Hostname: page-component-78c5997874-t5tsf Total loading time: 0 Render date: 2024-11-06T22:12:30.624Z Has data issue: false hasContentIssue false

Screening glyphosate-alternative weed control options in important perennial crops

Published online by Cambridge University Press:  24 August 2021

Panagiotis Kanatas
Affiliation:
Teaching Staff (P.D. 407/80), Department of Crop Science, University of Patras, Mesolonghi, Greece
Nikolaos Antonopoulos
Affiliation:
Ph.D. Candidate, Agricultural University of Athens, Faculty of Crop Sciences, Athens, Greece
Ioannis Gazoulis
Affiliation:
Ph.D. Candidate, Agricultural University of Athens, Faculty of Crop Sciences, Athens, Greece
Ilias S. Travlos*
Affiliation:
Assistant Professor, Agricultural University of Athens, Faculty of Crop Sciences, Athens, Greece
*
Author for correspondence: Ilias S. Travlos, Agricultural University of Athens, Faculty of Crop Sciences, 75, Iera Odos str., GR11855, Athens, Greece. (Email: travlos@aua.gr)
Rights & Permissions [Opens in a new window]

Abstract

The current study aimed to screen glyphosate-alternative weed control methods in three perennial crops in Greece. Field trials were conducted and repeated (2018 to 2019 and 2019 to 2020) in a citrus orchard (Citrus clementina Hort. ex Tan), an olive grove (Olea europaea L.), and a vineyard (Vitis vinifera L.) under the randomized complete block design (nine treatments, four blocks). Glyphosate was applied in the citrus orchard (720 g ae ha−1), the olive grove (720 g ae ha−1), and the vineyard (1,800 g ae ha−1). Pelargonic acid (1,088 g ha−1; two times), barley (Hordeum vulgare L.) residues and white mustard (Sinapis alba L.) residues were evaluated in all sites. Mowing was evaluated in the citrus orchard (one time) and the vineyard (two times). Flazasulfuron (50 g ha−1), oxyfluorfen (144 g ha−1), and flumioxazin (150 g ha−1) were applied in the citrus orchard and the olive grove. Penoxsulam + florasulam (15 + 7.5 g ha−1) was also applied in the olive grove. Cycloxydim (200 g ha−1), quizalofop-p-ethyl (150 g ha−1) and propaquizafop (150 g ha−1) were applied in the vineyard. An untreated control was included in all sites. Flazasulfuron, oxyfluorfen, and flumioxazin resulted in similar normalized difference vegetation index (NDVI) and weed biomass to glyphosate in the citrus orchard in both years and evaluations. Pelargonic acid (two times) and mowing (one time) were effective on broadleaf weeds. Flazasulfuron and penoxsulam + florasulam were the most promising glyphosate-alternative weed control methods against hairy fleabane [Conyza bonariensis (L.) Cronquist] in the olive grove. Cover crop residues showed their suppressive ability as in the citrus orchard. All selective herbicides resulted in similar NDVI and johnsongrass [Sorghum halepense (L.) Pers.] dry weight values in the vineyard in both years. Negative and strong correlations were observed in all sites and years between crop yield and weed dry weight (R2 = 0.543 to 0.924).

Type
Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2021. Published by Cambridge University Press on behalf of the Weed Science Society of America

Introduction

In perennial cropping systems, weed presence contributes to soil conservation, improves soil fertility, and provides food and shelter to beneficial organisms (Mia et al. Reference Mia, Massetani, Murri, Facchi, Monaci, Amadio and Neri2020). However, weed management is an essential agronomic practice in fruit orchards and vineyards (Onen et al. Reference Onen, Akdeniz, Farooq, Hussain and Ozaslan2018; Travlos et al. Reference Travlos, Bilalis, Katsenios and De Prado2018). Weeds can act as pathogen and insect pest hosts (Ligoxigakis et al. Reference Ligoxigakis, Vakalounakis and Thanassoulopoulos2002; Thanou et al. Reference Thanou, Kontogiannis and Tsagkarakis2021). Left uncontrolled, weeds can be established in high densities between tree rows and trees in the row and compete with the growing trees for water and nutrients (Mia et al. Reference Mia, Massetani, Murri, Facchi, Monaci, Amadio and Neri2020). Weed competition during crucial growth stages can affect tree establishment, limit fruit production, and downgrade product quality (Mia et al. Reference Mia, Massetani, Murri, Facchi, Monaci, Amadio and Neri2020; Travlos et al. Reference Travlos, Bilalis, Katsenios and De Prado2018; Tursun et al. Reference Tursun, Işık, Demir and Jabran2018).

In the European Union (EU), and indeed worldwide, the use of glyphosate remains by far the most common and effective practice for controlling weeds in perennial orchards (Duke and Powles Reference Duke and Powles2008; Kudsk and Mathiassen Reference Kudsk and Mathiassen2020). Glyphosate is a nonselective, systemic herbicide acting as an inhibitor of 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) of the shikimic acid pathway. It is the world’s most important herbicide, with unique properties such as its excellent uptake and translocation and its toxicologically and environmentally safe profile (Duke and Powles Reference Duke and Powles2008). This herbicide had a prominent role on the adoption of minimum-tillage and no-till cropping systems that maximized crop productivity even in semiarid lands where tillage leads to soil erosion and degradation (Farooq et al. Reference Farooq, Flower, Jabran, Wahid and Siddique2011). In perennial orchards and fallow areas, it is the only herbicide that can be applied to perennial weeds at late growth stages and still be effective (Miller et al. Reference Miller, Griffin and Richard1998). For all these reasons and more, glyphosate has been highlighted as the once-in-a-century herbicide (Duke and Powles Reference Duke and Powles2008). However, the agricultural sector seems to be overreliant on this specific herbicide. Consecutive applications in perennial orchards and fallow areas pose a common phenomenon during the last decades and have resulted in the spread of glyphosate-resistant weed populations. This is noted especially for the fruit orchards and vineyards of the Mediterranean region (Tahmasebi et al. Reference Tahmasebi, Alebrahim, Roldán-Gómez, da Silveira, de Carvalho, Alcántara-de la Cruz and De Prado2018; Travlos and Chachalis Reference Travlos and Chachalis2010; Travlos et al Reference Travlos, Kanatas, Tsekoura, Gazoulis, Papastylianou, Kakabouki and Antonopoulos2020a; Vazquez-Garcia et al. Reference Vazquez-Garcia, Palma-Bautista, Rojano-Delgado, De Prado and Menendez2020). Consequently, glyphosate applications should be integrated with nonchemical weed control practices, while alternative herbicides with different modes of action should be also evaluated (Scavo and Mauromicale Reference Scavo and Mauromicale2020; Travlos et al. Reference Travlos, Montull, Kukorelli, Malidza, Dogan, Cheimona, Antonopoulos, Kanatas, Zannopoulos and Peteinatos2019). Moreover, although glyphosate has lower environmental impact compared with other herbicides and there is not a causal association between glyphosate and genotoxicity to humans (Duke Reference Duke2020), the strict EU legislation for pesticides creates concerns regarding glyphosate’s future in Europe (Beckie et al. Reference Beckie, Flower and Ashworth2020; Fogliatto et al. Reference Fogliatto, Ferrero, Vidotto and Sparks2020). To maintain efficacy over time and partially counterbalance the economic consequences of the worst-case scenario (glyphosate phase-out), alternative weed management strategies should be investigated in the perennial orchards of Europe (Fogliatto et al. Reference Fogliatto, Ferrero, Vidotto and Sparks2020; Kudsk and Mathiassen Reference Kudsk and Mathiassen2020).

Glyphosate-alternative weed control strategies rely first upon pre- or early postemergence herbicides with soil residual activity that enable long-term weed control with a single application and the applications of selective herbicides on targeted weed species that have dominated a particular field (Scarabel et al. Reference Scarabel, Panozzo, Savoia and Sattin2014; Tahmasebi et al. Reference Tahmasebi, Alebrahim, Roldán-Gómez, da Silveira, de Carvalho, Alcántara-de la Cruz and De Prado2018; Travlos et al. Reference Travlos, Lysandrou and Apostolidis2014). Natural contact, burndown herbicides can also contribute to solutions for the management of annual weed species (Muñoz et al. Reference Muñoz, Torres-Pagán, Peiró, Guijarro, Sánchez-Moreiras and Verdeguer2020; Young Reference Young2004). Mowing operations are also among the recommended agronomic practices that reduce weed growth and the reproductive ability of a various weed species (Butler et al. Reference Butler, Brouder, Johnson and Gibson2013; Mia et al. Reference Mia, Massetani, Murri, Facchi, Monaci, Amadio and Neri2020). Winter-grown cover crops can outcompete weeds during the winter period. Moreover, the presence of their residues on the soil surface after cover crop termination can suppress the growth of summer weeds in the orchard (Alcántara et al. Reference Alcántara, Pujadas and Saavedra2011; Tursun et al. Reference Tursun, Işık, Demir and Jabran2018). All these practices have the potential to be combined with glyphosate applications in a rotational view under the concept of integrated weed management (Gerhards and Schappert Reference Gerhards and Schappert2020; Raimondi et al. Reference Raimondi, Constantin, Mendes, Oliveira and Rios2020). It should be noted that in several studies in which glyphosate-resistant weed populations were reported, alternative herbicides were tested for weed control (Tahmasebi et al. Reference Tahmasebi, Alebrahim, Roldán-Gómez, da Silveira, de Carvalho, Alcántara-de la Cruz and De Prado2018; Travlos and Chachalis Reference Travlos and Chachalis2010; Travlos et al Reference Travlos, Kanatas, Tsekoura, Gazoulis, Papastylianou, Kakabouki and Antonopoulos2020a; Vazquez-Garcia et al. Reference Vazquez-Garcia, Palma-Bautista, Rojano-Delgado, De Prado and Menendez2020). However, research is also needed to evaluate alternative weed control methods in fields where glyphosate is still effective to get an estimation of the strengths and the weaknesses of any alternative strategy compared with glyphosate.

