Volume application rate adapted to the canopy size in greenhouse tomato crops

Transcription

1 390 Volume Scienti ppliction Agricol rte in greenhouse crops Volume ppliction rte dpted to the cnopy size in greenhouse tomto crops Julián Sánchez-Hermosill *, Frncisco Páez 2, Víctor J. Rincón, José Pérez-Alonso University of Almerí Dept. of Agriculturl Engineering, Ctr. Scrmento s/n 0420 Almerí Spin. 2 Andlusin Institute for Reserch nd Trining in Agriculture (IFAPA-L Mojoner), Andlusin Government, Cmino de Sn Nicolás, L Mojoner, Almerí Spin. *Corresponding uthor <jusnche@ul.es> Edited by: Dniel Scherer de Mour Received November 28, 202 Accepted June 28, 203 ABSTRACT: The ppliction rte of plnt-protection products is indicted s concentrtion or mount of product per re. Greenhouse crops grow swiftly, nd n ppliction rte bsed on fi xed mount of product per hectre cn result either in lrge losses nd overdoses when the plnts re smll or to be insuffi cient when the plnts re fully developed. To solve these problems, the ppliction rtes of plnt-protection products need to be dpted to the plnt mss present in the greenhouse when the spry is pplied. Two models were developed to estimte the lef re bsed on esily mesured geometric dt of the vegettion in greenhouse tomto crop. The model bsed on the PRV (Plnt Row Volume) hd tht best results. The clcultion of the volume ppliction rte from the PRV hs resulted in reduction of more thn 30 % of the quntity of plnt protection product spryed, without decresing yield. The PRV of greenhouse tomto (Lycopersicon esculentum Mill.) is n esily mesured prmeter tht enbles the estimtion of the lef re index nd the use of ppliction strtegies dpted to the chnges in the plnt cnopy, sving mjor mounts of plnt protection product used, compred to the conventionl system. Introduction Quntifying vegettion nd its subsequent reltion to the spry-ppliction rte re the bsis for techniques of crop dpted sprying (CAS) used by reserchers (Sutton nd Unrth, 988; Giles et l., 989; Rüegg et l., 999; Gil et l., 2007; Pergher nd Petris, 2008) for pplying plnt protection products in different crops. The gol is to chieve more rtionl use of sprys ccording to the rel needs of the crop while seeking to protect the environment nd humn helth. Different methods hve been developed to solve the problems of djusting the ppliction rte to the level of crop development. One method is the Tree-Row Volume (TRV) (Byers et l., 97), whereby the ppliction rte is determined by compring the crop cnopy volume per surfce-re unit with the rte pplied to stndrd crop. This method specifies the ppliction rtes in more methodicl wy nd hs rendered good results in orchrds (Sutton nd Unrth, 988; Rüegg et l., 999; Gil et l., 2007; Siegfried et l., 2007). A simpler system ws proposed by Furness et l. (998), clled Unit Cnopy Row (UCR), in which the ppliction rte ws determined by ech unit of 00 m 3. Brni et l. (2008) used the UCR in vineyrds to determine the ppliction rtes nd obtined good results. To mtch ppliction rtes nd spry volume, Weisser nd Koch (2002) proposed the treted re concept, defined s the re tht is spryed nd oriented between nozzles nd trgets. Therefore, for high crops the ppliction rte is to be expressed in terms of the lef wll re (LWA) (Koch, 2007). Another wy of estblishing the ppliction rte s quntity of spry product per unit of lef surfce re is from the mesurement of the (Lef Are Index). This prmeter is used to determine the ppliction rtes (Dmmer et l., 2008; Pergher nd Petris, 2008; Gil nd Escolà, 2009). However, the use of the to djust the ppliction rte is not very prcticl becuse its vlue is difficult to determine for the frmer under field conditions. This study describes model to estimte the lef re bsed on esily mesured geometric dt of the vegettion in greenhouse tomto crop, such s the cnopy height