Services on Demand
- Cited by SciELO
- Access statistics
- Cited by Google
- Similars in SciELO
- Similars in Google
Print version ISSN 0034-737X
Rev. Ceres vol.59 no.4 Viçosa July/Aug. 2012
Nitrogen fertilizer (15N) leaching in a central pivot fertigated coffee crop*
Lixiviação de nitrogênio (15N) do fertilizante em uma cultura de café fertirrigada
Rafael Pivotto BortolottoI; Isabeli Pereira BrunoI; Klaus ReichardtII; Luís Carlos TimmIII; Telmo Jorge Carneiro AmadoIV; Ademir de Oliveira FerreiraV
IAgronomist Engineer, Doctor Science. Escola Superior de Agricultura "Luiz de Queiroz" (ESALQ), Avenida Pádua Dias, 11, São Dimas, Caixa Postal 9, 13418-900, Piracicaba, São Paulo, Brazil. email@example.com (corresponding author); firstname.lastname@example.org
IIAgronomist Engineer, Doctor Science. Centro de Energia Nuclear na Agricultura, Universidade de São Paulo (CENA/USP), Avenida Centenario, 303, Bairro São Dimas, Caixa Postal 96, 13400-970, Piracicaba, São Paulo, Brazil. email@example.com
IIIAgrgricultural Engineer, Doctor Science. Departmento de Engenharia Agrícola, Universidade Federal de Pelotas (UFPel), Campus Capão do Leão, s/n, Caixa Postal 354, 96001-970, Pelotas, Rio Grande do Sul, Brazil. firstname.lastname@example.org
IVAgronomist Engineer, Doctor Science. Departmento de Solos, Universidade Federal de Santa Maria (UFSM), Avenida Roraima, 100, 97105-900, Santa Maria, Rio Grande do Sul, Brazil. email@example.com
VAgronomist Engineer, Master of Science. Doctor Science student at Universidade Federal de Santa Maria (UFSM), Avenida Roraima, 100, 97105-900, Santa Maria, Rio Grande do Sul, Brazil. firstname.lastname@example.org
Nitrogen has a complex dynamics in the soil-plant-atmosphere system. N fertilizers are subject to chemical and microbial transformations in soils that can result in significant losses. Considering the cost of fertilizers, the adoption of good management practices like fertigation could improve the N use efficiency by crops. Water balances (WB) were applied to evaluate fertilizer N leaching using 15N labeled urea in west Bahia, Brazil. Three scenarios (2008/2009) were established: i) rainfall + irrigation the full year, ii) rainfall only; and iii) rainfall + irrigation only in the dry season. The water excess was considered equal to the deep drainage for the very flat area (runoff = 0) with a water table located several meters below soil surface (capillary rise = 0). The control volume for water balance calculations was the 0 - 1 m soil layer, considering that it involves the active root system. The water drained below 1 m was used to estimate fertilizer N leaching losses. WB calculations used the mathematic model of Penman-Monteith for evapotranspiration, considering the crop coefficient equal to unity. The high N application rate associated to the high rainfall plus irrigation was found to be the main cause for leaching, which values were 14.7 and 104.5 kg ha-1 for the rates 400 and 800 kg ha-1 of N, corresponding to 3.7 and 13.1 % of the applied fertilizer, respectively.
Key words: Penman-Monteith, evapotranspiration, deep drainage, urea.
