Agronomía Colombiana ISSN: Universidad Nacional de Colombia Colombia

Transcription

1 Agonomía Colombiana ISSN: Univesidad Nacional de Colombia Colombia Conteas B., Ángela P.; Cayón S., Geado; Cochuelo R., Gemán s to estimate the bunch dy weight in Afican oil palm (Elaeis guineensis Jacq.), Ameican oil palm (Elaeis oleifea H.B.K. Cotes) and the intespecific hybid (E. oleifea x E. guineensis) Agonomía Colombiana, vol. 0, núm. 1, eneo-abil, 01, pp Univesidad Nacional de Colombia Bogotá, Colombia Available in: How to cite Complete issue Moe infomation about this aticle Jounal's homepage in edalyc.og Scientific Infomation System Netwok of Scientific Jounals fom Latin Ameica, the Caibbean, Spain and Potugal Non-pofit academic poject, developed unde the open access initiative

2 s to estimate the bunch dy weight in Afican oil palm (Elaeis guineensis Jacq.), Ameican oil palm (Elaeis oleifea H.B.K. Cotes) and the intespecific hybid (E. oleifea x E. guineensis) os paa estima el peso seco del acimo en palma aficana (Elaeis guineensis Jacq.), palma ameicana (Elaeis oleifea H.B.K. Cotes) e híbido inteespecífico (E. oleifea x E. guineensis) Ángela P. Conteas B. 1, Geado Cayón S. 1,, and Gemán Cochuelo R. 1 ABSTRACT Statistical models wee used to estimate the bunch dy weight though indiect nondestuctive methods in Afican oil palm (Elaeis guineensis) Ameican oil palm (Elaeis oleifea) and the intespecific hybid (E. oleifea x E. guineensis); and compaed with the fomula poposed by Coley. The studies wee conducted at Santa Bábaa and Chapaal-Cuenavaca on the Unipalma plantation, located in the easten palm egion of Colombia. Ten palms wee selected fo each goup and 0 bunches wee sampled fo six months. Polynomial and exponential statistical models wee postulated, with the best being linea without intecept. The esults confim and validate the usefulness of the model fomulated to estimate the bunch dy weight of Afican oil palm (E. guineensis), Ameican oil palm (E. oleifea) and the intespecific hybid (OxG); howeve, it poved moe convenient to use the models poposed in this study because they ae tailoed to the specific envionmental conditions of the easten palm egion of Colombia. Key wods: oil palm, plant physiology, gowth paametes, statistical models. RESUMEN Se deteminaon modelos estadísticos paa estima el peso seco del acimo po métodos indiectos no destuctivos en palma aficana (Elaeis guineensis), palma ameicana (Elaeis oleifea) y el híbido inteespecífico (E. oleifea x E. guineensis) compaándolos con la fómula popuesta po Coley. Los estudios se ealizaon en las Haciendas Santa Bábaa y Chapaal-Cuenavaca de Unipalma, ubicadas en la zona palmea oiental de Colombia. Se escogieon 10 palmas de cada mateial y se muesteaon 0 acimos duante seis meses. Se postulaon modelos estadísticos polinomiales y exponenciales, encontando que los mejoes fueon de tipo lineal sin intecepto. Los esultados confiman y validan la utilidad del modelo paa estima el peso seco del acimo en palma aficana (E. guineensis), palma ameicana (E. oleifea) y el híbido inteespecífico (OxG) de éstas; sin embago, es más conveniente utiliza los modelos popuestos en este tabajo poque están ajustados a las condiciones ambientales específicas de la zona oiental palmea colombiana. Palabas clave: palmas oleaginosas, fisiología de plantas, paámetos de cecimiento, modelos estadísticos. Intoduction To estimate the poductive potential of oil palm, gowth analysis is needed, detemining the size and activity of the ogans called souces (leaves) and demands (bunches). The fist analysis of oil palm gowth included the use of mathematical functions to descibe the dy weight and leaf aea and calculate the elative gowth ate (RGR), net