The objective of the current study was to conduct an initial screening of glyphosate-alternative weed control options in three important perennial crops in southern Greece. Field trials were conducted to evaluate the efficacy of the alternative weed control strategies in comparison with glyphosate. Alternative herbicides currently registered and available to farmers in the EU were tested. Mechanical weed control and treatments related to cover crop residue management were also evaluated.

Materials and Methods

Site Description

Field trials were conducted in a citrus orchard (Citrus clementina Hort. ex Tan, ‘Marisol’), an olive grove (Olea europaea L., ‘Koroneiki’), and a vineyard (Vitis vinifera L., ‘Merlot’) in southern Greece. Each separate trial was conducted during 2018 to 2019 and repeated during 2019 to 2020. New plots were established in new areas of the sites in 2019 to 2020 to repeat each experiment in time. The citrus orchard was located outside Pyrgos, Elis (37.644°N, 21.473°E). Soil type was loam. Soil texture (0 to 30 cm) was: 24.7% clay, 39.8% silt, and 35.5% sand, with a pH of 7.3 and 1.6% organic matter content. The trees were 7-yr-old, grafted onto ‘Swingle citrumelo’ rootstock (Citrus paradisi Macf. × Poncirus trifoliata L. Raf.), spaced 4.5 by 4.5 m apart, and under a drip-irrigation system. The dominant weed species were barnyardgrass [Echinochloa crus-galli (L.) P. Beauv.], common lambsquarters (Chenopodium album L.), black nightshade (Solanum nigrum L.), and redroot pigweed (Amaranthus retroflexus L.) and were highly persistent for both years.

The olive grove was located in Kyparissia, Messinia (37.227°N, 21.632°E). Soil type was sandy loam. Soil texture (0 to 30 cm) was: 18.1% clay, 28.3% silt, and 52.9% sand, with a pH of 7.6 and 1.4% organic matter content. The trees were 13-yr-old, self-rooted, spaced 6 by 6 m apart, and grown under rainfed conditions. For both years, hairy fleabane [Conyza bonariensis (L.) Cronquist] dominated this site. The vineyard was located in Korakochori, Elis (37.676°N, 21.310°E). Soil type was sandy clay loam. Soil texture (0 to 30 cm) was: 28.4% clay, 25.1% silt, and 46.5% sand with a pH of 7.5 and 1.3% organic matter content. The vines were 5-yr-old, grafted onto ‘1103 Paulsen’ rootstock, and under a drip-irrigation system. Row spacing was 2.20 m. Vine spacing was 1.25 m. Johnsongrass [Sorghum halepense (L.) Pers.] was the dominant species in this site for both years. Similar climatic conditions prevailed in the experimental fields during both years (Table 1). This was expected for Elis fields and also for the olive grove, as it is located in the northwestern part of Messinia, neighboring the Elis district.

Table 1. Mean monthly temperature (C) and overall monthly precipitation (mm) at each site during 2018–2019 and 2019–2020.

a Mean T, mean temperature.

Experimental Setup

Herbicide application and the management of two cover crops’ residues were common weed control treatments in all trials. All herbicides were applied with a pressurized Gloria® 405 T sprayer (Gloria Haus & Gartengeraete GMBH, Witten, Germany) calibrated to deliver 300 L ha−1 of spray solution at constant pressure through five 8002 flat-fan spray nozzles (spray angle: 80o). Direct spraying was carried out between rows and plants in the row. Details regarding herbicide products used in each field are presented in Table 2. Cover crop establishment and termination procedures were also the same in all sites. Barley (Hordeum vulgare L., ‘Finola’) (K+N Efthymiadis S.A., Thessaloniki, Greece) and white mustard (Sinapis alba L., ‘Iris’) (Vandinter Semo B.V., Scheemda, Groningen, The Netherlands) were selected, due to evidence suggesting the suppressive ability of their residues on summer weeds (Dhima et al. Reference Dhima, Vasilakoglou, Eleftherohorinos and Lithourgidis2006; Didon et al. Reference Didon, Kolseth, Widmark and Persson2014). Seedbed preparation included tillage to 20-cm depth with a harrow pass and fertilizer incorporation with 50 kg ha−1 N and 60 kg ha−1 P2O5. Seeds were drilled at 0.5-cm depth. Seeding rates were 140 and 30 kg ha−1 for barley and white mustard, respectively. Rows were spaced 20 cm apart. Cover crop establishment was carried out in early October for both years. Termination was done by mowing in mid-April (Table 3). Cover crop residues were left on the soil surface to suppress summer weeds. Mowing was also a common weed control treatment in the citrus orchard and the vineyard. An intermediate-type mower with hydraulic motion control, 35 HP power, and 150-cm mowing width (OMA SGK/S 150, Panagrotiki S.A., Lamia, Greece) was towed on the field. Untreated plots were also included in all trials.

Table 2. Information regarding herbicides applied in each site.

a Glyphosate was applied at 720 g ae ha−1 in the citrus orchard and the olive grove. The herbicide was applied at the rate of 1,800 g ae ha−1 in the vineyard due to the presence of Sorghum halepense.

Table 3. Dates of key field activities carried out at each site for each experimental year (2018–2019 and 2019–2020).

a CC, cover crop; NDVI, normalized difference vegetation index.

A randomized complete block design (RCBD) with nine treatments and four blocks (replications) was established in the citrus orchard. Treatments included the applications of glyphosate, pelargonic acid (two times), flazasulfuron, oxyfluorfen, flumioxazin, mowing (one time), barley residue management, white mustard residue management, and an untreated control. Glyphosate was applied once at 720 g ae ha−1. Weeds were treated twice with pelargonic acid in a 15-d interval. Both pelargonic acid applications were carried out at an application rate of 1,088 g ha−1. The application timings of the single mowing treatment, the second pelargonic acid application, and the glyphosate application were the same (Table 3). These treatments were applied when the weeds were between the 4- and 6-leaf growth stages. Flazasulfuron, oxyfluorfen, and flumioxazin were applied as preemergence herbicides at 50, 144, and 150 g ha−1, respectively. Rainfall events occurred 2 d after treatment in April 2019. Precipitation of 26.8 mm was sufficient to ensure herbicide incorporation. For this purpose, plots were sprinkler-irrigated in April 2020. Water volume of 25 mm was applied to wet the top 15 cm of soil to activate herbicides. Glyphosate and flazasulfuron were applied at 200 kPa. Pelargonic acid, oxyfluorfen, and flumioxazin were applied at 250 kPa. Plot size was 7.5-m long by 7.5-m wide to include four trees per plot. A border of 1.5 m was kept between adjacent plots. In the second experimental year (2019 to 2020), new plots were established in this site to assure that the experiment was repeated in time.

In the olive grove, An RCBD was implemented with nine treatments assigned in four blocks. Treatments were: glyphosate, pelargonic acid (2x), flazasulfuron, oxyfluorfen, flumioxazin, penoxsulam + florasulam, barley residue management, white mustard residue management, and an untreated control. Glyphosate was applied as a single treatment at 720 g ae ha−1. Pelargonic acid was applied twice at 1,088 g ai ha−1 at 2-wk intervals. The application timings of the second pelargonic acid application and the single glyphosate application were the same (Table 3). These treatments were applied when the weeds had either small-sized (7- to 9-cm diameter) or large-sized rosettes (12- to 16-cm diameter). Preemergence herbicides included flazasulfuron (50 g ha−1), oxyfluorfen (144 g ha−1), flumioxazin (150 g ha−1), and the application of a prepackage mixture of penoxsulam + florasulam at 15 + 7.5 g ha−1. Herbicides were activated with 28.2 mm of water deposited on the soil surface during a single rain event in April 2019, 3 d after treatment. For the same reason, 13 mm of water was applied with sprinklers in April 2020 to incorporate herbicides. Application pressures for glyphosate, pelargonic acid, flazasulfuron, oxyfluorfen, and flumioxazin were the same as in the citrus orchard. Penoxsulam + florasulam mixture was applied at 250 kPa. Plots were 8-m long by 8-m wide. A 2-m border was kept between plots. In 2019 to 2020, the experiment was repeated in new plots in this site to repeat the experiment in time.

An RCBD was established in the vineyard with nine treatments repeated four times. Treatments were: glyphosate, pelargonic acid (2x), cycloxydim, quizalofop-p-ethyl, propaquizafop, mowing (two times), barley’s residue management, white mustard residue management, and an untreated control. Glyphosate was applied once at 1,800 g ae ha−1. Pelargonic acid was applied twice (1,088 g ha−1) at 2-wk intervals. Cycloxydim, quizalofop-p-ethyl, and propaquizafop were applied once at 200, 150, and 150 g ha−1, respectively. Two mowing operations were carried out in a 2-wk interval. The application timings of the second mowing treatment, the second pelargonic acid application, and the single applications of the other herbicides were the same (Table 3). These treatments were applied when the weeds were between the 4- and 6- leaf growth stages (30- to 40-cm height). Herbicides were applied at 200 kPa. Plot size was 4.5-m long by 3.5-m wide to include eight vines per plot. A border of 0.9 m was kept among neighboring rows. Borders among plots were 0.4 m. New plots were established in a new area in the vineyard in 2019 to 2020 to repeat the experiment in time.

