| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agricultural Research Council, Agronomical Research Institute, Modena Section, Viale Caduti in Guerra 134, I-41100 Modena, Italy
b Agricultural Research Council, Institute for Tobacco Research, Bovolone Section, Via Canton 14, I-37051 Bovolone (Verona), Italy
c Agricultural Research Council, Institute for Tobacco Research, Via Pasquale Vitiello 66, I-84018 Scafati (Salerno)
* Corresponding author (rosa.marchetti{at}entecra.it)
Received for publication April 11, 2005.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: AGN, plant aboveground N removal • AGNmax, plant maximum aboveground N removal • DAT, days after tobacco transplanting • DATmax, days to reach AGNmax • DM, dry matter • DTH, days to harvest • Nfert, fertilizer N input • 
Ninorg, change in soil inorganic N storage • Norg, change in soil organic N storage
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The high quality of certain tobacco types is demonstrated by the remarkable amount of tobacco (59.4%, by weight) exported to countries requiring good quality standards (EU members, USA, and Japan; Nomisma, 2003). Flue-cured tobacco constitutes 38% of the total tobacco production, which is located mainly in the Umbria and Veneto regions.
In the Veneto region, the Verona province produces 17 500 t of flue-cured tobacco (Coop. Tabacchicoltori Associati Veneti, personal communication, 2005). Cultivation areas are concentrated in the province's southern plain. Soils in this area are light, coarse, porous, and deep. Water abundance in this area permits frequent irrigations.
Flue-cured tobacco is very exacting in its N requirement. The regulation of the amount and timing of N availability is extremely important (Akehurst, 1981; Flowers, 1999; Peedin, 1999; Reed and Jones, 2002; Smith and Wood, 2005). Available N is needed to sustain full growth until flowering, because a N deficiency in this period is likely to result in a yield reduction. After this stage, N overfertilization leads to a reduction in cured-leaf quality (harsh tissue, darkening, reduction of strip yield) and commercial value, makes mechanical harvesting and curing more difficult, gives rise to N pollution of soils and underground waters, and increases levels of health-harmful constituents in tobacco leaves, like nitrates and nitrites. Therefore, available N in the soil should range within relatively narrow limits (Hawks and Collins, 1983).
According to scientific literature and to fertilizer recommendations of the local extension services, the most suitable N rate for flue-cured tobacco usually ranges between 40 and 90 kg ha–1, with differences depending on the soil type (Parker et al., 1993; Ramachandram et al., 1995; Adams and Mitchell, 2000; Kidder et al., 2002; Crozier et al., 2004; Smith and Wood, 2005). On loamy sand soils an application of supplemental N to replace presumably leached N during the early growth season is often suggested.
The usefulness of soil tests for estimating the residual soil N availability at the beginning of the crop growth season has been debated (Peedin, 1999; Reed and Jones, 2002; Moore and Harris, 2005). In fact N can also become plant available throughout the crop growth season, due to mineralization of soil organic N (Standford, 1982) and of N from previous crop residues (Aulakh et al., 2001).
In Italy, there are not any official recommendations. The common trend is to prevent soil N deficiencies. This goal is pursued by supplying dairy manure to soil early in the growth season. Inorganic fertilizer N is then applied before or at transplanting and then side-dressed again within a month.
The lack of response to N fertilization for standard N rates has sometimes been observed in local management practice. We hypothesized that more N may be available in soil for tobacco than expected on the basis of soil tests at the beginning of crop growth season. The aims of this work were therefore: (i) to verify the influence of N rate on flue-cured tobacco yields and (ii) to estimate the contribution of the soil organic N stock to crop nutrition. This contribution was evaluated by means of an N balance applied to the soil–plant system.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The study site is located in the Po river alluvial plain, with an elevation of 24 m. Soils of this area are flat, with 0.1% slopes, and belong to the evolutionary sequence of fine sand terraced areas. The climate is temperate suboceanic, typic ustic, with an annual temperature of 13.0°C, potential annual evapotranspiration of 765 mm, and annual precipitation of 699 mm (means of the 1956–2004 period). In the location of the experiment the water table levels always stay well below the cultivated soil layer. More information on the study site, climate, and soil can be found in Costantini et al. (2002).