assimilation ate (NAR) and cop gowth ate (CGR); which showed that the gowth is exponential and the lage the oil palm, the faste the gowth ate (Coley and Tinke, 009). Goudiaan and Monteith (1990) developed an expolinea equation fo gowth that descibes the tansition fom the exponential phase to the linea model phase, showing that this could accommodate oil palm gowth and yield data, the tansition occuing appoximately thee yeas afte tansplant to the field. The pimay data fo analysis of gowth can be collected fom individual plants, o deivatives of whole canopies, although the destuctive natue of the technique equies the use of homogeneous goups of plants o plots (Beadle, 1985). Destuctive methods cannot be used in peennial cops such as oil palm (Coley and Tinke, 009) because it has a single gowth meistem in the apical egion of the tunk whee the leaves and infloescences oiginate in egula succession (Coley and Gay, 198), but Hadon et al. (1969) and Coley et al. (1971) developed nondestuctive Received fo publication: 8 Mach, 011. Accepted fo publication: Novembe, Depatment of Agonomy, Faculty of Agonomy, Univesidad Nacional de Colombia. Bogota (Colombia). Coesponding autho. dgcayons@unal.edu.co Agonomía Colombiana 0(1), 46-51, 01

3 methods fo estimating leaf aea and annual poduction of dy matte. Although nondestuctive methods only conside the dy weight of leaves, stems and bunches, these stuctues constitute ove 96% of total dy matte (Coley et al., 1971). Bunch dy weight is a faily constant faction of fesh weight (5%), and as such, is outinely used in most studies on oil palm (Coley and Tinke, 009). Seveal nondestuctive methods have been developed to estimate leaf aea, bunch dy weight and dy matte poduction of oil palm (Hadon et al., 1969; Coley et al., 1971; Hadon et al., 197; Awal et al., 004). They have seveal advantages ove destuctive techniques: 1) measuements can be epeated seveal times on the same oil palms maintaining the minimum equied expeiment aea, wheeas with destuctive methods, expeiment size must be inceased fo each set of equied measuements, ) nondestuctive methods can be used fo selection in impovement pogams fo oil palms without destoying potentially poductive oil palms ) the technique consumes less time and is less laboious than destuctive analysis with plants as voluminous as oil palms (Coley, 1976; Chiaiello et al., 1989; Noman and Campbell, 1989). Gowth analysis also equies the use of equations to pedict the incease in aea and dy weight of the plant without the use of destuctive sampling, saving plant mateial by deceasing the aea equied fo the expeimental plots. Coley et al. (1971) developed the equation D = F to estimate bunch dy weight (D), whee F is bunch fesh weight (kg), which has been widely used in oil palm eseach. Howeve, fo moe expeimental pecision, equation constants should be checked and adjusted befoe applying them in othe pats of the wold (Mendham, 1971; Hatley, 1988; Henson, 199). The aim of the study was to develop, unde the conditions of the easten palm egion, statistical models to estimate the bunch dy weight fo Afican oil palm (E. guineensis), Ameican oil palm (E. oleifea) and the intespecific hybid (OxG), compaed with the fomula poposed by Coley et al. (1971). Mateials and methods The study was conducted at two locations in the municipalities of Cumaal (Meta) and Paatebueno (Cundinamaca), unde topical ain foest conditions (Bh-T) at 05 m. The fist site was at the La Cabaña fam, located in the Pesentando inspection zone, with an aveage tempeatue of 7 C, 80% elative humidity, annual ainfall of,500 mm and 1,500 to,000 h yea -1 of sunshine. The second site was at the Santa Babaa and Chapaal-Cuenavaca fams on the Unipalma