or the Plnt Row Volume (PRV), nd its use to clculte the volume ppliction rte of plnt protection product from the verge folir deposit. Mterils nd Methods Mesurement of geometric prmeters of the cnopy This study ws conducted with the geometric dt from greenhouse tomto plnts (Lycopersicon esculentum Mill.), mesured t three loctions from vrious commercil cultivrs nd in different types of greenhouses, from 2007 to 200. Tble lists the loctions of the greenhouses, cultivrs, nd plnting pttern of the crop. In ll cses, the crop ws plnted in twin-rows (two plnts per position) on perlite substrte. In the greenhouses, the geometric prmeters of height nd width of the cnopy s well s the folir surfce re were recorded using the methodology detiled below. These prmeters were mesured periodiclly (every 5-20 dys) ccording to the crop cycle, development, nd cultivtion tsks. Determintion of the Plnt Row Volume The TRV is defined s the volume of vegettion per unit of surfce re cultivted. In the present study, this is clled Plnt Row Volume (PRV), s this refers to nnul plnts nd not to the trees, with the implictions tht this involves. The geometric vribles needed to determine this index re height nd width of the cnopy,

2 39 Tble Geogrphicl coordintes, seson, nd cultivr t the loctions. Loctions 2 3 Geogrphicl coordintes WGS84 Lt º Long. -2.4º Lt º Long -2.72º Lt º Long -2.7º Nº of greenhouses 8,000 2,780 2,920 Are (m 2 ) Sesons Cultivr Plnting pttern (m) Nº of smplings Autumn-Winter Autumn-Winter Beel Divine Divine Divine Rcimo Rcimo Delici nd the distnce seprting the plnt rows. Becuse the width of the cnopy vries with plnt height, PRV ws clculted considering the volume of the vegettion corresponding to ech of the three heights defined to mesure the width (Figure ). Thus, the PRV ws clculted ccording to Eq H B B M M A A PRV 3D () where PRV is the Plnt-Row Volume (in m 3 h ), H is the cnopy height (in m), B nd B 2 re the width of the plnts t /6 the height (in m), M nd M 2 width of the plnts t /2 the height (in m), A nd A 2 width of the plnts t 5/6 the height (in m) nd D is the distnce between the rows (in m). The height of the cnopy ws mesured perpendiculr to the soil. In the first stges totl height ws considered. After the different clering of leves closest to the soil during the crop cycle, usul in cultivtion tsks, this mesurement refers only to the zone with the folir mss. The cnopy height ws mesured in four rndomly ssigned pirs of plnts representtive of the rest of the greenhouse. As the plnts were pired, the width ws the men length prllel to the soil, from the lterl exterior of the right-hnd plnt to the lterl exterior of the left-hnd plnt, discounting the spce between them, but without overestimting this geometric prmeter (Figure ). Determintion of the lef re index The surfce re of the leflets comprising the leves ws mesured, using destructive method, which consisted of cutting off ll the leves of the individul, mesuring their surfce re with the use of plnimeter (WinDis, Delt-T Devices Ltd. Cmbridge). Thus, this method involves the destruction of the studied smple. This prmeter ws mesured on six plnts ssigned Figure Mesurement of width; B low width (B +B 2 ) (/6 height), M middle width (M +M 2 ) (/2 height), nd A high width (A +A 2 ) (5/6 height). t rndom from mong the four pirs of plnts used to mesure the height nd width. Becuse of the destruction involved over the crop cycle, this prmeter ws not mesured on ll the plnts. Development of the model to estimte the Models were developed to relte the cnopy height, PRV, nd, with the im of finding correltion between them, nd to be ble to estimte from the cnopy height or from the PRV. For the development of these models, the men vlues of H, PRV, nd were used for ech greenhouse nd dy smpled, nd were mesured ccording to the methodology described bove. Correltion nd regression nlyses were clculted with the softwre SPSS v5.0 (SPSS Inc., n IBM Compny, Chicgo, IL, USA). The ccurcy of the developed models ws determined by mesuring the height of the cnopy nd quntifying the nd PRV, ccording to the methodology

