurement in rice. Front. Plant Sci. 7:719.
Zhang, G.R., Newman, M.A., and Bradley, C.A. 2012. First report of the soybean frogeye leaf spot fungus (Cercospora sojina) resistant to quinone outside inhibitor fungicides in North America. Plant Dis. 96:767.
Zhang, D., Zhou, X., Zhang, J., Huang, L., and Zhao, J. 2017. Developing a small uav platform to detect sheath blight of rice. In Proceedings of the 2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Fort Worth, TX, USA. 3190–3193.
Zhang, G.R., Allen, T.W., Bond, J.P., Fakhoury, A.M., Dorrance, A.E., Weber, L., Faske, T.R., Giesler, L.J., Hershamn, D.E., Kennedy, B.S., et al. 2018. Widespread occurrence of quinone outside inhibitor fungicide-resistant isolates of Cercospora sojina, causal agent of frogeye leaf spot of soybean, in the United States. Plant Health Prog. 19:295–302.
103
Appendix. Tables and Figures
Table A1. Least square meansz of the plant count (PC) measured from experiments in Iowa during 2018 and 2019 across site years and growing degree day group (GDDG). PC Site year GDDG-1 GDDG-2 GDDG-3 GDDG-4 GDDG-5 GDDG-6 P value Kanawha 15.8 a, A 11.6 a, B 10.6 a, B 11.2 b, B 0.001 2018 Lewis 2018 11.3 a, A 11.9 ab, 11.4 a, A 11.9 A 0.906 A Kanawha 10.5 b, B 12.2 a, A 11.3 a, 12.4 a, A 11.0 a, 0.203 2019 AB AB P value < 0.001 0.502 0.615 0.195 0.566 z Means within a column followed by the same lower-case letter and means within a row followed by the same upper-case letter do not differ significantly at α = 0.1. Means were separated using “Lines” option in LSMEANS statement.
104
Table A2. Least square meansy of normalized difference vegetation index (NDVI)z values measured from field experiments in 2018 and 2019 in Iowa across growing degree day group (GDDG). NDVI Site year GDDG-1 GDDG-2 GDDG-3 GDDG-4 GDDG- GDDG- P value 5 6 Kanawha 0.87 a, A 0.88 b, A 0.88 a, A 0.72 b, B 0.010 2018 Lewis 2018 0.88 a, A 0.89 a, A 0.76 a, B 0.50 C < 0.001 Kanawha 0.88 a, AB 0.89 a, A 0.88 a, AB 0.87 a, B 0.54 b, C < 0.001 2018 P value 0.418 < 0.001 0.990 < 0.001 < 0.001 y Means within a column followed by the same lower-case letter and means within a row followed by the same upper-case letter do not differ significantly at α = 0.1. Means were separated using “Lines” option in LSMEANS statement. z Normalized difference vegetation index values collected with a handheld GreenSeeker (Trimble, Sunnyvale, CA, USA).
105
Table A3. Least Square meansy of soil plant analysis development (SPAD)z values measured from field experiments in 2018 and 2019 in Iowa across growing degree day group (GDDG). SPAD Site year GDDG-1 GDDG-2 GDDG-3 GDDG-4 GDDG-5 GDDG-6 P value Kanawha 40.0 B 40.8 B 45.3 a, A 38.0 b, B 0.045 2018 Lewis 45.9 a, A 48.0 a, A 45.0 a, A 17.3 B < 0.001 2018 Kanawha 24.0 b, A 21.1 c, B 6.2 b, C < 0.001 2019 P value < 0.001 < 0.001 < 0.001 y Means within a column followed by the same lower-case letter and means within a row followed by the same upper-case letter do not differ significantly at α = 0.1. Means were separated using “Lines” option in LSMEANS statement. z In 2018, a SPAD-502 meter (Konica-Minolta, Japan) was used and in 2019 a CCM-200 (Opti- Sciences, Hudson, NH, USA) was used for readings.
106
Figure A1. Workflow of the of the entire RUE2019 tool composed of the three groups; “COMBINE BANDS”, “RASTER CALCULATOR (EXG)”, and “EXTRACT TO EXCEL” in ArcGIS Pro ModelBuilder. Blue ovals = input variables; yellow rectangles = tools; green ovals = output; P = parameter.
107
Figure A2. The RUE2019 tool dialogue box in ArcGIS Pro. The input variables are: Band 1, Band 2, Band 3, and Plots RUE. The outputs are: Summarize B123FO and ExG DNs 31.
108
Figure A3. Workflow of the entire Targets tool composed of four groups; “COMBINE BANDS”, “WHITE TARGET”, and “GRAY TARGET” in ArcGIS Pro ModelBuilder. Blue ovals = input variables; yellow rectangles = tools; green ovals = output; P = parameter. 109
Figure A4. The dialogue box of the Targets tool in ArcGIS Pro. The input variables are: Band 1, Band 2, Band 3, Black Target, Gray Target, and White Target. The outputs are: Black Target xlsx, Gray Target xlsx, and White Target xlsx.
110
CHAPTER 4. GENERAL CONCLUSION
Overall, radiation use efficiency (RUE) of soybean, partially estimated by unmanned aerial vehicle (UAV) imagery fell within the range of previously published values. The lack of an increase in yield with prophylactic application of strobilurin fungicides is in agreement with previous investigations. RUE values did not reflect the impact on yield of this specific pathogen.
The application of foliar fungicides on soybean at low levels of frogeye leaf spot caused by the pathogen Cercospora sojina Hara, did not impact yield. Applications of fungicides and their promoted health benefits have been inconsistent across previous investigations. In our study, we found foliar fungicides increased canopy cover, but this did not manifest in additional yield.
Applications of certain foliar fungicides protected yield from frogeye leaf spot. Fungicide resistance was present within the C. sojina population, as has been show in previous reports to exist in Iowa. Normalized difference vegetation index (NDVI) values and soil plant analysis development (SPAD) readings did not reflect the significant differences between fungicides with regards to frogeye severity. This suggests that symptomatic (green) leaf area is not affected by C. sojina. Estimated RUE did not differ significantly across treatments when yield was different although trends indicated higher RUE with fungicide application compared to the non-treated control. A combination of experimental errors may have convoluted the parameters needed for
RUE estimation. Similarly, no significant correlations between yield, frogeye leaf spot severity, and RUE were detected.
This research demonstrated RUE estimation using UAV imagery can be performed on a small plot scale. Since it is directly related to yield, RUE could be an appropriate parameter to elucidate the impact of plant diseases and other stresses on yield. However, factors such as the 111 overabundance of leaves that soybean has and the impact that pathogens may have on non-lesion green leaf area are issues that require further investigation.