plantation; Santa Babaa is located in the Veacuz inspection zone, with an aveage tempeatue of 6 C, elative humidity of 78%, annual ainfall of,77 mm and 1,50 h yea -1 of sunshine; Chapaal-Cuenavaca ae located in the municipality of Paatebueno, with an aveage tempeatue of 6 C, elative humidity of 78%, annual ainfall of,990 mm and 1,50 h yea -1 of sunshine. The mateials used wee: Afican oil palm (E. guineensis) Ténea Unileve Cameún sowed in 1989 on Santa Babaa, Nolí (E. oleifea) code 557, sowed in 1991 on Chapaal- Cuenavaca, and the hybid (OxG) code 5, sowed in 1991 on La Cabaña 10 palms wee chosen and maked pe mateial (with good mophological conditions, length and numbe of suitable leaves, thick leaflets and a nomal dak geen colo). Bunch sampling was conducted fo six months, with a fequency of 8 d fo the Ténea mateial, and 0 d fo the Nolí and hybid mateials, fo a total of 0 bunches pe mateial. The bunches wee havested at the physiological matuity stage, that is, when the fist ipe fuit detached, and wee taken to the laboatoy fo bunch analysis and pocessed to obtain thee bunch weights: 1) actual bunch dy weight (ABDW), ) calculated bunch dy weight () and ) estimated bunch dy weight (). Actual bunch dy weight (ABDW) To calculate ABDW, fesh weight was taken fo each bunch and each bunch was theshed sepaating the stem fom the infloescence elements without sepaating the fuit, then the fesh weight of the stem was taken and a subsample of the stem was taken and died in a convection oven at 100 C fo 4 h until a constant dy weight was eached. Samples of the infloescence elements, without sepaating the fuit, wee taken at 5.0 kg in Afican oil palm and.5 kg in Nolí and the hybid because the latte failed to achieve sufficiently ample infloescence elements; a subsample was taken fom the infloescence elements of 50 g of nomal fuit and anothe of 00 g of pathenocapic fuit and died in an oven until constant dy weight was eached. ABDW was indiectly estimated with the following equation: bunch fesh weight x bunch subsample dy weight ABDW = bunch subsample fesh weight Calculated bunch dy weight () Each havested bunch was weighed fo fesh weight and the dy weight was calculated () with the equation Conteas B., Cayón S., and Cochuelo R.: s to estimate the bunch dy weight in Afican oil palm (Elaeis guineensis Jacq.), Ameican oil... 47

4 D = F, fomulated by Coley et al. (1971), whee F equals bunch fesh weight in kg. Results and discussion Afican oil palm Statistical models poposed fo commecial palm mateial ae pesented in Tab.. You can see six models with intecept and six without intecept. In models with intecept, the R anged between and 0.91, consideed high. The R of cubic models showed no elevance, wheeas the R of quadatic models showed elevance and non-elevance, unlike that seen with the linea models which wee highly elevant in thei coefficients. The MSE anged between 0.40 and, the latte being associated with the linea exponential model. Fom a statistical standpoint, this would be a good model, while the statistical evidence was lowe fo the cubic model. In pevious wok (Conteas et al., 1999), it has been epoted that statistical models with intecept have elatively low eliability and ae not adequate to estimate the dy weight of the Afican oil palm, contay to the model poposed by Coley et al. (1971) which has intecept. Estimated bunch dy weight () With the data fo actual (ABDW) and calculated bunch dy weight (), polynomial models and exponentialpolynomial with and without intecept (Tab. 1) wee postulated to detemine the estimated dy weight (). Table 1. Postulated statistical models to estimate bunch dy weight ( y i ) fo Afican oil palm (E. guineensis), Ameican oil palm (E. oleifea) and