3 392 described bove, in tomto crop t times over the crop cycle (Tble 2). The tomto cultivr ws Divine, cultivted in greenhouse in El Ejido, Almerí, Spin (36º 48 N, 2º 43 W nd 5 m elevtion), during the utumn-winter seson For ech of the developmentl stges of the crop in which the dt were tken, the ws estimted ccording to models estblished previously, nd the prediction residul sum of squres (PRESS) ws clculted using Eq. 2. N j ˆ 2 j j PRESS (2) where N is the totl number of dt (); j the known for the j-th dt nd LÂI j the estimted of j. The PRESS vlues provide mesure of how well the models predict the. The prediction error of ws clculted s the reltive stndrd error (RSE) of the prediction (Eq. 3). RSE(%) N ˆ 2 2 j j j 00 N 2 j j Estimtion of the volume ppliction rte Pergher nd Petris (2008) proposed Eq. 4 to estimte the product ppliction rte (Q, in g h ) bsed on, the verge folir deposit (d, in µg cm 2 lef re) nd the spry frction deposited on the cnopy (). Q = d. / (4) The use of Eq. 4 to djust the product ppliction rte is not very prcticl becuse the vlues of, d nd re difficult to determine by the frmer under field conditions. In the present study, the d vlue corresponds to the men deposition chieved by frmers in their routine prctice, consisting of spry gun pplictions t high pressures nd volumes. Tking into ccount the vlue Tble 2 Dt used to vlidte the models. Assy dt PRV b m 3 h 36 4, , , , , , , , , , , dt = dys fter trnsplnting; b Plnt Row Volume. (3) of verge folir deposition in µl cm 2 lef re (d v ) the volume ppliction rte (V, in L h ) cn be determined by Eq. 5. V = d V. /ε (5) The vlue of depends on spryer type, crop structure, nd wether conditions. In severl studies the vlues of d v nd hve been clculted in greenhouse tomto crops (Cerruto et l., 2009, b;, 20, 202), s indicted in Tble 3. In considertion of these vlues, our proposl is to use vlue of.5 µl cm 2 d v nd, for the spry frction deposited on the cnopy () 0.50 for spry-gun pplictions or 0.75 for spry boom pplictions. The vlue of in Eq. 5 cn be clculted with the model for estimting s function of geometric prmeters of the crop, developed under the previous point. Evlution of the method to determine the volume ppliction rte To evlute the method to determine the volume ppliction rte, two greenhouses of 960 m 2 (24 m 40 m) were selected in L Mojoner, Almerí, Spin, (36º48 N, 2º4 W nd 42 m elevtion) for field experiment during two sesons (spring-summer 20 nd utumnwinter ). The crop used ws tomto grown in perlite nd plnted in twin-row system. Ech twin-row ws 2 m prt nd hd 50 pirs of plnts spced on 0.4 m (2.5 plnts m 2 ). Every greenhouse ws divided into two blocks following rndomized block design with four blocks nd two tretments (two ppliction strtegies). The tretments were the conventionl (CON) nd estimted (EST) ppliction rte. The CON tretment involved sprying throughout the seson ccording to the frmer s criteri (ppliction rte, pressure, trvel speed, etc.). The EST tretment ment sprying ccording to the volume ppliction rte clculted using the Eq. 5. During the ppliction of the sprys, the concentrtion Tble 3 Vlues of verge folir volume deposit (d v ) with spry gun nd the spry frction deposited on the cnopy () for greenhouse tomto crop in different studies. d v (µl cm 2 ) b Spry gun Verticl boom Sánchez-Hermosill et l. (20) c Sánchez-Hermosill et l. (202) Beulke et l. (20) Beulke et l. (20) Cerruto et l. (2009) Cerruto et l. (2009) Cerruto et l. (2009b) Lef re index; b Men deposition by frmers in their routine prctice; c Vlue clculted from the dt published in, 20.