Data Collection

For all trials, four 1-m2 quadrats were placed in each plot. Normalized difference vegetation index (NDVI) and weed dry weight were measured from two quadrats in the first evaluation (Eval 1) and from the remaining two quadrats in the second evaluation (Eval 2). For NDVI measurements, the sensor was held stable at approximately 25 cm above the weed canopy for 5 seconds (Kong et al. Reference Kong, Si, Feng, Li, Wang and Sayre2009). A hand-held Trimble® GreenSeeker® (Trimble Agriculture Division, Westminster, CO, USA) sensor was used. The sensor unit has self-contained illumination in both red and near-infrared (NIR) bands and measures reflectance in the red and NIR regions of the electromagnetic spectrum (Tremblay et al. Reference Tremblay, Wang, Ma, Belec and Vigneault2009) according to Equation 1:

([1]) $${\rm{NDVI}} = \;{{{\rm{NIR}} - {\rm{Red}}} \over {{\rm{NIR}} + {\rm{Red}}}}$$

For weed dry weight evaluations, weeds were harvested at a height of 5 cm with scissors. The samples were classified by weed species and oven-dried (DHG-9025, Knowledge Research S.A., Athens, Greece) at 65 C for 48 h to measure weed biomass. Data of the dry weight of the dominant species and total weed dry weight are presented for the citrus orchard. For the olive grove and the vineyard, only the dry weights of the dominant species are presented. Conyza bonariensis and S. halepense were the dominant species at these sites, contributing more than 90% of the total weed biomass (data not shown). For all trials, cover crop samples were also harvested from two 0.25-m2 areas in each plot the day before termination. The samples were oven-dried and weighed to get an estimate of biomass production for both species.

The first evaluations (Eval 1) of NDVI and weed dry weight were carried out at fruit set (BBCH 71, 69, and 71 for the citrus trees, the olive trees, and the vines, respectively). The second evaluations (Eval 2) were conducted during the rapid growth stage of the produced fruits/grapes (BBCH 71 for the citrus trees, 71 to 75 for the olive trees and the vines). At these crucial growth stages, water stress and nutrient deficiencies can affect the number and size of the produced fruits/grapes, thus affecting the final yield of these crops more than stress conditions at later growth stages (García-Tejero Reference García-Tejero, Durán-Zuazo, Arriaga-Sevilla and Muriel-Fernández2011; Gómez-del-Campo Reference Gómez-del-Campo, Pérez-Expósito, Hammami, Centeno and Rapoport2014; Wenter et al. Reference Wenter, Zanotelli, Montagnani, Tagliavini and Andreotti2018).

Crops were harvested at fruit/grape maturity stage. In the citrus orchard, fruits were cut with scissors and measured per tree, and fruit yield per tree was estimated after fruit weight was measured (Bassal Reference Bassal2009). In the olive grove, fruits were harvested electromechanically and fruit weight per tree and per plot was measured (Bourazanis et al. Reference Bourazanis, Roussos, Argyrokastritis, Kosmas and Kerkides2016). In the vineyard, all clusters from each plot’s vines were harvested and counted. μean berry weight was also measured to estimate grape yield per vine (Taskos et al. Reference Taskos, Stamatiadis, Yvin and Jamois2019). The final fruit/grape yield values per tree/vine are presented. Exact dates for crop harvest, weed biomass, and NDVI data collection are shown in Table 3. To protect crops from insect pests and fungal infections, the mandatory spraying programs had been implemented according to the special needs of each crop.

Statistical Analysis

Data from each evaluation were subjected to an initial two-way ANOVA. Treatments, years, and their interactions were considered as fixed effects, while blocks were considered random effects. When years affected the studied parameters (P ≤ 0.05), data were analyzed to compare means between treatments in each separate year. All analyses were performed at α = 0.05 significance level using STATGRAPHICS Centurion XVI (Statgraphics Technologies Inc., The Plains, VA, USA). Mean comparisons were performed according to Fisher’s LSD test.

Weed dry weight was correlated with NDVI according to the linear model:

([2]) $$Y = A + B*X$$

where Y represents the values of weed dry weight, X represents NDVI values, A represents the intercept of the regression, and B represents the slope of the regression. These correlations were carried out between data obtained in the first (Eval 1) and second evaluations (Eval 2) in each year. For each separate year, crop yield was also correlated with weed dry weight values that were recorded in the second evaluation (Eval 2) according to the linear model (Equation 2). For the citrus orchard, total weed dry weight data (as recorded in Eval 2 of each separate year) were used to conduct the correlations. As for the other fields, the dry weight of the dominant species (as recorded in Eval 2 of each separate year) was used to conduct the correlations.

Results and Discussion

Weed Control in the Citrus Orchard

The effects of treatments and years on the dry weight of dominant weeds in the citrus orchard were significant (P ≤ 0.05; Supplementary Table S1). The effects of year on weed dry weight were significant for all the dominant species (P ≤ 0.05). Therefore, data were analyzed separately for 2018 to 2019 and 2019 to 2020 to separate the means between treatments in each year. This was carried out for every species mentioned.

Glyphosate application resulted in the lowest values of E. crus-galli dry weight in both evaluations and years. In both evaluations of 2018 to 2019 and 2019 to 2020, lower values of E. crus-galli dry weight corresponded to preemergence herbicide applications compared with cover crop residues, pelargonic acid, and mowing treatments (Figure 1A). Flazasulfuron, oxyfluorfen, and flumioxazin can effectively control summer annual grasses such as E. crus-galli that are common in fruit orchards (Plaza et al. Reference Plaza, Hoyos, Vázquez-García, Alcántara-de la Cruz and De Prado2021; Seale et al. Reference Seale, Bararpour, Bond, Gore and Golden2020). Barley and white mustard residues suppressed E. crus-galli more than pelargonic acid and mowing in the first (Eval 1) and second evaluations (Eval 2) of both years. Similar results have been reported by Kruidhof et al. (Reference Kruidhof, Bastiaans and Kropff2009). In addition, barley residues have been reported to effectively suppress E. crus-galli in Greece (Dhima et al. Reference Dhima, Vasilakoglou, Eleftherohorinos and Lithourgidis2006). Pelargonic acid (two times) was characterized by lower efficacy on E. crus-galli. This result agrees with the findings of Travlos et al. (Reference Travlos, Rapti, Gazoulis, Kanatas, Tataridas, Kakabouki and Papastylianou2020b), who reported low efficacy of pelargonic acid on grass weeds. Mowing did not reduce E. crus-galli biomass compared with the untreated in both years and evaluations. This can be attributed to the increased regrowth ability of this species after clipping (Butler et al. Reference Butler, Brouder, Johnson and Gibson2013).

Figure 1. Dry weight (g m−2) of (A) Echinochloa crus-galli, (B) Chenopodium album, (C) Solanum nigrum, and (D) Amaranthus retroflexus as recorded in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020. For each evaluation, means were compared between treatments according to Fisher’s LSD test at α = 0.05 significance level (vertical bars indicate the standard errors of the measurements).

As for C. album, the double pelargonic acid application reduced its dry weight compared with barley residues and the untreated in both evaluations of 2018 to 2019 and 2019 to 2020 (Figure 1B). The efficacy of pelargonic acid on broadleaf weeds has recently been recently highlighted (Travlos et al. Reference Travlos, Rapti, Gazoulis, Kanatas, Tataridas, Kakabouki and Papastylianou2020b). Mowing reduced C. album dry weight compared with cover crop residues and the untreated in both years and evaluations. Mowing can provide solutions in the control of annual broadleaf weeds in fruit orchards (Mia et al. Reference Mia, Massetani, Murri, Facchi, Monaci, Amadio and Neri2020). Barley and white mustard residues also showed some suppressive ability on this species, especially in 2019 to 2020. These results are partially consistent with another study in which oat (Avena sativa L.) and canola (Brassica napus L.) residues reduced C. album emergence (Campiglia et al. Reference Campiglia, Radicetti and Mancinelli2012; Radicetti et al. Reference Radicetti, Mancinelli and Campiglia2013). Glyphosate and preemergence herbicides resulted in the lowest C. album biomass. In the second evaluations of both years, the performance of preemergence herbicides was comparable with that of glyphosate. These results agree with those of other studies in which flazasulfuron, oxyfluorfen, and flumioxazin provided sufficient control of C. album under European soil and climatic conditions (Jursík et al. Reference Jursík, Andr, Holec and Soukup2011; Pacanoski et al. (Reference Pacanoski, Boškov and Mehmeti2020).