the hybid OxG. s with intecept s without intecept + b xi + b xi + b xi + b xi + b x + b xi i b0 +b1xi +b xi +b xi b0 +b1xi +b xi +b xi b0 + b1 xi + b xi b1xi +b xi b1xi b0 +b1xi The poposed intecept models fo palm had an R of, consideed outstanding. The SME was low () fo the cubic and quadatic exponential models and highest (0.41) fo the simple egession model. The egession coefficients estimated fo the cubic models wee not elevant, in the quadatic models some coefficients wee highly elevant and othes wee not, wheeas in the simple linea models, all wee highly elevant. b0 = intecept; b1, b, b = estimated paametes; x = bunch fesh weight. To choose the best model, the study took into account fo each: the ability to be applied in the field, its goodness of fit expessed by coefficient of detemination (R ), mean squae eo (MSE) and the impotance of each vaiable in the model. Coelations wee estimated between, and ABDW to see association between the model poposed in this pape and the model of Coley et al. (1971). Simple linea, quadatic and simple exponential models would be adequate to estimate bunch dy weight but, since Table. Estimate statistical models fo bunch dy weight ( y i ) of Afican oil palm (E. guineensis). R SME b0 b1 b b + b x + b x NS.5 NS NS NS + b xi NS 0.77 * - NS * 0.65 ** i i b0 +b1xi +b xi +b xi NS 0.51 NS -0.0 NS b0 + b1 xi + b xi NS 0.5 ** - ** b0 +b1 xi ** 0.5 ** + b xi + b xi + b xi b0 +b1xi +b xi +b xi NS 0.0 NS ** NS ** 0. ** NS b1xi +b xi 0. ** - ** b1xi ** NS NS NS b0 = intecept; b1, b, b = estimated paametes; x = bunch fesh weight; MSE = mean squae eo ** Significance 1%; * Significance 5%; NS, not significant. 48 Agon. Colomb. 0(1) 01

5 the idea is to use the simplest and most economical and pactical one, we chose the simple egession model = 0.58 x (R = and SME = 0.41). To detemine if this model would be a good estimate fo bunch dy weight, we estimated values fo BDW and, and coelated each with the ABDW values. The esults ae pesented in Tab. whee one can see a positive coelation, highly significant and equal fo both models. without intecept also showed high R ( 0.89) and SME vaied between 0.8 and ; the R fo the cubic models showed elevance and non-elevance, and in the quadatic models, some R showed elevance and othes not. Fom a statistical standpoint, the best models ae simple linea and simple linea exponential, of which, because of its simplicity and ease of implementation in the field, the selected model is = x (R = 0.98 and CME = 0.04). Table. Coelation coefficients of simple linea () fo the actual bunch Coelations between estimated dy weight () and calculated dy weight () with the actual dy weight (ABDW) ae expessed in Tab. 5, showing that the coelation coefficients () wee highly significant (P ) but highe in the ABDW- association (0.91). This indicates that although both models ae eliable fo estimating the dy weight of Ameican oil palm bunches, the simple egession model selected in this study ( = x) is moe suitable fo being adapted to the Colombian palm egion. dy weight (ABDW) with calculated bunch dy weight () and estimated bunch dy weight () in Afican oil palm (E. guineensis) ** 0.95 ** Based on this, it is moe convenient to use the estimated model in this pape, as adjusted especially fo Colombian conditions with efeence to the taditional model oiginally developed by Coley et al. (1971). Ameican oil palm Statistical models fo estimating the dy weight of the Ameican oil palm Nolí bunch ae pesented in Tab. 4. The R fo models with intecept vaied between 0.8 and 0.89 and the SME between 0.0 and. The R fo the cubic and quadatic models showed no elevance, while fo the simple linea models some did and othes did not, theefoe the model with the highest appaent pediction is the simple linea ( ). The models fo the Nolí oil palm Table 5. Coelation coefficients () of simple linea fo actual bunch dy weight (ABDW) with calculated bunch dy weight () and estimated bunch dy weight () in Ameican oil palm (E. oleifea) ** 0.91 ** Hybid (OxG) The estimations of the 1 models poposed fo the hybid mateial (Tab. 6) show that the R of models with intecept Table 4. Estimate statistical models fo bunch dy weight of ( y i ) Ameican oil palm (E. oleifea). + b x + b x i i R SME b0 b1 b b NS.51 NS NS 0.05 NS 0.07 NS + b x NS -0.7 NS * 0.70 ** NS 0.98 NS -0.1 NS NS 0.15 NS NS NS 0.17 ** i b0 +b1xi +b xi +b xi b0 + b1 xi + b xi b0 +b1 xi + b x + b x ** NS + b xi ** 0.0 ** i i ** b0 +b1xi +b xi +b xi 0.6 ** -0.0 NS b1xi +b xi 0.1 ** NS 0.19 ** b1 xi NS NS NS b0 = intecept; b1, b, b = estimated paametes; x = bunch fesh weight; MSE = mean squae eo. ** Significance 1%; * Significance 5%; NS, not significant. Conteas B., Cayón S., and Cochuelo R.: s to estimate the bunch dy weight in Afican oil palm (Elaeis guineensis Jacq.), Ameican oil... 49

6 Table 6. Statistical models to estimate bunch dy weight of the ( y i) hybid (OxG). R SME b0 b1 b b + b x + b x NS.80 NS -0. NS 0 NS + b xi * 1.4 ** * NS 0.56 ** i i b0 +b1xi +b xi +b xi * 1.9 ** -0.1 * * 0.4 ** -0.0 ** ** 0.1 ** + b x + b x NS 0.07 NS + b x ** 0.00 * ** 0.18 ** NS 0.5 ** - ** ** b0 +b1xi +b xi b0 +b1xi i i i b0 +b1xi +b xi +b xi b1xi +b xi b1 xi 0.00 NS NS NS MSE = mean squae eo; ** Significance 1%; * Significance 5%; NS, not significant. was geneally low (between 0.61 and ), some of the R fo the cubic models showed high elevance (**) and othes did not. The simple linea and quadatic models showed significant (*) and highly significant (**) coefficients, of which quadatic, quadatic exponential and simple linea exponential had the bette statistical behavios. Fom the pactical viewpoint fo use in the field, the exponential simple linea model would be best. It is infeed that the coelation coefficient of the two models egading the obseved values is the same and highly significant, theefoe, the poposed model in this pape is simila to that poposed by Coley et al. (1971) but moe suitable as it was adjusted to Colombian topical conditions. The degee of association between and the pevious model ae seen in Tab. 7, showing that the coelation coefficients () wee simila and highly significant (P ), so we conclude that both models ae eliable fo estimating bunch dy weight in the hybid OxG. Geneal model Finally, a geneal model was estimated fo the thee studied mateials, with the same statistical tests descibed above fo each mateial. Since models without intecept wee the best fo the thee mateials, a model that uses bunch fesh weight (x) was postulated, finding that the best model is = x which povides an R of and SME of 0.. To detemine the degee of association, coelation analysis was pefomed between the estimated values of bunch weight () and calculated values of bunch weight () with the actual values (ABDW), pesented in Tab. 8. The geneal model has a geate association of the vaiables than the models of each of the mateials sepaately and also a geate association than the model poposed by Coley et al. (1971); confiming that this is a eliable model to estimate the bunch dy weight of Afican oil palm, Ameican oil palm and the hybid (OxG). Table 7. Coelation coefficients () of simple linea fo actual bunch dy Table 8. Coelation coefficients () of simple linea fo actual bunch dy The models fo the hybid without intecept exhibited high R ( 0.988) and SME anged between 0.9 and 0, the R fo cubic models was highly elevant and not elevant, wheeas in the quadatic models R was elevant and highly elevant. Fom a statistical standpoint, the best models ae simple linea and simple linea exponential, of which, and fom the pactical point of view of field application due to simplicity and ease of calculation and economy, the selected model is = x (R = and MSE = 0.1). weight (ABDW) with calculated bunch dy weight () and estimated bunch dy weight () in the hybid (OxG) ** ** 50 weight (ABDW) with calculated bunch dy weight () and estimated bunch dy weight () in the geneal model fo the thee studied mateials. Geneal model 0.95 ** 0.96 ** Agon. Colomb. 0(1) 01

7 The esults of this study confim the impotance and validation of the model fomulated by Coley et al. (1971) to estimate the bunch dy weight in Afican oil palm (E. guineensis), Ameican oil palm (E. oleifea) and the intespecific hybid (OxG). Howeve, the models poposed in this wok pesent a viable option because they comply with statistical paametes and ae adjusted to Colombian topical conditions, theefoe poviding data that ae moe consistent with the Colombian envionmental eality. Similaly, fo a good estimate of bunch dy weight, it would be bette to use the geneal model due to the highe degee of association with the actual values of bunch dy weight. Liteatue cited Awal, M.A., I. Wan, J. Endan, and M. Haniff Detemination of specific leaf aea and leaf aea-leaf mass elationship in an oil palm plantation. Asian J. Plant Sci. (), Beadle, C.L Plant gowth analysis. pp In: Coombs, J., D.O. Hall, S.P. Long, and M.O. Seulock (eds.). Techniques in biopoductivity and photosynthesis. Pegamon Pess, Oxfod, UK. Conteas, A.P., G. Cochuelo, O. Matínez, and G. Cayón Estimación del áea y del peso seco folia en Elaeis guineensis, Elaeis oleifea y el híbido inteespecífico E. guineensis x E. oleifea. Agon. Colomb. 16(1-), 4-9. Coley, R.H.V., J.J. Hadon, and G.Y. Tan Analysis of gowth of the oil palm (Elaeis guineensis Jacq.) I. Estimation of gowth paametes and application in beeding. Euphytica 0, Coley, R.H.V Photosynthesis and poductivity. pp In: Coley, R.H.V., J.J. Hadon, and B.J. Wood (eds.). Oil palm eseach. Elsevie, Amstedam, The Nethelands. Coley, R.H.V. and B.S. Gay Gowth and mophology. pp In: Coley, R.V.H., J.J. Hadon, B.J. Wood (eds.). Oil palm eseaches. Elseive, Amstedam, The Nethelands. Coley, R.H.V. and P.B. Tinke La palma de aceite. 4 th ed. Fedepalma, Bogota. Chiaiello, N.R., H.A. Mooney, and K. Williams Gowth, cabon allocation and cost of plant tissues. pp In: Peacy, R.W., J.R. Ehleinge, H.A. Mooney, and P. Runde (eds.). Plant physiological ecology: field methods and instumentation. Chapman and Hall, London. Goudiaan, J. and J.L. Monteith A mathematic function fo gowth based on light inteception and leaf aea expansion. Ann. Bot. 66, Hadon, J.J., C.N. Williams, and I. Watson Leaf aea and yield in the oil palm in Malaya. Expl. Agic. 5, 5-. Hadon, J.J., R.H.V. Coley, and S.C. Ool Analysis of gowth in oil palm. II. Estimation of genetic vaiances of gowth paametes and yield of fuit bunches. Euphytica 1, Hatley, C.W.S The oil palm. Longmans Geen. New Yok, NY. Henson, I.E Assessing fond dy matte poduction and leaf aea development in young palms. pp In: Poceedings 1991 PORIM Institute Palm Oil Reseach Confeence Agicultue, Palm Oil Reseach Institute of Malaysia. Kuala Lumpu. Mendham, N.J Note on leaf aea measuement in oil palms. Papua New Guinea Agic. J., 0-1. Noman, J.M. and S.G. Campbell Canopy stuctue. pp In: Peacy, R.W., J.R. Ehleinge, H.A. Mooney,and P.W. Rundel (eds.). Plant physiological ecology: field methods and instumentation. Chapman and Hall, London. Conteas B., Cayón S., and Cochuelo R.: s to estimate the bunch dy weight in Afican oil palm (Elaeis guineensis Jacq.), Ameican oil... 51