4 393 of the ctive ingredient (.i.) in the tnk ws the sme for the CON s for the EST tretment nd ws determined following the recommendtions indicted on the product lbel, using the volume ppliction rte selected by the frmer (V CON ) s reference. Ech tretment ws pplied in n experimentl plot mesuring 0 m 24 m, 4-5 crops lines nd seprted by plstic sheets (Figure 2). The spry equipment used in this study ws mnully pulled trolley with two verticl spry booms (Crretills Amte S.L., Almerí, Spin), which ws connected through 25-m long hose (dimeter 7 mm) to wheelbrrow with tnk of 00 L nd membrne pump (M-30, Imovilli Pompe s.r.l., Reggio Emili, Itly) tht provided mximum pressure of 3000 kp nd mximum flow of 33 L min. Ech boom hd four stndrd flt fn nozzles (XR 0 03) inserts spced 0.50 m prt nd the nozzles were fitted t n offset ngle of 7º nd oriented 5º upwrds. The verticl distnce between the lowest nozzle nd the ground ws 0.27 m. Tble 4 shows the working prmeter for the pplictions. Yield evlution To quntify yield, two groups of 2 plnts were selected from ech experimentl plot (Figure 2). One ws situted in the northern hlf of the greenhouse nd the other in the southern hlf in order to void the effects of orienttion. In ech hrvest, the fruits from ech group were weighed using scle (HW-00KV WP, A&D Compny Ldt., Jpn). Results nd Discussion Development of the model to estimte the The correltions between nd cnopy height were nlysed (Figure 3), s well s between nd PRV (Figure 4), the results being significnt in both cses t confidence level of 95 %, with determintion coefficients (R 2 ) of 0.9 nd 0.94, respectively. Esily mesured prmeters in the plnt mss of the tomto crop permit the estimtion of n index s complex s. This is becuse the norml development of the tomto plnt hs regulr node distribution nd lef development. However, the hs closer reltionship with PRV thn with cnopy height, since the PRV vlue, prt from tking into ccount cnopy height lso tkes into ccount the development of the leves in n indirect wy from the width of the vegettion. The model relting with PRV corresponds to power eqution, similr to Figure 2 Greenhouse plnt with the comprtmentliztion in experimentl blocks nd plots y = 0.99x.49 R² = Cnopy height (m) Figure 3 Cnopy height- Lef re index () regression curve. Tble 4 Working prmeters for the pplictions. Seson Applictions PRV Spring-summer 20 Conventionl ppliction Estimted ppliction b c d Pressure Speed V CON V EST Pressure Speed V F m 3 h kp m s L h kp m s L h Autumn-winter / Plnt Row Volume V CON ; b ppliction rte clculted with the frmer s criteri; c V EST, ppliction rte clculted with Eq. 6; d V F, ppliction rte clculted under fi eld conditions.

5 y = x.25 R 2 = PRV (m 3 h ) Figure 4 Plnt row volume (PRV)- Lef re index () regression curve. the model reported by Siegfried et l. (2007) to relte nd VRV (Vine Row Volume) in grpevine. To determine which of the models offered the best estimtion of the, clcultion ws mde of the Reltive Stndrd Error (RSE) from the dt ( dt in totl) of series of plnts smpled (Tble 2), which were not included in the models developed. For ech model used to estimte, Tble 5 shows the PRESS, the RSE, nd the eqution. These models to estimte were vlidted only for greenhouse tomto crops with norml vegettive development (internodes of c. 5-6 cm, depending on the growth stge), cultivted with dequte irrigtion nd nutrients s well s proper mngement (lef thinning, brnch thinning, nd tip pinching). The nlysis of the vlues estimted for ech dtum used for vlidtion (Tble 6) indicted tht the model bsed on the cnopy height hd greter reltive errors between tests 3 nd 5, with vlues higher thn 0 %. This ws becuse just t 56 dt (dys fter trnsplnting) the plnts were lef thinned. The thinning, i.e. clipping off the lower lefs of the plnts, decresed the lef