In the first evaluation (Eval 1) of 2018 to 2019, preemergence herbicide application resulted in similar S. nigrum biomass compared with the double pelargonic acid application and the single mowing treatment. In the second evaluation (Eval 2), flazasulfuron, oxyfluorfen, and flumioxazin resulted in lower S. nigrum biomass than pelargonic acid (two times) and mowing (Figure 1C). These results suggest that although weeds can escape mechanical weed control operations or postemergence herbicide applications, the soil residual activity of preemergence herbicides enables the long-term management of weeds with an extended emergence window in fruit orchards (Ribeiro et al. Reference Ribeiro, Maia, Arneson, Oliveira, Read, Ané, dos Santos and Werle2021). In the second evaluation of 2019 to 2020, oxyfluorfen and glyphosate applications resulted in the lowest values of S. nigrum biomass. These findings are in line with other studies highlighting oxyfluorfen efficacy on S. nigrum and other weeds belonging to Solanum spp. (Liang et al. Reference Liang, Jia, Dong, Li, Liu, Cao, Wang and Yang2014). As for cover crops, their residues reduced weed dry weight compared with the untreated in both years and evaluations. This outcome is partially consistent with Campiglia et al.’s (Reference Campiglia, Radicetti and Mancinelli2012) observation of the suppressive ability of barley residues on S. nigrum. Oat and canola residues were also reported to suppress S. nigrum (Radicetti et al. Reference Radicetti, Mancinelli and Campiglia2013).

Mowing, pelargonic acid (two times), and cover crop residues resulted in similar A. retroflexus biomass values, as observed in the first evaluation (Eval 1) of 2018 to 2019. Similar values corresponded to pelargonic acid (two times), mowing, and white mustard residues in the second evaluation (Eval 2). These treatments reduced A. retroflexus dry weight compared with the untreated (Figure 1D). Similar observations were made in 2019 to 2020. The satisfactory performance of mowing is in line with the recommendations of Donald (Reference Donald, Singh, Batish and Kohli2006) that timely mowing operations can be highly effective on annual broadleaf species. The low biomass values recorded for A. retroflexus after pelargonic acid applications (two times) agree with Muñoz et al. (Reference Muñoz, Torres-Pagán, Peiró, Guijarro, Sánchez-Moreiras and Verdeguer2020). Preemergence herbicides and glyphosate resulted in the lowest dry weight values for this species in both years and evaluations. Flazasulfuron, oxyfluorfen, and flumioxazin were reported in previous studies to control A. retroflexus and other weeds belonging to Amaranthus spp. (Mahoney et al. Reference Mahoney, Shropshire and Sikkema2014; Ronchi and Silva Reference Ronchi and Silva2004). Moreover, white mustard residue efficacy on A. retroflexus was comparable to that of preemergence herbicides in the second evaluation of 2019 to 2020. The suppressive effects of white mustard residues on A. retroflexus growth were also observed in a previous study by Alcántara et al. (Reference Alcántara, Pujadas and Saavedra2011) conducted in perennial orchards.

Both NDVI and total weed dry weight were affected by treatments and years (P ≤ 0.001; Supplementary Table S2) in their first (Eval 1) and second evaluations (Eval 2), as shown by the two-way ANOVAs that were conducted. Because the effects of years on the studied parameters were significant, data for each evaluation were analyzed separately to separate means between treatments in each year.

Total weed dry weight was positively correlated with NDVI in the first evaluation (Eval 1) of 2018 to 2019 (P ≤ 0.001; n = 36; Y = −189.06 + 1,131.98 * X; R = 0.966; R2 = 0.934; RMSE [root mean-square error] = 48.286; Figure 2A) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = −153.859 + 930.565 * X; R = 0.975; R2 = 0.951; RMSE = 32.972; Figure 2A). The strong and positive correlation between total weed dry weight and NDVI was also obvious in the second evaluation (Eval 2) of 2018 to 2019 (P ≤ 0.001; n = 36; Y = −285.012 + 1,590.43 * X; R = 0.973; R2 = 0.947; RMSE = 59.641; Figure 2B) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = −218.929 + 1,318.02 * X; R = 0.987; R2 = 0.975; RMSE = 33.088; Figure 2B). These correlations indicate that NDVI can be a reliable, nondestructive tool for weed control evaluation and agree with previous findings (Travlos et al. Reference Travlos, Tsekoura, Antonopoulos, Kanatas and Gazoulis2021).

Figure 2. Linear regression analyses between (A) total weed dry weight (g m−2) and normalized difference vegetation index (NDVI) in the first evaluation (Eval 1) of 2018–2019 and 2019–2020; (B) total weed dry weight (g m−2) and NDVI in the second evaluation (Eval 2) of 2018–2019 and 2019–2020; and (C) fruit yield (kg tree−1) and total weed dry weight (g m−2) in 2018–2019 and 2019–2020. For the two correlations presented in C, total weed dry weight data are derived from the second evaluation (Eval 2) of each separate year. RMSE, root mean-square error.

In both years and evaluations, glyphosate and preemergence herbicides resulted in lower values of NDVI and total weed dry weight compared with other treatments (Table 4). In the second evaluation of 2018 to 2019, oxyfluorfen performance was comparable to that of glyphosate. Flazasulfuron and oxyfluorfen performances were comparable to that of glyphosate in the second evaluation of 2019 to 2020. The efficacy of preemergence herbicides is encouraging in a time when glyphosate resistance is a growing problem in the perennial orchards of the Mediterranean region (Tahmasebi et al. Reference Tahmasebi, Alebrahim, Roldán-Gómez, da Silveira, de Carvalho, Alcántara-de la Cruz and De Prado2018; Travlos et al Reference Travlos, Kanatas, Tsekoura, Gazoulis, Papastylianou, Kakabouki and Antonopoulos2020a; Vazquez-Garcia et al. Reference Vazquez-Garcia, Palma-Bautista, Rojano-Delgado, De Prado and Menendez2020).

Table 4. Normalized difference vegetation index (NDVI), total weed dry weight (g m−2) in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020 in the citrus orchard with fruit yield (kg tree−1) data also presented as recorded in 2018–2019 and 2019–2020. a

a For each evaluation, mean separation was conducted according to Fisher’s LSD test. P-values are shown as derived from ANOVA conducted for each growing season at α = 0.05 probability level. Different letters in the same column indicate significant differences among treatments for each evaluation conducted in each separate growing season.

***P ≤ 0.001.

Pelargonic acid (two times) reduced NDVI and total weed dry weight compared with barley residues and the untreated in the first and second evaluations of 2018 to 2019. Our findings indicate that repeated applications of pelargonic acid to small-sized weeds can provide solutions for weed control in fruit orchards (Young Reference Young2004). In the second evaluation of 2019 to 2020, white mustard residues reduced NDVI and total weed biomass compared with mowing. No significant differences were detected between barley residues, mowing, and pelargonic acid. The higher weed- suppression ability of both species can be attributed to the higher rainfall in the second growing season than in the first (Table 1). The higher rainfall in the winter of 2019 to 2020 resulted in higher biomass production in both species (data not shown) and consequently higher weed suppression. These results agree with those of MacLaren et al. (Reference MacLaren, Swanepoel, Bennett, Wright and Dehnen-Schmutz2019). Our findings indicate also that mowing and cover crop residue management cannot achieve sufficient weed control as single treatments. However, these practices can contribute to weed management in perennial orchards, especially when combined with the applications of glyphosate or alternative herbicides (Alcántara et al. Reference Alcántara, Pujadas and Saavedra2011; Mia et al. Reference Mia, Massetani, Murri, Facchi, Monaci, Amadio and Neri2020; Raimondi et al. Reference Raimondi, Constantin, Mendes, Oliveira and Rios2020).

Fruit yield was also affected by treatments and years (P ≤ 0.001; Supplementary Table S2), as shown in the results of the initial two-way ANOVA. The significantly higher fruit yield per tree in the second growing season is mainly attributed to the phenomenon of alternate bearing exhibited by the citrus trees (Bassal Reference Bassal2009). Because years affected fruit yield, data were analyzed separately for each year to compare means among treatments in each year.

In 2018 to 2019, glyphosate, flazasulfuron, oxyfluorfen, and flumioxazin resulted in the highest fruit yields per tree. Intermediate fruit yield values corresponded to pelargonic acid (two times), mowing, and barley residues and white mustard residues; the lowest yields were recorded in the untreated plots. The same observations were made in 2019 to 2020. The highest fruit yields in plots treated with glyphosate and preemergence herbicides are attributed to the higher levels of weed control provided by these treatments. This is further supported by the negative correlations obtained between fruit yield and total weed dry weight in both 2018 to 2019 (P ≤ 0.001; n = 36; Y = 35.387 − 0.00627696 * X; R = −0.887; R2 = 0.787; RMSE = 0.845; Figure 2C) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = 41.0395 − 0.00874668 * X; R = −0.737; R2 = 0.543; RMSE = 1.675; Figure 2C). These results are consistent with those of Merwin and Ray (Reference Merwin and Ray1997), who observed 62% yield loss due to weed interference in a newly planted fruit orchard. In addition, our findings suggest that weed competition can be an important obstacle limiting citrus crops’ productivity under Mediterranean soil and climatic conditions, as also mentioned by Onen et al. (Reference Onen, Akdeniz, Farooq, Hussain and Ozaslan2018).

Weed Control in the Olive Grove

NDVI and C. bonariensis dry weight were affected by treatments and years in both their first and second evaluations (P ≤ 0.001; Supplementary Table S3) given the results of the two-way ANOVAs that were conducted. Because the effects of years on both parameters were significant, data for each evaluation were analyzed separately for each experimental year.