O nitrogênio possui uma dinâmica complexa no sistema solo-planta-atmosfera. Considerando o elevado custo dos adubos, é fundamental o desenvolvimento de manejos da adubação nitrogenada que visem ao melhor aproveitamento do N pelas culturas, como é o caso da fertirrigação e o mínimo impacto ambiental. Balanços hídricos e a lixiviação de N derivado do fertilizante são apresentados para um cafezal sob fertirrigação com uréia marcada com 15N no oeste baiano, em três cenários para um ciclo da cultura 2008/2009: i) precipitação + irrigação no ano inteiro, ii) apenas precipitação; e iii) precipitação + irrigação apenas na estação seca. Nos balanços hídricos os componentes ascensão capilar e escoamento superficial foram considerados nulos por se tratar de solo arenoso em declive praticamente nulo, com lençol freático profundo. A irrigação foi realizada por pivô-central e no balanço hídrico o volume de controle considerou a camada 0 - 1 m responsável pela disponibilidade de água pela cultura. A água drenada abaixo de 1 m foi considerada para os cálculos da lixiviação do nitrogênio do fertilizante. O balanço hídrico utilizado calculou a evapotranspiração baseado no modelo matemático de Penman-Monteith, considerando um coeficiente de cultura unitário. Foi possível verificar que a alta quantidade de N, associada precipitação concentrada são os grandes responsáveis pela lixiviação, cujos valores foram 14,7 e 104,5 kg ha-1 de N do fertilizante para as doses de 400 e 800 kg ha-1 de N aplicadas na forma de uréia, correspondendo a 3,7 e 13,1 % da quantidade total do fertilizante aplicado.
Palavras-chave: Penman-Monteith, evapotranspiração, drenagem profunda, uréia.
The dynamics of nitrogen in the soil-plant-atmosphere system is complex. The N fertilizer is subject to a series of chemical and biological transformations that can lead to significant N losses to the environment. Therefore, it is important to search for agricultural practices that enable a more efficient use of the applied N fertilizer, such as fertigation. Advanced farming practices have been adopted for coffee production to increase crop yields, such as denser planting, harvest mechanization and fertigation (Coelho & Silva, 2005). This has been the case of the western Bahia, in Brazil, in which the coffee production is only viable with irrigation (Silva et al., 2005).
Fertigation has several advantages in relation to conventional cropping practices, allowing for the control, monitoring and split of fertilization according to the plant requirements along the productive cycle (Coelho & Silva, 2005), although increasing the risk of losses to the environment (Oliveira et al., 2002); the remaining N stays in the soil, mainly in the organic form (Scivittaro et al., 2003; Silva et al., 2006). The N fertilization efficiency is also influenced by the irrigation management, N rates and intervals of applications (Quiñones et al., 2007). Results of the most different crop scenarios show that the absorption of the total N applied rarely exceeds 60% (Reichardt et al., 2009).
Several studies of N uptake have used 15N as a tracer to quantify this plant nutrient in the different compartments of a soil-plant- system (Lara Cabezas et al., 2000; Boaretto et al., 1999; Boaretto et al., 2007; Fenilli et al., 2004; Oliveira et al., 2002; Lima Filho & Malavolta, 2003). However, the widespread adoption of advanced farming practices such as N fertigation requires further studies on the interactions climate-soil-coffee. This study uses the 15N tracer to quantify N fertilizer leaching, and the consequent environmental risk of fertigation associated with high N fertilizer rates in a "cerrado" or savanna area, where the supplemental irrigation is used over the whole year.
MATERIAL AND METHODS
An area of fertigated coffee crop was chosen for this study due to the high fertilizer rate routinely used (600 kg ha-1 year-1 of N) over the last seven years, showing a high risk of N leaching below the crop root zone. This central pivot irrigated coffee crop belongs to a commercial farm (Fazenda Morena) located in Barreiras, BA, Brazil (11° 46' S and 45° 43' W, 740 m asl). The soil, previously covered by "cerrado" or savanna vegetation, was classified as a "LATOSSOLO VERMELHO-AMARELO Alumínico típico" (Embrapa, 2006) and as a Typic Hapludox (Soil Survey Staff, 2010), of low natural fertility, with 75% sand, 3% silt and 22 % clay, of sandy texture. The climate, according to Köppen's (1931) classification, is tropical sub-humid (Aw) with yearly rainfall ranging from 800 to 1800 mm concentrated between October and April, with a well defined dry season that for perennial crops, which requires supplemental irrigation, andannual average air temperature of 25 ºC. Table 1 shows the climate data obtained from the meteorological station of the "Instituto Nacional de Meteorologia" (INMET) of Barreiras. Rainfall and irrigation were measured in situ at the farm.
Soil water retention properties were evaluated for each 0.2 m layers (Table 2), leading to an available water capacity (AWC) of 86.4 mm for the 0 - 1.0 m soil layer. This soil depth was considered for water balance (WB) calculations, assuming that it contains 100% of the active coffee root system, so that all water fluxes below the 1.0 m depth are considered as deep drainage (Qi), used to estimate fertilizer N leaching. The chemical characterization of the soil (Table 3) indicates that this Oxisol has a very low natural fertility level that requires heavily fertilizer input for crop production.
Coffee plants (Coffea arabica L.), variety Catuaí Vermelho, were planted on January 2001, with a spacing of 3.8 m between circles (central pivot arrangement) and 0.5 m between plants in order to form a tier. During the experimental period, August 2008 to July 2009, plants were adult, 7 to be 8 years old, with about 3 m height and 1.9 m width, leaving a free inter-row of 1.9 m for machinery traffic. This cropland has been fertilized since 2002, with 600 kg ha-1 year-1 of N. The crop yield was an average of 56 bags for ha-1 year-1green bean (60 kg), almost three times higher than the Brazilian average (20 bags ha-1year-1).
Fertigation was performed with low energy precision application (LEPA) sprinklers, which distribute the solution in a localized form directly over the circular coffee rows, with minimum wetting of the inter-row. Irrigation depth is of the order of 3 to 4 mm day-1 applied in alternated days. During regular operation of fertigation N, the experimental row (row 4) did not receive the usual urea N fertilization in order to allow the application of labeled 15N urea, used to estimate N fertilizer leaching.
In order to evaluate WB components the sequential water balance (SWB) program suggested by Rolim et al., (1998) was used for five-day intervals and then monthly values were obtained. Considering the elemental volume of 1 m soil depth (assumed to contain 100% of the active root zone) the changes in soil water storage (ï²SWS) were calculated by equation (1):
P = rainfall (mm);
I = irrigation (mm);
ETr = actual evapotranspiration (mm);
CR = capillary rise (mm);
Qi = internal drainage (mm);
RO = runoff (mm).
The SWB program was set to estimate the potential evapotranspiration through the methodology of Penman-Monteith, adapted by Allen et al. (1989), and to consider the sum CR + Qi + RO = excess (EXC). Because the experimental area was flat (slope approximately zero) and has a very deep water table, we consider CR = 0 and RO = 0, so that the EXC given by the balance is equal to Qi.
In order to evaluate the effect of, I on Qi, WBs were carried out in the real scenario i) considering P + I, and, additionally, in two alternative scenarios: ii) no irrigation, only P; and iii) P + irrigation only in the dry season (Idry).
The application of 15N urea was made every 15 days (counting was made in days after beginning (DAB), starting September 1, 2008) on plots consisting of 3 plants of the cycle N° 4 of the central pivot, using a ladder and a watering can in order to simulate the LEPA sprinkler. The fertilizer was diluted in a volume of water corresponding to an irrigation of 4 mm. Plants bordering experimental plots received fertilization of solid urea on the soil surface. Two treatments with four replicate were tested, one below the normal N fertilization rate of the farm (600 kg ha-1 of N) and the other above, as following: 1) T400 (400 kg ha-1 of N), corresponding to 76 g plant-1 of N or 169 g plant-1 of urea; 2) T800 (800 kg ha-1 of N), corresponding to 152 g plant-1 of N or 338 g plant-1 of urea. All other fertilization and management practices were maintained as usually performed on the farm, and are shown elsewhere (Bruno et al., 2011).
Under the central plant of each plot, 0.2 m from the trunk were installed porous soil solution extraction probes at the depth of 1m, the lower boundary of the WB control volume (Figure 1A), to measure nitrate concentration by flow injection analysis (Giné et al., 1980) and abundance of 15N by mass spectrometry (ANCA SL Mass Spectrometer). Soil solution extractions were made at least once a week before of the application of N fertilizer.
Average nitrate ion concentrations (Ci) and average abundances (15Ni), for month i, were estimated taking averages of solutions collected at each month.
Because of the low Ci values and the need of having a minimum N quantity for the isotope analysis, solutions collected from the replicates were joined in a single sample for each date. This fact did not allow a statistical analysis of the variability due to replicates. The quantity of leached N for month i, derived from fertilizer (QNdffi, kg ha-1) was calculated using the equation (2):
Where: Qi (kg ha-1) is the value calculated by the SWB program, first given in mm and then transformed in kg of drained water per ha;
Ci initially expressed in mg L-1 of the nitrate ion, was transformed in kg of N per kg of water, i.e., kg kg-1 of N;
Ndffi is the fraction of nitrogen derived from the fertilizer, given by the equation (3):
Because our leaching calculations were based on the hectare, and the fertilizer N leaching occurred only in the area of fertilizer application, which is the area effectively used by the plants ( 0.5 x 1.90 = 0.95 m2 , which is less than the area occupied by one plant 0.5 x 3.8 = 1.90 m2, Figure 1A), it was assumed the effective area per plant for leaching calculations was 1.425 m2, based on the assumption that the N displacement from the soil surface to the depth of 1 m is dispersed, having 100% as probe concentration at the edge of the plant canopy and 0% in the middle of the inter-row (Figure 1B) that does not receive N fertilizer.
The annual leaching loss PL (kg ha-1 year-1 of N) is simply the sum of the monthly values, given by equation (4):
The significance level of R2 was determined by the JMP IN software version 3.2.1 (Sall et al., 2005).
RESULTS AND DISCUSSION
Losses of nitrogen by leaching are important and need to be estimated in order to improve fertilizer recommendation in coffee, especially in sandy soils of the "cerrado" region of western Bahia with low soil fertility, where large amounts of N fertilizers are usually applied every year (400; 800 kg ha-1 year-1 of N). The inward flow Qi and drain of N soil to a depth of 1 m are shown in Table 4 for scenario i, providing evidence of the potential pollution of the ground water by nitrate.
During the complete coffee crop cycle, for scenario i, Qi amounted to 1010.5 mm (Table 4), with P + I = 2232.3 mm, corresponding to 45.3% of the total water inflow, and as expected, showing that the management of irrigation can partially control Qi. Irrigation amounted to 697.3 mm and the actual evapotranspiration to 1270.4 mm. It is clear that irrigation (31.2% of P+I) is a strong contributor to Qi, mainly during the wet season (Figure 2A) when I is not necessary for demand of water in the plant, but that is still performed as fertilizer application (fertigation). For the alternative scenario ii, with only P, Qi was reduced to 811.5 mm, and scenario iii with irrigation only in the dry season, Qi was reduced to 873.1 mm.
Monthly averages of Ci in soil solution (Figure 2B) do not follow either the P + I distribution, or the continuous and cumulative applications of fertilizer. Data were very scattered and show only a tendency of an increase in time through the linear regressions (Figure 2B), with low R2 coefficients.
The abundant average monthly of % 15N abundances (Figure 3) also do not present the expected distribution that would be an asymptotic behavior with decreasing increments and a tendency of stabilization due to the progressive 15N applications. Data are less scattered as Ci and also show a tendency to increase with time, and linear regressions with significant R2 coefficients. Figure 3 also shows the continuous lines of the urea applications, starting at 1 DAB with 15.4 for T400 and 30.8 kg ha-1 of N for T800, and ending at 350 DAB with 400 and 800 kg ha-1 of N, respectively.
The distributions of Ci and %15N in the deep drainage water are actually a function of several processes occurring in the atmosphere (rainfall and irrigation), in the plant (absorption and redistribution of N), and in the soil (physico-chemical and biological N transformations), which explain the scattering of the data. Details of plant N uptake evaluated at the same site and time, can be found in Bruno et al. (2011).
With the monthly data of Ci, 15Ni and Qi , the leached QNddfi was calculated by equation (2). Figure 4A shows that the peak of leached N happens in March (227 DAB), when 66.7% of the fertilizer had already been applied by fertigation. The difference between treatments T800 and T400 at this leaching peak was of 33 kg ha-1 month-1of N, or 9.3 times higher for T800 in relation to T400. The great difference between treatments is also a strong argument to justify the no use of statistical differences between them, because replicates were lost when making composite samples. The results showed that doubling the rate from 400 to 800 kg ha-1 year-1 of urea N led to about seven-time more leached N fertilizer. Bruno et al., (2011) demonstrated that the rate of 800 kg ha-1 is too high in terms of plant uptake efficiency and recommended the rate of 200 kg ha-1, which would also reduce leaching losses significantly.
The accumulated leached N data (Figure 4B) also showed great differences between the treatments T400 and T800, which at the end of the cropping cycle amounted to 14.7 and 104.5 kg ha-1, respectively, or about 10 times more in favor to T800. The accumulated data of QNdffi presented a good relationship with accumulated Qi, for both treatments (Figure 5A), which was expected because calculations of the first included the second. Anyway, it was an interesting result, mainly in view of the scatter of Ci and 15Ni data. Although the best regressions are third-order polynomials, the very low coefficients of x3 and x2 (Figure 5A) indicate that the relationship is essentially linear.
Similarly good relations were found for accumulated data of QNdffi and P + I (Figure 5B) showing the importance of the water input in the amount of leached N. Again, the coefficients of x3 and x2 (Figure 5B) indicate almost linear relations.
Nario et al. (2003), stress the fact that the leaching process depends on irrigation management. Quiñones et al. (2005) reported that the response to N fertilizer is influenced by irrigation methods, frequency and application timing, as well as by the processes of nitrification, imobilization, denitrification, volatilization and leaching. Boaretto et al. (2007) report that for citrus trees, the absorption of N is hindered in rainy periods due to leaching losses. According to Oliveira et al. (2002) and Franco et al. (2008) N losses by leaching may be negligible because most of the 15N studies in sugarcane indicate very little leaching losses, as also recently reported by Ghiberto et al. (2011), however, for Silva et al. (2006) and Duete et al. (2008) these low losses cannot be disregarded.
In our study, to understand better the impact of irrigation on QNdffi, three scenarios were analyzed: i, the first that really occurred under field conditions and the simulations ii and iii (Table 5).
For scenario i, in T400, the leaching amounted to 14.7 kg ha-1 of N fertilizer, which corresponds only to 3.7% of the total N applied. For scenario ii, this amount was only reduced to 12.1 per ha or 3.0%, showing that the major contributor to QNdff is P not I. Irrigation only during the dry season does not change very much this trend (scenario iii, Table 5). Because irrigation is widely spread for coffee plantations in this region, fertigation can be also applied during the wet season, however, following the recommendations of Bruno et al. (2011). Fenilli et al. (2008) reported a leaching of 6.5 kg ha-1 of N, corresponding to 2.3% of the total rate, in an area where 280 kg ha-1 of N rate was applied. In the second year, with a dose of 350 kg ha-1 of N, the leaching remained at the same 2.3% of the total N applied. Oliveira et al. (2007) also reported very low leaching losses in a pasture, however under very low N rates.
Results of treatment T800 were more than five times higher, showing that doubling the N fertilizer rate the effect on leaching is highly significant, leading to a loss of about 12% of the total applied urea. Again, for scenarios ii and iii the N leaching losses were not significantly reduced showing that the main responsible for this result is the rainfall, and that a reduction in N leaching can only be achieved by reducing the amount of N applications. As already mentioned, in other regions and other experiments the N leaching component was always of minor importance. Reichardt et al. (2009) also working with coffee found N leaching of 2.3% of the total N applied, Gava et al. (2006) for corn and Boaretto et al. (2004) for wheat reported losses of 1%, and Quiñones et al. (2005 and 2007) for citrus under controlled environment found only 0.1% of N losses.
Sequential water balance calculations using the model of Penman-Monteith for potential evapotranspiration and measured values of rainfall and irrigation, indicate that high amount of N fertilizer applied in fertigated coffee plantations in association with the high volume of precipitation (rainfall plus irrigation) are the main causes of N leaching in western Bahia, in Brazil. N fertilizer leaching of 14.7 and 104.5 kg ha-1 year-1 were recorded for the rates 400 and 800 kg ha-1 year-1of N fertigated during the whole year, corresponding to 3.7 and 13.1% of the total amount of fertilizer input, respectively. Restricting irrigation only to the dry season, reduced N losses only to 12.3 and 91.0 ha-1year-1, showing that rainfall is the major determinant of N leaching.
The authors express their gratitude to CNPq and FAPESP for financial support and fellowships.
Allen RG, Jensen ME, Wright JL & Burman RD (1989) Operational estimates of reference evapotranspiration. Agronomy Journal, 81:650-662. [ Links ]
Boaretto AE, Shiavinatto-Neto P, Muraoka T, Oliveira MW & Trivelin PCO (1999) Foliar nitrogen supply to young citrus plants. Scientia Agricola, 56:621-626. [ Links ]
Boaretto AE, Spolidorio ES, Freitas JG, Trivelin PCO, Muraoka T & Cantarella H (2004) Fate of 15N-urea applied to wheat-soybean succession crop. Bragantia, 63:265-274. [ Links ]
Boaretto RM, Matos Junior D, Trivelin PCO, Muraoka T & Boaretto AE (2007) Nutrient accumulation and fate of nitrogen (15N) in young bearing orange trees. Revista Brasileira de Fruticultura, 29:600-605. [ Links ]
Bruno IP, Unkovich MJ, Bortolotto RP, Bacchi OOS, Dourado-Neto D & Reichardt K (2011) Fertilizer nitrogen in fertigated coffee crop: absorption changes in plant compartments over time. Field Crops Research, 124:369-377. [ Links ]
Coelho G & Silva AM (2005) Irrigation seasons and splitting fertilizer effects on coffee plant productivity of three crops consecutive. Ciência e Agrotecnologia, 29:400-408. [ Links ]
Duete RRC, Muraoka T, Silva EC, Trivelin PCO & Ambrosano EJ (2008) Nitrogen fertilization management and nitrogen (15N) utilization by corn crop in red latosol. Revista Brasileira da Ciência do Solo, 32:161-171. [ Links ]
Embrapa - Brazilian Agricultural Research Corporation (2006) Brazilian System of Soil Classification. 2ª ed. Rio de Janeiro, National Research Center for Soil. 306p. [ Links ]
Fenilli TAB, Boaretto AE, Bendassolli JA, Trivelin PCO & Muraoka T (2004) Nitrogen dynamics in young orange trees grown in nutrient solution. Laranja, 25:461-472. [ Links ]
Fenilli TAB, Reichardt K, Favarin JL, Bacchi OOS, Silva AL & Timm LC (2008) Fertilizer 15N balance in a coffee cropping system: a case study in Brazil. Revista Brasileira de Ciência do Solo, 32:1459-1469. [ Links ]
Franco HCJ, Trivelin PCO, Faroni CE, Vitti AC & Otto R (2008) Utilization by sugarcane cane of nitrogen applied at planting. Revista Brasileira de Ciência do Solo, 32:2763-2770. [ Links ]
Gava GJC, Trivelin PCO, Oliveira MW, Heinrichs R & Silva MA (2006) Balance of nitrogen from urea (15N) in the soil-plant system at the establishment of no-till in maize. Bragantia, 65:477-486. [ Links ]
Ghiberto PJ, Libardi PL, Brito AS & Trivelin PCO (2011) Nitrogen fertilizer leaching in an oxisol cultivated with sugarcane. Scientia Agricola, 68:86-93. [ Links ]
Giné MF, Bergamin H, Zagatto EAG & Reis BF (1980) Simultaneous determination of nitrate and nitrite by flow injection analysis. Analytica Chimica Acta, 114:191-197. [ Links ]
Köppen WP (1931) Grundriss der Klimakunde. Berlin, Walter de Gruyter. 388p. [ Links ]
Lara Cabezas WAR, Trivelin PCO, Kondörfer GH, Pereira S (2000) Nitrogen balance of solid and liquid nitrogen sources on side-dressed no-till corn system at Triângulo Mineiro. Revista Brasileira de Ciência do Solo, 24:363-376. [ Links ]
Lima Filho OF & Malavolta E (2003) Studies on mineral nutrition of the coffee plant (Coffea arabica L. cv. Catuaí Vermelho) LXIV. Remobilization and re-utilization of nitrogen and potassium by normal and deficient plants. Brazilian Journal of Biology, 63:418-490. [ Links ]
Nario A, Pino I, Zapata F, Albornoz MP & Baherle P (2003) Nitrogen (15N) fertilizer use efficiency in peach (Prunus persica L.) cv. Goldencrest trees in Chile. Scientia Horticulturae, 97:279-287. [ Links ]
Oliveira MW, Trivelin PCO, Boaretto AE, Muraoka T, Mortatti J (2002) Leaching of nitrogen, potassium, calcium, and magnesium in a sandy soil cultivated with sugarcane. Pesquisa Agropecuária Brasileira, 37:861-868. [ Links ]
Oliveira PPA, Trivelin PCO & Oliveira WS (2007) Urea 15N balance in the fractions of a Brachiaria brizantha cv. Marandu pasture under recovery with different liming dates. Revista Brasileira de Zootecnia, 36:1982-1989. [ Links ]
Quiñones A, Bañuls J, Primo-Millo E & Legas F (2005) Recovery of the 15N-labelled fertilizer in citrus trees in relation with timing of application and irrigation system. Plant and Soil, 268:367-376. [ Links ]
Quiñones A, Martínes-Alcántara & B Legas, F (2007) Influence of irrigation system and fertilization management on seasonal distribution of N in the soil profile and on N-uptake by citrus trees. Agriculture, Ecosystems and Environment, 122:399-409. [ Links ]
Raij B van, Andrade JC, Cantarella H, Quaggio JA (2001) Chemical analysis for evaluation of the fertility of tropical soils. Campinas, Instituto Agronômico. 285p. [ Links ]
Reichardt K, Silva AL, Fenilli TAB, Timm LC, Bruno IP & Volpe CA (2009) Relation between nitrogen fertilization and water soil conditions for a coffee plantation from Piracicaba, SP. Coffee Science, 4:41-55. [ Links ]
Rolim GS, Sentelhas PC & Barbieri V (1998) Spreadshets in ExcelTM environment to calculation of water balance: normal, sequential, culture, and potential, real productivity. Revista Brasileira de Agrometeorologia, 6:133-137. [ Links ]
Scivittaro WB, Muraoka T, Boaretto AE & Trivelin PCO (2003) Transformations of nitrogen from velvet bean and urea used as fertilizers in corn crop. Pesquisa Agropecuária Brasileira, 38:1427-1433. [ Links ]
Silva AM, Coelho G & Silva RA (2005) Irrigation timing and split application of fertilizer on productivity of the coffee plant in 4 harvests. Revista Brasileira de Engenharia Agrícola e Ambiental, 9:314-319. [ Links ]
Silva EC, Muraoka T, Buzetti S, Veloso MEC & Trivelin PCO (2006) Utilization of nitrogen (15N) from sun hemp and millet by corn crop under no-tillage in Red Latosol of Cerrado. Ciência Rural, 36:739-746. [ Links ]
Soil Survey Staff (2010) Keys to Soil Taxonomy. 11ª ed. Washington DC, USDA-Natural Resources Conservation Service. 338p. [ Links ]
Sall J, Creighton L & Lehman A (2005) JMP start statistics: A guide to statistics and data analysis using JMP and JMP IN software. 3ª ed. Cary, Duxbury Press. 580p. [ Links ]
Received for publication on March 5th, 2012 and approved on August 10th, 2012.
* Part of the first author's Doctor Science Thesis. Work funded by CNPq and FAPESP.