surfce re, reducing the by 4 % from ssy 3 to 4. However, lef thinning did not lower the cnopy height, which between ssy 3 nd 4 incresed 8 %, resulting in errors of more thn the 0 % indicted bove. On the other hnd, the cnopy height ws shortened when tip pinched, which consisted of removing the tip of the min stem, limiting plnt growth. Tip pinching ws performed t 88 dt, lthough this opertion did not led to mjor errors in estimting the with the cnopy height, (ll were under 0 %) becuse the loss in lef re ws minor, since new leves were locted t the tip of the stem, representing minor surfce-re development. For the model used to estimte bsed on the PRV, ll the errors were lower thn 9 % (Tble 6), even fter lef thinning nd tip pinching, which prompted mjor chnges in plnt rchitecture. Nevertheless, mjor errors (c. 8 %) lso rose just fter lef thinning. The two models enbled good estimtion of the with reltive stndrd errors lower thn 9 %. Nevertheless, the model tht offered the best estimtion of the ws the one bsed on the PRV with RSE of 5.44 Tble 5 Vlues of the prediction residul sum of squres (PRESS), the reltive stndrd error (RSE) nd eqution for ech model. Models to estimte PRESS RSE Model eqution R (%) 2 Bsed on the cnopy height = 0.99 (H) Bsed on PRV = (PRV) Tble 6 Reltive errors in estimting the lef re index () with ech of the models. Assy dt PRV Cnopy Height %. Tking into ccount the PRV- reltionship, Eq. 5 enbled the volume ppliction rte to be determined s function of PRV (Eq. 6.): V = d V. PRV.25 / (6) Evlution of the method to determine the volume rte for sprying An evlution ws mde of the plnt protection chieved over two sesons (spring-summer 20, utumn-winter 20-2). Tble 7 lists the products pplied in ech seson nd the dose pplied for the CON nd EST strtegies, clculted on the bsis of the volume shown in Tble 4 for ech strtegy. The im of these ssys ws to compre two strtegies to clculte the ppliction volumes: the CON rte, clculted with the criteri of the frmer (V CON ), nd the EST rte, clculted from the PRV using Eq. 6 (V EST ). For this, the yield corresponding to ech tretment ws compred for possible differences. Tble 7 shows the percentge of svings of the chemicl when Eq. 6 ws used to estimte the volume ppliction rte. The svings in both sesons consistently exceeded 30 %, with men vlues of % nd 35.5 % in the first nd second seson, respectively. In terms of yield, Tble 8 shows the men vlues clculted in the res where ech control strtegy ws evluted. The two strtegies gve similr yield in both sesons. There were no differences (p < 0.05) with Reltive Error (%) b RE H c RE PRV d m 3 h m Plnt Row Volume; b Reltive error (%) = -LÂI / 00; c Reltive error with the model height of plnt mss- ; d Reltive error with the model PRV-.

6 395 Tble 7 List of plnt protection product pplied in ech seson. Applied dose for the conventionl nd estimted strtegies nd svings. Seson Applictions Active ingredients Conventionl rte Estimted rte Svings (%) Spring-summer 20 Autumn-winter 20/2 Cyprodinil 37.5 % + Fludioxonil 25 % g h g h 38 Dimethomorph 7.5 % + Mncozeb 66.7 % g h g h 38 2 Indoxcrb 30 % g h g h 34 3 Chlorntrniliprole 35 % g h g h 32 4 Chlorntrniliprole 35 % 4.26 g h g h 33 Spiromesifen 24 % ml h ml h 33 Cyzofmid 40 % ml h ml h 33 2 Cyzofmid 40 % ml h ml h Fenhexmid 50 % g h g h 36 Chlorothlonil 50 % ml h ml h 36 Mefenoxm 2.5 % + Copper oxychloride 40 % g h g h 37 Spiromesifen 24 %.30 ml h ml h 37 Tble 8 Commercil yield in ech seson. Appliction strtegy spring-summer 20 utumn-winter 20/ kg m Conventionl ppliction rte Estimted ppliction rte respect to the min fctors (control strtegy, block nd greenhouse) or between their interctions. The similrity of the yield in both tretments, V CON nd V EST, mens tht the crop ws suitbly protected despite the decrese in volume pplied. If the volume