Conyza bonariensis dry weight was positively correlated with NDVI in the first evaluation (Eval 1) of 2018 to 2019 (P ≤ 0.001; n = 36; Y = −182.396 + 833.892 * X; R = 0.901641; R2 = 0.812; RMSE = 72.446; Figure 3A) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = −158.547 + 727.451 * X; R = 0.901; R2 = 0.812; RMSE = 63.764; Figure 3A). The correlation between C. bonariensis biomass and NDVI was strong and positive in the second evaluation (Eval 2) of 2018 to 2019 (P ≤ 0.001; n = 36; Y = −203.347 + 910.336 * X; R = 0.894; R2 = 0.799; RMSE = 82.823; Figure 3B) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = −166.913 + 769.842 * X; R = 0.908; R2 = 0.825; RMSE = 64.731; Figure 3B). These results suggest that NDVI values were in accordance with those of weed biomass and agree with other studies in which this vegetation index was effectively used to estimate herbicide efficacy (Lewis et al. Reference Lewis, Jeffries, Gannon, Richardson and Yelverton2014; Travlos et al. Reference Travlos, Tsekoura, Antonopoulos, Kanatas and Gazoulis2021).

Figure 3. Linear regression analyses between (A) Conyza bonariensis dry weight (g m−2) and normalized difference vegetation index (NDVI) in the first evaluation (Eval 1) of 2018–2019 and 2019–2020; (B) Conyza bonariensis dry weight (g m−2) and NDVI in the second evaluation (Eval 2) of 2018–2019 and 2019–2020; and (C) fruit yield (kg tree−1) and Conyza bonariensis dry weight (g m−2) in 2018–2019 and 2019–2020. For the two correlations presented in C, weed dry weight data are derived from the second evaluation (Eval 2) of each separate year. RMSE, root mean-square error.

In both years and evaluations, the lowest NDVI and C. bonariensis dry weight values were recorded in plots treated with glyphosate (Table 5). These results highlight the role of glyphosate in the control of this noxious species, which can cause severe yield losses in important perennial crops (Bajwa et al. Reference Bajwa, Sadia, Ali, Jabran, Peerzada and Chauhan2016; Curt et al. Reference Curt, Mauri, Sanz, Cano-Ruiz, Del Monte, Aguado and Sánchez2017). Flazasulfuron resulted in lower NDVI and C. bonariensis biomass than oxyfluorfen, penoxsulam + florasulam, and flumioxazin. These results were common in both evaluations of 2018 to 2019 and 2019 to 2020 and highlight the potential of flazasulfuron as an alternative herbicide for C. bonariensis control and are consistent with those of another recent study (Tahmasebi et al. Reference Tahmasebi, Alebrahim, Roldán-Gómez, da Silveira, de Carvalho, Alcántara-de la Cruz and De Prado2018). The same authors also reported that flazasulfuron was more effective on C. bonariensis compared with oxyfluorfen and flumioxazin, as also observed in the present study. Penoxsulam + florasulam reduced NDVI and C. bonariensis biomass compared with flumioxazin. This was observed in the first (Eval 1) and second evaluations (Eval 2) of 2018 to 2019 and validated in the corresponding evaluations of 2019 to 2020. The higher efficacy of the prepackaged triazolopyrimidine mixture on C. bonariensis compared with flumioxazin was reported in previous studies conducted in olive groves in western Greece (Travlos et al. Reference Travlos, Lysandrou and Apostolidis2014).

Table 5. Normalized difference vegetation index (NDVI) and Conyza bonariensis dry weight (g m−2) in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020 in the olive grove with fruit yield data (kg tree−1) also presented as recorded in 2018–2019 and 2019–2020. a

a For each evaluation, mean separation was conducted according to Fisher’s LSD test. P-values are shown as derived from ANOVA conducted for each growing season at α = 0.05 probability level. Different letters in the same column indicate significant differences between treatments for each evaluation conducted in each separate year.

***P ≤ 0.001.

Pelargonic acid (two times) did not cause noticeable reductions in weed biomass in all the evaluations carried out in the olive grove, in contrast to the results obtained in the citrus orchard. This difference between sites can be attributed to the presence of C. bonariensis as the dominant species in the olive grove. The presence of buds in the C. bonariensis taproot enable the rapid regrowth of this species after clipping (Wu Reference Wu2007). Regarding cover crop residues, they caused some reductions in NDVI and weed biomass compared with pelargonic acid and the untreated. These results were recorded in 2018 to 2019 and confirmed in 2019 to 2020. Recent studies reported C. bonariensis suppression after mulching with the residues of winter-grown cereal and cruciferous cover crops (Cholette et al. Reference Cholette, Soltani, Hooker, Robinson and Sikkema2018; Pittman et al. Reference Pittman, Barney and Flessner2019). In these studies, cover crop residues suppressed C. bonariensis in the short time after cover crop termination, and then weeds were suppressed in the subsequent cash crops. Because this is not possible in perennial orchards, supplementary glyphosate applications are required to control this species later in the growing season. However, integrated weed management systems including combinations of cultural practices and alternative herbicides should be further evaluated for the management of this species. This is further supported by glyphosate-resistance cases reported for C. bonariensis accessions in the perennial crops of the Mediterranean region (Amaro-Blanco et al. Reference Amaro-Blanco, Fernández-Moreno, Osuna-Ruiz, Bastida and De Prado2018; Tahmasebi et al. Reference Tahmasebi, Alebrahim, Roldán-Gómez, da Silveira, de Carvalho, Alcántara-de la Cruz and De Prado2018; Travlos and Chachalis Reference Travlos and Chachalis2010).

The results of the two-way ANOVA indicated that treatments and years influenced fruit yield per tree at a significant point (P ≤ 0.001; Supplementary Table S3). Because the effects of years on fruit yield per tree were significant, yield data were analyzed separately for the first and second experimental years to separate means among treatments in each year. Olive trees were in the high-yield year of their alternate bearing cycle, and this mainly explains the higher yields of 2018 to 2019 (Bourazanis et al. Reference Bourazanis, Roussos, Argyrokastritis, Kosmas and Kerkides2016).

In 2018 to 2019, glyphosate resulted in the highest fruit yield. Flazasulfuron, penoxsulam + florasulam, oxyfluorfen, and white mustard residues increased fruit yield compared with pelargonic acid (two times), flumioxazin, and barley residues (Table 5). In 2019 to 2020, glyphosate and flazasulfuron provided the highest yields. The yield values that corresponded to penoxsulam + florasulam, oxyfluorfen, and white mustard residues were close to the value that corresponded to flazasulfuron. Pelargonic acid (two times), flumioxazin, and barley residues resulted in lower fruit yield. The lowest yield values were recorded in the untreated plots in both years. These results trend with other studies in which C. bonariensis control was essential to avoid severe yield losses in perennial crops (Curt et al. Reference Curt, Mauri, Sanz, Cano-Ruiz, Del Monte, Aguado and Sánchez2017). The effects of C. bonariensis competition on the productivity of the olive trees is further validated by the strong and negative correlations obtained between fruit yield per tree and C. bonariensis dry weight in 2018 to 2019 (P ≤ 0.001; n = 36; Y = 15.0611 − 0.0065342 * X; R = −0.953; R2 = 0.909; RMSE = 0.380; Figure 3C) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = 9.82386 − 0.0074612 * X; R = −0.961; R2 = 0.924; RMSE = 0.330; Figure 3C).

Weed Control in the Vineyard

The effects of treatments and years on NDVI and S. halepense dry weight in both their first and second evaluations were significant (P ≤ 0.001; Supplementary Table S4), as shown in the results of the two-way ANOVAs that were carried out to check whether data could be pooled across growing seasons. As a result, data of each evaluation were analyzed to compare means between treatments in each experimental year.

Sorghum halepense dry weight correlations with NDVI were noticeable in the first evaluation (Eval 1) of 2018 to 2019 (P ≤ 0.001; n = 36; Y = −293.29 + 1,324.79 * X; R = 0.982; R2 = 0.966; RMSE = 56.632; Figure 4A) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = −249.16 + 1,227.34 * X; R = 0.979; R2 = 0.958; RMSE = 59.421; Figure 4A). The correlations between S. halepense dry weight and NDVI were strong and positive in the second evaluation (Eval 2) of 2018 to 2019 (P ≤ 0.001; n = 36; Y = −334,788 + 1,395.94 * X; R = 0.986; R2 = 0.973; RMSE = 53.286; Figure 4B) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = −264.444 + 1,283.92 * X; R = 0.977; R2 = 0.955; RMSE = 65.777; Figure 4B). In both years and evaluations, glyphosate, cycloxydim, quizalofop-p-ethyl, and propaquizafop resulted in the lowest NDVI and S. halepense dry weight values compared with mowing (two times), pelargonic acid (two times), barley residues, white mustard residues, and the untreated (Table 6). All the acetyl-CoA carboxylase inhibitors tested were as effective as glyphosate. The satisfactory performance of cycloxydim is consistent with the findings of Scarabel et al. (Reference Scarabel, Panozzo, Savoia and Sattin2014), who also observed excellent efficacy of this herbicide on S. halepense. The low NDVI and S. halepense dry weight values in plots treated with quizalofop-p-ethyl agree with Johnson et al. (Reference Johnson, Li and Wait2003). Propaquizafop achieved sufficient S. halepense control in the field trials of Haitas et al. (Reference Haitas, Kotoula-Syka and Eleftherohorinos1995), as also observed in the current study.

Figure 4. Linear regression analyses between (A) Sorghum halepense dry weight (g m−2) and normalized difference vegetation index (NDVI) in the first evaluation (Eval 1) of 2018–2019 and 2019–2020; (B) Sorghum halepense dry weight (g m−2) and NDVI in the second evaluation (Eval 2) of 2018–2019 and 2019–2020; and (C) fruit yield (kg tree−1) and Sorghum halepense dry weight (g m−2) in 2018–2019 and 2019–2020. For the two correlations presented in C, weed dry weight data are derived from the second evaluation (Eval 2) of each separate year. RMSE, root mean-square error.

Table 6. Normalized difference vegetation index (NDVI) and Sorghum halepense dry weight (g m−2) in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020 in the vineyard with grape yield data (kg vine−1) are also presented as recorded in 2018–2019 and 2019–2020. a

a For each evaluation, mean separation was conducted according to Fisher’s LSD test. P-values are shown as derived from ANOVA conducted for each growing season at α = 0.05 probability level. Different letters in the same column indicate significant differences between treatments for each evaluation conducted in each separate year.

***P ≤ 0.001.

Mowing (two times) and pelargonic acid (two times) caused slight reductions in S. halepense biomass compared with the untreated. No significant differences were detected in NDVI between these treatments. These results were common in the first (Eval 1) and the second evaluations (Eval 2) of 2018 to 2019 and 2019 to 2020. Mowing was more effective in the citrus orchard, where it controlled erect, annual broadleaf weeds. In contrast, the regrowth ability of S. halepense after clipping can explain the higher NDVI and weed biomass values recorded after mowing in the vineyard. It should be also noted that the poor performance of mowing for controlling grass weeds was similar in these two sites, given the regrowth ability of S. halepense and E. crus-galli after clipping (Butler et al. Reference Butler, Brouder, Johnson and Gibson2013; Horowitz Reference Horowitz1972). Pelargonic acid was also much less effective in this field compared with the other field trials presented. Although this natural herbicide could control annual broadleaves in the citrus orchard and C. bonariensis individuals that were in very early growth stages in the olive grove (visual observations), S. halepense regrew rapidly after treatment in the vineyard. Another explanation for this difference among sites is that pelargonic acid is a contact-type herbicide that tends to be more effective on broadleaf weeds with increased leaf area compared with narrow-leaved grass species (Travlos et al. Reference Travlos, Rapti, Gazoulis, Kanatas, Tataridas, Kakabouki and Papastylianou2020b).

Barley and white mustard residues caused some reductions in NDVI and weed dry weight in comparison to the untreated, mowing (two times), and pelargonic acid (two times) treatments in both years and evaluations. Cover crop residue might have suppressed the emergence of S. halepense plants derived from seed, while these treatments could not suppress the emergence of S. halepense plants derived from rhizome. This can also explain the lower efficacy of cover crop residues in this site compared with sites dominated by annual species reproduced exclusively by seed (i.e., citrus orchard, olive grove). Regarding the effects of mulching on S. halepense, an interesting approach by Uremis et al. (Reference Uremis, Arslan, Uludag and Sangun2009) suggested that the residues of cover crops belonging to the botanical family of Brassicaceae can be incorporated into the soil to suppress rhizome growth. Further research is required to evaluate such practices for S. halepense control under different soil and climatic conditions. This applies also for other nonchemical control methods such as mowing. Although two mowing passes were not effective in this site, repeated and frequent mowing controls S. halepense seedlings and reduces rhizome growth and shoot regrowth after clipping (Travlos et al. Reference Travlos, Montull, Kukorelli, Malidza, Dogan, Cheimona, Antonopoulos, Kanatas, Zannopoulos and Peteinatos2019; Warwick and Black Reference Warwick and Black1983). For instance, mowing S. halepense at the ground surface every 2 wk from spring to autumn caused significant reductions on S. halepense dry weight and rhizome length in early studies (Horowitz Reference Horowitz1972). The need to optimize integrated management systems to eradicate S. halepense is further supported by cases reporting the presence of S. halepense accessions with resistance to selective herbicides and glyphosate in the perennial orchards of the Mediterranean zone (Heap Reference Heap2021; Vazquez-Garcia et al. Reference Vazquez-Garcia, Palma-Bautista, Rojano-Delgado, De Prado and Menendez2020). To maintain efficacy over time, timely herbicide applications should be combined with cultural practices such as mowing (Travlos et al. Reference Travlos, Montull, Kukorelli, Malidza, Dogan, Cheimona, Antonopoulos, Kanatas, Zannopoulos and Peteinatos2019). Such systems have been reported to provide sufficient control of perennial rhizomatous weed species that are troublesome in perennial crops (Raimondi et al. Reference Raimondi, Constantin, Mendes, Oliveira and Rios2020).

As for grape yield, it was also affected by treatments and years (P ≤ 0.001; Supplementary Table S4), as shown in the results of the initial two-way ANOVA. Therefore, means were compared among treatments in each separate year. In 2018 to 2019, the highest grape yields were observed in plots treated with glyphosate and cycloxydim. Quizalofop-p-ethyl and propaquizafop applications led to grape yield values that were comparable to those with cycloxydim. Lower yields were recorded in plots mulched with cover crop residues. The lowest yields corresponded to mowing (two times), pelargonic acid (two times), and the untreated. These results were also observed in 2019 to 2020. The herbicide treatments that most reduced S. halepense dry weight resulted in the highest grape yields, as highlighted by the strong and negative correlations observed between grape yield per vine and S. halepense dry weight in both 2018 to 2019 (P ≤ 0.001; n = 36; Y = 5.27244 − 0.00258846 * X; R = −0.955846; R2 = 0.913642; RMSE = 0.260154; Figure 4C) and 2019 to 2020 (P ≤ 0.001; n = 36; Y = 5.35936 − 0.00276311 * X; R = −0.959; R2 = 0.921; RMSE = 0.252; Figure 4C). The significant effect of weed competition on grape yield was expected, because S. halepense is a troublesome species that can compete for water and nutrients and subsequently reduce crop productivity in vineyards (Travlos et al. Reference Travlos, Bilalis, Katsenios and De Prado2018).

Acknowledgments

This research received no specific grant from any funding agency or the commercial or not-for-profit sectors. No conflicts of interest have been declared.

Supplementary material

To view supplementary material for this article, please visit https://doi.org/10.1017/wsc.2021.55

Footnotes

Associate Editor: William Vencill, University of Georgia

References

Alcántara, C, Pujadas, A, Saavedra, M (2011) Management of Sinapis alba subsp. mairei winter cover crop residues for summer weed control in southern Spain. Crop Prot 30:12391244 10.1016/j.cropro.2011.04.007CrossRefGoogle Scholar
Amaro-Blanco, I, Fernández-Moreno, PT, Osuna-Ruiz, MD, Bastida, F, De Prado, R (2018) Mechanisms of glyphosate resistance and response to alternative herbicide-based management in populations of the three Conyza species introduced in southern Spain. Pest Manag Sci 74:19251937 10.1002/ps.4896CrossRefGoogle ScholarPubMed
Bajwa, AA, Sadia, S, Ali, HH, Jabran, K, Peerzada, AM, Chauhan, BS (2016) Biology and management of two important Conyza weeds: a global review. Environ Sci Pollut Res 23:2469424710 10.1007/s11356-016-7794-7CrossRefGoogle ScholarPubMed
Bassal, MA (2009) Growth, yield and fruit quality of ‘Marisol’ clementine grown on four rootstocks in Egypt. Sci Hortic 119:132137 10.1016/j.scienta.2008.07.020CrossRefGoogle Scholar
Beckie, HJ, Flower, KC, Ashworth, MB (2020) Farming without glyphosate? Plants 9:96 10.3390/plants9010096CrossRefGoogle ScholarPubMed
Bourazanis, G, Roussos, PA, Argyrokastritis, I, Kosmas, C, Kerkides, P (2016) Evaluation of the use of treated municipal waste water on the yield, oil quality, free fatty acids’ profile and nutrient levels in olive trees cv Koroneiki, in Greece. Agric Water Manag 163:18 10.1016/j.agwat.2015.08.023CrossRefGoogle Scholar
Butler, RA, Brouder, SM, Johnson, WG, Gibson, KD (2013). Response of four summer annual weed species to mowing frequency and height. Weed Technol 27:798802 10.1614/WT-D-12-00112.1CrossRefGoogle Scholar
Campiglia, E, Radicetti, E, Mancinelli, R (2012) Weed control strategies and yield response in a pepper crop (Capsicum annuum L.) mulched with hairy vetch (Vicia villosa Roth.) and oat (Avena sativa L.) residues. Crop Prot 33:6573 10.1016/j.cropro.2011.09.016CrossRefGoogle Scholar
Cholette, TB, Soltani, N, Hooker, DC, Robinson, DE, Sikkema, PH (2018) Suppression of glyphosate-resistant Canada fleabane (Conyza canadensis) in corn with cover crops seeded after wheat harvest the previous year. Weed Technol 32:244250 10.1017/wet.2018.19CrossRefGoogle Scholar
Curt, MD, Mauri, PV, Sanz, M, Cano-Ruiz, J, Del Monte, JP, Aguado, PL, Sánchez, J (2017) The ability of the Arundo donax crop to compete with weeds in central Spain over two growing cycles. Ind Crops Prod 108:8694 10.1016/j.indcrop.2017.06.015CrossRefGoogle Scholar
Dhima, KV, Vasilakoglou, IB, Eleftherohorinos, IG, Lithourgidis, AS (2006) Allelopathic potential of winter cereals and their cover crop mulch effect on grass weed suppression and corn development. Crop Sci 46:345352 10.2135/cropsci2005-0186CrossRefGoogle Scholar
Didon, UM, Kolseth, AK, Widmark, D, Persson, P (2014) Cover crop residues—effects on germination and early growth of annual weeds. Weed Sci 62:294302 10.1614/WS-D-13-00117.1CrossRefGoogle Scholar
Donald, WW (2006) Mowing for weed management. Pages 357400 in Singh, HP, Batish, DR, Kohli, KR, eds. Handbook of Sustainable Weed Management. Boca Raton, FL: CRC Press Google Scholar
Duke, SO (2020) Glyphosate: environmental fate and impact. Weed Sci 68:201207 10.1017/wsc.2019.28CrossRefGoogle Scholar
Duke, SO, Powles, SB (2008) Glyphosate: a once-in-a-century herbicide. Pest Manag Sci 64:319325 10.1002/ps.1518CrossRefGoogle ScholarPubMed
Farooq, M, Flower, KC, Jabran, K, Wahid, A, Siddique, KH (2011) Crop yield and weed management in rainfed conservation agriculture. Soil Till Res 117:172183 10.1016/j.still.2011.10.001CrossRefGoogle Scholar
Fogliatto, S, Ferrero, A, Vidotto, F, Sparks, D (2020) Current and future scenarios of glyphosate use in Europe: are there alternatives? Pages 219–278 in Sparks DL, ed. Advances in Agronomy. Volume 163. Cambridge: Academic Press10.1016/bs.agron.2020.05.005CrossRefGoogle Scholar
García-Tejero, I, Durán-Zuazo, VH, Arriaga-Sevilla, J, Muriel-Fernández, JL (2011) Impact of water stress on citrus yield. Agron Sustain Dev 32:651659 10.1007/s13593-011-0060-yCrossRefGoogle Scholar
Gerhards, R, Schappert, A (2020) Advancing cover cropping in temperate integrated weed management. Pest Manag Sci 76:4246 10.1002/ps.5639CrossRefGoogle ScholarPubMed
Gómez-del-Campo, M, Pérez-Expósito, , Hammami, SB, Centeno, A, Rapoport, HF (2014) Effect of varied summer deficit irrigation on components of olive fruit growth and development. Agric Water Manag 137:8491 10.1016/j.agwat.2014.02.009CrossRefGoogle Scholar
Haitas, VC, Kotoula-Syka, E., Eleftherohorinos, IG (1995) Influence of propaquizafop application rate and time on Sorghum halepense (L.) Pers. control and cotton (Gossypium hirsutum) yield. Weed Res 35:16 10.1111/j.1365-3180.1995.tb02009.xCrossRefGoogle Scholar
Heap, I (2021) The International Herbicide-Resistant Weed Database. http://weedscience.org/Home.aspx. Accessed: July 29, 2021Google Scholar
Horowitz, M (1972) Effects of frequent clipping on three perennial weeds, Cynodon dactylon (L.) Pers., Sorghum halepense (L.) Pers. and Cyperus rotundus L. Exp Agric 8:225234 10.1017/S0014479700005299CrossRefGoogle Scholar
Johnson, WG, Li, J, Wait, JD (2003) Johnsongrass control, total nonstructural carbohydrates in rhizomes, and regrowth after application of herbicides used in herbicide-resistant corn (Zea mays). Weed Technol 17:3641 10.1614/0890-037X(2003)017[0036:JCTNCI]2.0.CO;2CrossRefGoogle Scholar
Jursík, M, Andr, J, Holec, J, Soukup, J (2011) Efficacy and selectivity of post-emergent application of flumioxazin and oxyfluorfen in sunflower. Plant Soil Environ 57:532539 10.17221/285/2011-PSECrossRefGoogle Scholar
Kong, L, Si, J, Feng, B, Li, S, Wang, F, Sayre, K (2009) Differential responses of two types of winter wheat (Triticum aestivum L.) to autumn-and spring-applied mesosulfuron-methyl. Crop Prot 28:387392 10.1016/j.cropro.2008.12.008CrossRefGoogle Scholar
Kruidhof, HM, Bastiaans, L, Kropff, MJ (2009) Cover crop residue management for optimizing weed control. Plant Soil 318:169184 10.1007/s11104-008-9827-6CrossRefGoogle Scholar
Kudsk, P, Mathiassen, SK (2020) Pesticide regulation in the European Union and the glyphosate controversy. Weed Sci 68:214222 10.1017/wsc.2019.59CrossRefGoogle Scholar
Lewis, D, Jeffries, M, Gannon, T, Richardson, R, Yelverton, F (2014) Persistence and bioavailability of aminocyclopyrachlor and clopyralid in turfgrass clippings: recycling clippings for additional weed control. Weed Sci 62:493500 10.1614/WS-D-13-00119.1CrossRefGoogle Scholar
Liang, Y, Jia, H, Dong, X, Li, J, Liu, Y, Cao, Z, Wang, C, Yang, D (2014) The effects of four soil-applied herbicides on control of Solanum nigrum L. and safety evaluation. Acta Agric Univ Jiangxi 36:102108 Google Scholar
Ligoxigakis, EK, Vakalounakis, DJ, Thanassoulopoulos, CC (2002) Weed hosts of Verticillium dahliae in Crete: Susceptibility, symptomatology and significance. Phytoparasitica 30:511518 10.1007/BF02979756CrossRefGoogle Scholar
MacLaren, C, Swanepoel, P, Bennett, J, Wright, J, Dehnen-Schmutz, K (2019) Cover crop biomass production is more important than diversity for weed suppression. Crop Sci 59:733748 10.2135/cropsci2018.05.0329CrossRefGoogle Scholar
Mahoney, KJ, Shropshire, C, Sikkema, PH (2014) Weed management in conventional- and no-till soybean using flumioxazin/pyroxasulfone. Weed Technol 28:298306 10.1614/WT-D-13-00128.1CrossRefGoogle Scholar
Merwin, IA, Ray, JA (1997) Spatial and temporal factors in weed interference with newly planted apple trees. Hort Sci 32:633637 Google Scholar
Mia, MJ, Massetani, F, Murri, G, Facchi, J, Monaci, E, Amadio, L, Neri, D (2020) Integrated weed management in high density fruit orchards. Agronomy 10:1492 10.3390/agronomy10101492CrossRefGoogle Scholar
Miller, DK, Griffin, JL, Richard, EP Jr (1998) Johnsongrass (Sorghum halepense) control and rainfastness with glyphosate and adjuvants. Weed Technol 12:617622 10.1017/S0890037X00044468CrossRefGoogle Scholar
Muñoz, M, Torres-Pagán, N, Peiró, R, Guijarro, R, Sánchez-Moreiras, AM, Verdeguer, M (2020) Phytotoxic effects of three natural compounds: pelargonic acid, carvacrol, and cinnamic aldehyde, against problematic weeds in Mediterranean crops. Agronomy 10:791 10.3390/agronomy10060791CrossRefGoogle Scholar
Onen, H, Akdeniz, M, Farooq, S, Hussain, M, Ozaslan, C (2018) Weed flora of citrus orchards and factors affecting its distribution in western Mediterranean region of Turkey. Planta Daninha 36:018172126 10.1590/s0100-83582018360100036CrossRefGoogle Scholar
Pacanoski, Z, Boškov, K, Mehmeti, AB (2020) Replace of the EPOST glyphosate with pre herbicides and application of different LPOST glyphosate rates for weed control in established vineyard. Acta Agric Slov 116:299310 10.14720/aas.2020.116.2.1910CrossRefGoogle Scholar
Pittman, KB, Barney, JN, Flessner, ML (2019) Horseweed (Conyza canadensis) suppression from cover crop mixtures and fall-applied residual herbicides. Weed Technol 33:303311 10.1017/wet.2018.111CrossRefGoogle Scholar
Plaza, G, Hoyos, V, Vázquez-García, J. G., Alcántara-de la Cruz, R, De Prado, R (2021) First case of multiple resistance to EPSPS and PSI in Eleusine indica (L.) Gaertn. collected in rice and herbicide-resistant crops in Colombia. Agronomy 11:96 10.3390/agronomy11010096CrossRefGoogle Scholar
Radicetti, E, Mancinelli, R, Campiglia, E (2013) Impact of managing cover crop residues on the floristic composition and species diversity of the weed community of pepper crop (Capsicum annuum L.). Crop Prot 44:109119 10.1016/j.cropro.2012.10.017CrossRefGoogle Scholar
Raimondi, RT, Constantin, J, Mendes, RR, Oliveira, RS Jr, Rios, FA (2020) Glyphosate-resistant sourgrass management programs associating mowing and herbicides. Planta Daninha 38:020215928 10.1590/s0100-83582020380100033CrossRefGoogle Scholar
Ribeiro, VHV, Maia, LSG, Arneson, NJ, Oliveira, MC, Read, HW, Ané, JM, dos Santos, JB, Werle, R (2021) Influence of PRE-emergence herbicides on soybean development, root nodulation and symbiotic nitrogen fixation. Crop Prot 144:105576 10.1016/j.cropro.2021.105576CrossRefGoogle Scholar
Ronchi, CP, Silva, AA (2004) Weed control in young coffee plantations through post emergence herbicide application onto total area. Planta Daninha 22:607615 10.1590/S0100-83582004000400016CrossRefGoogle Scholar
Scarabel, L, Panozzo, S, Savoia, W, Sattin, M (2014) Target-site ACCase-resistant johnsongrass (Sorghum halepense) selected in summer dicot crops. Weed Technol 28:307315 10.1614/WT-D-13-00137.1CrossRefGoogle Scholar
Scavo, A, Mauromicale, G (2020) Integrated weed management in herbaceous field crops. Agronomy 10:466 10.3390/agronomy10040466CrossRefGoogle Scholar
Seale, JW, Bararpour, T, Bond, JA, Gore, J, Golden, BR (2020) Evaluation of preemergence and postemergence herbicide programs on weed control and weed seed suppression in Mississippi peanut (Arachis hypogea). Agronomy 10:1058 10.3390/agronomy10081058CrossRefGoogle Scholar
Tahmasebi, BK, Alebrahim, MT, Roldán-Gómez, RA, da Silveira, HM, de Carvalho, LB, Alcántara-de la Cruz, R, De Prado, R (2018) Effectiveness of alternative herbicides on three Conyza species from Europe with and without glyphosate resistance. Crop Prot 112:350355 10.1016/j.cropro.2018.06.021CrossRefGoogle Scholar
Taskos, D, Stamatiadis, S, Yvin, JC, Jamois, F (2019) Effects of an Ascophyllum nodosum (L.) Le Jol. extract on grapevine yield and berry composition of a Merlot vineyard. Sci Hortic 250:2732 10.1016/j.scienta.2019.02.030CrossRefGoogle Scholar
Thanou, ZN, Kontogiannis, EG, Tsagkarakis, AE (2021) Impact of weeds on Auchenorrhyncha incidence and species richness in citrus orchards. Phytoparasitica 49:333347 10.1007/s12600-020-00857-wCrossRefGoogle Scholar
Travlos, I, Kanatas, P, Tsekoura, A, Gazoulis, I, Papastylianou, P, Kakabouki, I, Antonopoulos, N (2020a) Efficacy of different herbicides on Echinochloa colona (L.) Link control and the first case of its glyphosate resistance in Greece. Agronomy 10:1056 10.3390/agronomy10071056CrossRefGoogle Scholar
Travlos, I, Rapti, E, Gazoulis, I, Kanatas, P, Tataridas, A, Kakabouki, I, Papastylianou, P (2020b) The herbicidal potential of different pelargonic acid products and essential oils against several important weed species. Agronomy 10:1687 10.3390/agronomy10111687CrossRefGoogle Scholar
Travlos, I, Tsekoura, A, Antonopoulos, N, Kanatas, P, Gazoulis, I (2021) Novel sensor-based method (quick test) for the in-season rapid evaluation of herbicide efficacy under real field conditions in durum wheat. Weed Sci 69:147160 10.1017/wsc.2021.8CrossRefGoogle Scholar
Travlos, IS, Bilalis, DJ, Katsenios, N, De Prado, R (2018) Sustainable weed control in vineyards. Pages 526–542 in Korres NE, Burgos NR, Duke SO, eds. Weed Control: Sustainability, Hazards, and Risks in Cropping Systems Worldwide. Boca Raton, FL: CRC Press10.1201/9781315155913-28CrossRefGoogle Scholar
Travlos, IS, Chachalis, D (2010) Glyphosate-resistant hairy fleabane (Conyza bonariensis) is reported in Greece. Weed Technol 24:569573 10.1614/WT-D-09-00080.1CrossRefGoogle Scholar
Travlos, IS, Lysandrou, M, Apostolidis, V (2014) Efficacy of the herbicide GF-2581 (penoxsulam + florasulam) against broadleaf weeds in olives. Plant Soil Environ 60:574579 Google Scholar
Travlos, IS, Montull, JM, Kukorelli, G, Malidza, G, Dogan, MN, Cheimona, N, Antonopoulos, N, Kanatas, PJ, Zannopoulos, S, Peteinatos, G (2019) Key aspects on the biology, ecology and impacts of Johnsongrass [Sorghum halepense (L.) Pers] and the role of glyphosate and non-chemical alternative practices for the management of this weed in Europe. Agronomy 9:717 10.3390/agronomy9110717CrossRefGoogle Scholar
Tremblay, N, Wang, Z, Ma, BL, Belec, C, Vigneault, P (2009) A comparison of crop data measured by two commercial sensors for variable-rate nitrogen application. Precis Agric 10:145161 10.1007/s11119-008-9080-2CrossRefGoogle Scholar
Tursun, N, Işık, D, Demir, Z, Jabran, K (2018) Use of living, mowed, and soil-incorporated cover crops for weed control in apricot orchards. Agronomy 8:150 10.3390/agronomy8080150CrossRefGoogle Scholar
Uremis, I, Arslan, M, Uludag, A, Sangun, M (2009) Allelopathic potentials of residues of 6 brassica species on johnsongrass [Sorghum halepense (L.) Pers.]. Afr J Biotechnol 8:34973501 Google Scholar
Vazquez-Garcia, JG, Palma-Bautista, C, Rojano-Delgado, AM, De Prado, R, Menendez, J (2020) The first case of glyphosate resistance in johnsongrass (Sorghum halepense (L.) Pers.) in Europe. Plants 9:313 10.3390/plants9030313CrossRefGoogle ScholarPubMed
Warwick, SI, Black, LD (1983) The biology of Canadian weeds: 61. Sorghum halepense (L.) Pers. Can J Plant Sci 63:9971014.10.4141/cjps83-125CrossRefGoogle Scholar
Wenter, A, Zanotelli, D, Montagnani, L, Tagliavini, M, Andreotti, C (2018) Effect of different timings and intensities of water stress on yield and berry composition of grapevine (cv. Sauvignon blanc) in a mountain environment. Sci Hortic 236:137145.10.1016/j.scienta.2018.03.037CrossRefGoogle Scholar
Wu, H (2007) The biology of Australian weeds 49. Conyza bonariensis (L.) Cronquist. Plant Prot Q 22:122131 Google Scholar
Young, SL (2004) Natural product herbicides for control of annual vegetation along roadsides. Weed Technol 18:580587 10.1614/WT-03-094R3CrossRefGoogle Scholar
Figure 0

Table 1. Mean monthly temperature (C) and overall monthly precipitation (mm) at each site during 2018–2019 and 2019–2020.

Figure 1

Table 2. Information regarding herbicides applied in each site.

Figure 2

Table 3. Dates of key field activities carried out at each site for each experimental year (2018–2019 and 2019–2020).

Figure 3

Figure 1. Dry weight (g m−2) of (A) Echinochloa crus-galli, (B) Chenopodium album, (C) Solanum nigrum, and (D) Amaranthus retroflexus as recorded in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020. For each evaluation, means were compared between treatments according to Fisher’s LSD test at α = 0.05 significance level (vertical bars indicate the standard errors of the measurements).

Figure 4

Figure 2. Linear regression analyses between (A) total weed dry weight (g m−2) and normalized difference vegetation index (NDVI) in the first evaluation (Eval 1) of 2018–2019 and 2019–2020; (B) total weed dry weight (g m−2) and NDVI in the second evaluation (Eval 2) of 2018–2019 and 2019–2020; and (C) fruit yield (kg tree−1) and total weed dry weight (g m−2) in 2018–2019 and 2019–2020. For the two correlations presented in C, total weed dry weight data are derived from the second evaluation (Eval 2) of each separate year. RMSE, root mean-square error.

Figure 5

Table 4. Normalized difference vegetation index (NDVI), total weed dry weight (g m−2) in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020 in the citrus orchard with fruit yield (kg tree−1) data also presented as recorded in 2018–2019 and 2019–2020.a

Figure 6

Figure 3. Linear regression analyses between (A) Conyza bonariensis dry weight (g m−2) and normalized difference vegetation index (NDVI) in the first evaluation (Eval 1) of 2018–2019 and 2019–2020; (B) Conyza bonariensis dry weight (g m−2) and NDVI in the second evaluation (Eval 2) of 2018–2019 and 2019–2020; and (C) fruit yield (kg tree−1) and Conyza bonariensis dry weight (g m−2) in 2018–2019 and 2019–2020. For the two correlations presented in C, weed dry weight data are derived from the second evaluation (Eval 2) of each separate year. RMSE, root mean-square error.

Figure 7

Table 5. Normalized difference vegetation index (NDVI) and Conyza bonariensis dry weight (g m−2) in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020 in the olive grove with fruit yield data (kg tree−1) also presented as recorded in 2018–2019 and 2019–2020.a

Figure 8

Figure 4. Linear regression analyses between (A) Sorghum halepense dry weight (g m−2) and normalized difference vegetation index (NDVI) in the first evaluation (Eval 1) of 2018–2019 and 2019–2020; (B) Sorghum halepense dry weight (g m−2) and NDVI in the second evaluation (Eval 2) of 2018–2019 and 2019–2020; and (C) fruit yield (kg tree−1) and Sorghum halepense dry weight (g m−2) in 2018–2019 and 2019–2020. For the two correlations presented in C, weed dry weight data are derived from the second evaluation (Eval 2) of each separate year. RMSE, root mean-square error.

Figure 9

Table 6. Normalized difference vegetation index (NDVI) and Sorghum halepense dry weight (g m−2) in the first (Eval 1) and second evaluations (Eval 2) of 2018–2019 and 2019–2020 in the vineyard with grape yield data (kg vine−1) are also presented as recorded in 2018–2019 and 2019–2020.a

Supplementary material: File

Kanatas et al. supplementary material

Tables S1-S4

Download Kanatas et al. supplementary material(File)
File 18.1 KB