The experimental design included five rates of fertilizer N (0, 20, 40, 60, and 80 kg N ha–1) in a randomized block with two replications, for a total of 10 plots. Only two replications were considered acceptable, their numerical scarcity being compensated by the soil sampling temporal intensity. Each plot measured 200 m2. The plant distance was 1 m between rows and 0.45 m within rows, corresponding to 444 plants per plot (22 222 plants ha–1).
The tobacco plants were transplanted on 13 May 1998 and 14 May 1999. Nitrogen fertilizer was side-dressed as calcium nitrate 2 wk after transplanting. Based on soil analysis, no P fertilizer was applied. Potassium was applied at the rate of 250 kg ha–1 (as potassium sulfate) before the transplanting time. Chemical treatments to control weeds, aphids and blue mold were also applied. Topping was performed in both years at the end of July, at the flowering button stage, and it was followed by three sprays of chemicals for sucker control (twice with fatty alcohol, once with maleic hydrazide). Sprinkler irrigation was applied from transplanting to the end (in 1998) or the middle (in 1999) of September. Frequent irrigations with reduced water volumes were performed, to avoid runoff water losses (12 irrigation events in 1998, for a total of 249 mm; 7 events in 1999, for a total of 158 mm). Leaves were hand-harvested as they ripened, from the end of July to the end of October, and four sequential harvests (primings) were performed in total in each plot, following the traditional practice of the region.
Plant Sampling and Analysis
Plant data were collected at two sampling scales, with two goals: (i) at plot scale, to measure selected crop production traits, and (ii) at plant scale, to evaluate the plant aboveground N uptake, for balance purposes.
At plot scale, the cured-leaf yield and total N content were measured for a plant row, including 50 individuals, in the central area of each plot. In each plot a two-row buffer zone was excluded from sampling. Days to harvest (DTH), as an index of tobacco-leaf early harvesting, were determined at plot scale, as follows:
 | [1] |
At plant scale, we assumed the plant aboveground N uptake to coincide with the plant maximum aboveground N removal during the crop growth season (AGNmax). The AGNmax was determined as follows. In both years of the experiment, starting about 1 mo after tobacco transplanting, when the small plants were well established, two contiguous plants were collected weekly in each plot, from random positions. Dry matter (DM) of leaves, stems, and flower heads (inflorescence and uppermost leaves) was measured for each individual plant. Dry plant samples (leaves, stems, flowers heads, and harvested mature leaves) were used for the tissue N concentration analysis. For each plant part, N content was calculated as DM times N concentration. At each sampling date, the aboveground N removal (AGN) for a given treatment and replicate was the sum of N content in leaves, stems, and flower heads.
For each year, treatment and replicate, the estimate of AGNmax, and the time needed to reach AGNmax, DATmax, was obtained by fitting the weekly AGN measurements to a parabola (Fig. 1
), by means of quadratic regression analysis:
 | [2] |
|
 | [3] |
Plant chemical analyses were performed with an automatic chemical analyzer Technicon AutoAnalyzer II (Technicon Industrial Systems, Tarrytown, NY 10591).The analysis of total ammonium N in plant tissues (leaves, stems, harvested mature leaves) after Kjeldahl digestion of plant material was based on a colorimetric method, given in Technicon Industrial Methods (1977). Nitrite N, and nitrate N after reduction to nitrite with hydrazine sulfate and Cu2+, were determined according to Cooperation Centre for Scientific Research Relative to Tobacco (1994). Endogenous nitrite N concentration in plant tissues was lower than the detection limit. All the samples were analyzed in duplicate.
Soil Sampling and Analysis
Selected properties of the soils of the experiment are reported in Table 1. Bulk density values were: 1.54 Mg m–3, in the Ap1 horizon (top 0.28-m soil layer); 1.69 Mg m–3, in the Ap2 horizon (from 0.28 to 0.45 m); and 1.48 Mg m–3, in the Bw1 horizon (from 0.45 to 0.9 m) (Costantini et al., 1992).
|
Organic C (Walkley and Black), total N (Kjeldahl), Olsen P, and exchangeable K (ammonium acetate extractable) concentrations, and pH (water-saturated paste) were measured on the air-dried samples according to Page et al. (1982). Soil texture was measured according to Gee and Bauder (1986). Soil inorganic N was extracted from the frozen soil samples, after thawing them at room temperature, by using 2 M KCl (soil/solution ratio, 1:5). Nitrates were reduced to nitrites by passing them through a cadmium column; nitrites and ammonium were determined colorimetrically (with sulphanilide and dichloroisocyanuric acid, respectively) with an automatic analyzer (AutoAnalyzer 3; Bran + Luebbe GmbH, Norderstedt, Germany) according to Keeney and Nelson (1982). The total amount of inorganic N in the soil profile was obtained as the sum of nitrate, nitrite, and ammonium N in the four soil layers where measurements had been performed.
Results relevant to the data set collected at the plant-scale level (including both plant and soil data) were reported on a per square meter basis.
Nitrogen Balance
The general conservation of mass equation for any soil–crop system is: N inputs – N outputs = change in total N (organic N plus inorganic N) stored within the system (Meisinger and Randall, 1991). Thus, the change in the organic N storage (Norg) was calculated as follows:
 | [4] |
Ninorg is the change in the soil inorganic N storage. In our system the main N input was represented by the fertilizer N, Nfert. Due to lack of measurements of N leaching and denitrification losses, as well as of plant root-N uptake, the N balance took the following simplified form:
 | [5] |
Norg is the change in soil organic N storage, Nfert is the fertilizer N, AGNmax is the maximum plant aboveground N removal and Ninorg is the change in the soil inorganic N storage. As the fertilizer N was supplied in the nitrate form, ammonia losses to the atmosphere were judged as not influencing the whole N balance.
The N balance was applied to the top 0.8-m soil layer, between the transplanting time and the time the crop had reached the maximum AGNmax, that is DATmax. As AGNmax was reached at different times, depending on year, treatment, and replicate, in both years we decided to close the balance at the time when all the plots had reached AGNmax, that is at the DATmax of the plot showing the latest AGNmax. This choice was due to the need to avoid any underestimation of the crop N uptake, as it would occur by closing the balance at a time when not all the plots had reached AGNmax.
Statistical Data Analysis
The relationships between N rate and selected dependent variables (DTH, cured-leaf yield, cured-leaf N content, AGNmax, DATmax, Ninorg, and Norg) were studied by linear regression analysis. The fitted relationships in 1998 and 1999 were compared to check for differences in the two data sets. When there was evidence that the relationships for the two data sets were not identical, the possibility for their slopes being the same (that is, for lines to be parallel and the observed differences to be due only to the year effect) was also explored. Curvilinear and linear regression analyses were performed using the SAS regression procedure (SAS Institute, 1987).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In 1998, the effect of N rate on tobacco yield was not significant (Table 2). Higher fertilizer N rates allowed a significant (P < 0.05) increase in cured leaf yields to be obtained only in 1999. In 1999, a unit increase of fertilizer N produced an increase of 5.0 kg cured leaf ha–1 (slope of the regression). In other words, the N rate of 80 kg ha–1 increased cured-leaf yield by 402 kg ha–1, compared with the unfertilized control. This means that, in 1999, at the highest N rates, tobacco yield was stabilized on 3900 kg ha–1 (precisely, 3941 kg ha–1, obtained as the sum of the intercept, 3539 kg ha–1, and of the further 402 kg of cured leaves due to fertilizer). The maximum yield obtained in 1999 was lower than the mean yield obtained in 1998. Lower tobacco yields in 1999 were attributed to water-logging episodes that occurred in the weeks immediately following the tobacco transplant (Ceotto and Castelli, 2002).
|
The mean total N concentration in cured leaves was equal to 29.2 g N kg–1 DM in 1998, and to 22.1 g N kg–1 DM in 1999. These concentrations have the same magnitude as those reported by other authors (Tso, 1990; Parker et al., 1993; Ramachandram et al., 1995). The total N concentration in the cured leaves significantly increased for increasing N rates (Table 2), in both years of the experiment. In 1999, the N increase in leaves for a unit increase of supplied N was significantly lower than in 1998.
Nitrogen Balance vs. Fertilizer Nitrogen Rate
Plant Aboveground Nitrogen Removal
In 1998 the crop N removal was on the whole higher, and the N removal plateau was reached faster than in 1999 (Fig. 1). On the basis of the regressions coefficients for all the measured points, for 1998 an AGNmax equal to 17.0 g N m–2 (corresponding to 7.65 g plant–1) after 112 DAT was calculated; for 1999 the AGNmax value was of 15.9 g N m–2 (corresponding to 7.16 g N plant–1), after 122 DAT. In 1999 the aforementioned delay in the start of the linear growth phase, due to water logging, may have caused the observed differences in the amount and rate of plant N removal.
The measured N removals were higher than those more frequently found for flue-cured tobacco in the related literature (Goenaga et al., 1989; Prasada Rao et al., 1992; Peiguo et al., 1997; Mitchell, 1999; Ryding et al., 2004). As the N concentration in the plant tissues (data not reported) as well as the total N concentration in the cured leaves (Table 2) were average, the higher N removal observed in our experiment must be attributed not to higher N uptake per crop unit weight, but rather to a higher biomass (data not reported) and flue-cured leaf production.
On the basis of a regression-based estimate (Fig. 1), it was also possible to verify that 20% of the soil N was taken up between 73 and 112 DAT, in 1998, and between 81 and 122 DAT, in 1999, between the topping and the second-priming time. Our results differ from those of Raper and McCants (1967), who found that, in normal growth conditions, 95% of the soil N was taken up within 9 wk after transplanting. Our results agree with those of Goenaga et al. (1989), who demonstrated that more than 20% of the total N in the tissues (excluding roots) at harvest was taken up between crop Day 83 and 127. Differences in these results are likely to be linked to differences in cultivars and climate conditions of the experiments.
The maximum plant N removal (AGNmax) linearly increased for increasing N rates. No significant differences between the regressions on the 1998 and 1999 measurements were detected (Table 3). In both years no significant effect of the N rate was observed on the time taken to reach AGNmax, that is on DATmax. Therefore, the leaf ripening delay that was observed at the highest N rate does not seem related to an extension of the plant N uptake period, but rather to a higher amount of N taken up by the crop in the same period.
|
|
|
The amount of available inorganic N in soil at the transplanting time (Ninorg-ini) was on average equal to 8.66 g N m–2 in 1998, and 13.17 g N m–2 in 1999, and it largely exceeded the subsequent fertilizer N supply, especially in 1999 (Table 4). The amount of available soil inorganic N at the end of the balance period (Ninorg-end) was lower than that measured at the beginning, except for the 6 and 8 g m–2 N rates, in 1998.
The change in the soil inorganic N storage linearly increased during the balance period for increasing fertilizer N rates (positive values of the Ninorg regression slopes, Table 5). In the absence of fertilization, during the balance period the inorganic N storage decreased more in 1999 than in 1998, as shown by the 1998 and 1999 intercept values in the Ninorg regression vs. N rate, Table 5. Denitrification, favored by the aforementioned water-logging episodes, may be responsible for the reduction of the amount of nitrates that were in the soil at the time of transplanting, in 1999 (95.5% of Ninorg-ini, Table 4).
|
Changes in Soil Organic Nitrogen Storage
The estimated change in the organic N storage was always negative (i.e., the N storage decreased) in both years (Table 4), and it linearly decreased for increasing N rates. Even though the decrease in the organic N storage was higher in 1998, the temporal patterns in the 2 yr of the experiment were similar (slopes not significantly different, Table 5). Consequently, according to these balance results, a fertilizer N unit increase produced an average decrease of 1.55 units (mean of the 2-yr slopes) in soil organic N content.
Soil cultivation brings about a decline of the soil organic N. The amount of organic N that can mineralize each year is reported to be commonly in the range of 0.5 to 3.5% of the soil total N (Broadbent, 1984). In our experiment the total N content, measured with the Kjeldahl method at the beginning of the balance period in the top 0.8-m soil profile, in the 2 yr of the experiment was on average equal to 761 g m–2 (n = 20, without significant differences between years). Therefore, on the basis of the balance Norg values (Table 4), the decrease of the initial amount of total N in the top 0.8-m soil layer varied between 0.77 and 4.37% (2.06%, on average), depending on year and N rate. Mineralization rates varying between 0.51 and 2.44 kg N ha–1 d–1, depending on year and N rate, were also calculated for the balance period (from mid-May to mid-September, 120 d), on the basis of the balance-estimated decrease in soil organic N.
Key protagonists of the soil N transformations are microfloral and microfaunal populations, and soil temperature and moisture levels may be identified as the factors mostly favoring the soil biota metabolic activities (Jarvis et al., 1996). The Bovolone soil has a mesic temperature regime, which is common in temperate climatic regions. In 1998 the mean soil temperature of the June and July months in the top 0.2-m soil profile was of 25.8°C in 1998 and of 24.7°C in 1999. These temperatures are ideal, when soil moisture is high, for the microbial activities. In fact the soil moisture at Bovolone is high in summer, due to frequent irrigations. Moreover the Bovolone's soil texture lacks fine, clayey particles (Table 1). This feature may be regarded as predisposing to soil N mineralization, because it has been reported that net mineralization of soil organic matter is more rapid in sandy soils than in clay soils (Hassink, 1992; Bosatta and Ågren, 1997). Organic N from previous wheat root residues may partly have contributed to mineralized N.
Our balance probably underestimated soil organic N decrease, for two reasons. First, tobacco-root N uptake, that may vary from 15 to 30% of the AGN (Elliot, 1967; Goenaga et al., 1989; Prasada Rao et al., 1992), was not included in the balance sheet. Second, N losses by leaching as well as denitrification could be predicted, on the basis of the hydrological (not reported) data. Had all these output items been measured, and explicitly included in the balance calculation, the estimated decrease of the soil organic N storage would have been even higher than that which was estimated through our results. An unaccounted for influence on the N balance of atmospheric N depositions, as well as N losses due to the release of ammonia in the atmosphere by the plant itself, may also have occurred. This kind of N loss has been reported for certain crops (Harper and Sharpe, 1995).
When comparing the linear increase of crop N uptake (AGNmax) for increasing N rates, and the inorganic (Ninorg) and organic N (Norg) changes vs. N rates (Table 4), it becomes evident that both crop N uptake and inorganic N storage increased at the expense of the organic N storage, the fertilizer N contribution being modest. We must therefore admit that a more intense soil N mineralization occurred as the consequence of increasing fertilizer N rates. This phenomenon may be regarded as a positive priming effect (Kuzyakov et al., 2000), that is, the occurrence of considerable increases in the soil inorganic N levels, following the supplying of inorganic N fertilizers. The occurrence of the priming effect has been reported for a long time (Jansson and Persson, 1982), either directly, by using labeled 15N fertilizer, or indirectly. Raun et al. (1998), in long-term, non-isotope-using experiments of N fertilization in continuous wheat under conventional tillage, hypothesized a positive priming effect (i.e., an increased net N mineralization from the soil organic matter pool) at low fertilizer N rates (<67 kg ha–1).
The estimated organic N decrease in the unfertilized plots represents the N amount that became available from the soil during the crop growth season, in addition to the initial inorganic N reserve. This amount is influenced by the viability of the microbial populations colonizing the soil; it may vary depending on soil type, pedoclimate and crop management, and may be estimated by measuring the potentially mineralizable N, a useful index of biologically active soil N. However, the determination of the potentially mineralizable N, as well as that of residual soil nitrate N content at the beginning of the crop growth season, may still not be sufficient to correctly predict the most suitable N rate for optimization of crop yield. This is the case when the situation is complicated by the occurrence of priming effect phenomena, as happened in our experiment. Unfortunately, mechanisms of the priming effect need to be further studied and understood, when seeking to predict their consequences on the N availability for crop nutrition. Our results are not sufficiently detailed to ascertain if the addition of inorganic N fertilizer was per se the reason for a mineralization stimulation, or if other factors, such as the emission of plant root exudates, may have acted, with impact on soil microbial populations and their activities.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The amount of available soil N, coming either from the residual nitrate N or from the inorganic N that was formed ex-novo in soil during the crop growth season, largely exceeded the N fertilization supply. More N became available, due to the triggering in soil of organic matter mineralization phenomena by N fertilization. Even though a clear relationship between N rate and N mineralization levels in soil could be found, it remains to be clarified if this priming effect was only a consequence of N fertilization, or rather if it was determined by different contributing causes.
Fertilizer recommendations, when not considering the soil N contribution throughout the crop growth season, may give rise to soil N excesses. These excesses, even when they do not have a negative influence on crop yields, nevertheless may cause negative effects on the farmer's income, either directly, by causing higher management costs, or indirectly, by negatively influencing the cured-leaf quality. Moreover, the N not taken up by the crop may produce a negative impact on the environmental quality. When aiming to estimate the most suitable time and way for N fertilization it is therefore necessary not only to correctly quantify the residual nitrate N content and the organic N biological quality in soil at the beginning of the season, but also to explore the possibility and the triggering causes of priming effect phenomena.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