ppliction rtes in V EST hd not been enough to control the pest ttck, there would hve been reduction in the yield. Conclusions The model tht offered the best results ws bsed on the PRV. The errors in estimting the were reltively low even fter lef thinning nd tip pinching, which lter the plnt rchitecture. The estimtion of from the PRV defined strtegy by which volume ppliction rte could be mtched to the chrcteristics of the crop. This strtegy presented stisfctory results, providing reduction of more thn 30 % with respect to the conventionl volume ppliction rte. Acknowledgements This study ws supported by grnt P07-AGR from CICE-Junt de Andlucí (Spin), co-finnced with FEDER funds of the Europen Union. References Brni, A.; Frnchi, A.; Bugini, R.; Montermini, A Efficcy of unit cnopy row sprying system for control of Europen vine moth (Lobesi botrn) in vineyrds. Avilble t: cigrjournl.org/index.php/ejounrl/rticle/viewfile/244/02 [Accessed Nov. 28, 202]. Beulke, S.; vn Beinum, W.; Glss, R.; vn Os, E.; Holtermn, H.; Spouns, A.; Voogt, W.; vn de Znde, J.; de Zwrt, F.; Grrtt, J. 20. Estimtion/clcultion of emissions of plnt protection products from protected crops (greenhouses nd cultivtions grown under cover) to support the development of risk ssessment methodology under Regultion (EC) n 07/2009 nd Council Directive 9/44/EEC. Europen Food Sfety Authority (EFSA). Avilble t: eu/en/supporting/pub/5e.htm [Accessed Nov 28, 202]. Byers, R.E.; Hickey, K.D.; Hill, C.H. 97. Bse gllonge per cre. Virgini Fruit 60: Cerruto, E.; Emm, G.; Mnetto, G Spry pplictions to tomto plnts in greenhouses. I. Effect of wlking direction. Journl of Agriculturl Engineering 3: Cerruto, E.; Emm, G.; Mnetto, G. 2009b. Spry pplictions to tomto plnts in greenhouses. II. Effect of spry lnce type. Journl of Agriculturl Engineering 3: Dmmer, K.; Wollny, J.; Giebel, A Estimtion of the Lef Are Index in cerel crops for vrible rte fungicide sprying. Europen Journl of Agronomy 28: Furness, G.O.; Mgrey, P.A.; Miller, P.M.; Drew, H.J Fruit tree nd vine spryer clibrtion bsed on cnopy size nd length of row: Unit cnopy row method. Crop Protection 7: Gil, E.; Escolà, A Design of decision support method to determine volume rte for vineyrd sprying. Applied Engineering in Agriculture 25: Gil, E.; Escolà, A.; Rosell, J.R.; Plns, S.; Vl, L Vrible rte ppliction of plnt protection products in vineyrd using ultrsonic sensors. Crop Protection 26: Giles, D.K.; Delwiche, M.J.; Dodd, R.B Spryer control by sensing orchrd crop chrcteristics: Orchrd rchitecture nd spry liquid svings. Journl of Agriculturl Engineering Reserch 43: Koch, H How to chieve conformity with the dose expression nd spryer function in high crops. Pflnzenschutz- Nchrichten Byer 60/200 : Pergher, G.; Petris, R Pesticide dose djustment in vineyrd sprying nd potentil for dose reduction. Avilble t: viewfile/254/ [Accessed Nov 28, 202].

7 396 Rüegg, J.; Viret, O.; Risigl, U Adpttion of spry dosge in stone-fruit orchrds on the bsis of the tree row volume. Bulletin OEPP/EPPO 29: Sánchez-Hermosill, J.; Rincón, V.J.; Páez, F.; Agüer, F.; Crvjl, F. 20. Field evlution of self-propelled spryer nd effects of the ppliction rte on spry deposition nd losses to the ground in greenhouse tomto crops. Pest Mngement Science 67: Sánchez-Hermosill, J.; Rincón, V.J.; Páez, F.; Fernández, M Comprtive spry deposits by mnully pulled trolley spryer nd spry gun in greenhouse tomto crops. Crop Protection 3: Siegfried, W.; Viret, O.; Huber, B.; Wohlhuser, R Dosge of plnt protection products dpted to lef re index in viticulture. Crop Protection 26: Sutton, T.B.; Unrth, C.R Evlution of the tree-rowvolume model for full-seson pesticide ppliction on pples. Plnt Disese 72: Weisser, P.Y.; Koch, H Expression of dose rte with respect to orchrd spryer function. Aspects of Applied Biology 66: