| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agricultural Research Center of Gembloux, Crop Production Dep., 4 rue du Bordia, 5030 Gembloux, Belgium
b Catholic Univ. of Louvain, Faculty of Biologic, Agronomic and Environmental Engineering, 2 place Croix du Sud, 1348 Louvain-la-Neuve, Belgium
* Corresponding author (goffart{at}cra.wallonie.be)
Received for publication April 12, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: DAE, days after emergence • HNT, Hydro N Tester leaf chorophyll meter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Improvement of N efficiency in this crop should be possible by matching crop N requirements and mineral N supply (soil and fertilizer as complements) throughout the growing season. According to Vos and MacKerron (2000), MacKerron (2000), and Shepherd (2000), this could be achieved by splitting the N fertilizer applications, combined or not with the monitoring of the crop N needs and soil mineral N supply.
That is not the current N fertilization practice, which consists of applying the entire amount of N either before or just around planting time. Low uptake efficiency is essentially due to the shallow root system of the potato, which is less efficient in taking up N than other crops such as cereals, sugarbeet (Beta vulgaris L.), or maize (Zea mays L.). Nitrogen present in soil layers deeper than 50 cm is poorly available to the potato crop. Moreover, as the potato crop is considered to have high N needs (4–6 kg N t–1 of fresh plants, i.e., about 250 kg N ha–1 for 50 t), fertilizer N rates currently applied at planting are higher than needed. The N uptake by the potato crop, however, is low before plant emergence and only starts to become intensive about 15 d after emergence. Westermann and Kleinkopf (1985) showed that 80 to 90% of the N is taken up during the 2 mo following emergence. Thus, during three or more weeks after N application, the risk of N losses from the plant–soil system by volatilization or leaching is high.
Furthermore, overfertilization can be profitable for the leaf canopy to the detriment of tuber production. Millard and Marshall (1986) showed that potato can take up more N than is needed to satisfy immediate requirements for growth. High rates of N application stimulate excessive canopy dry-matter production (Baker et al., 1980).
To minimize such risks of loss and overfertilization, methods were developed at the field scale to assess the total required N rate at the beginning of the season (Hofman and Salomez, 2000). They integrate basic available information on specific field characteristics such as soil type, humus content, and field history to assess, prior to planting, the soil mineral N supply during the season. Even with such a suitable assessment, however, the climate during the growing season and its interaction with the soil determines the real schedule of mineral N production by the soil and the effective crop N demand. Applying all of the N fertilizer at planting makes it impossible to make corrections according to the actual crop requirement during the growing period.
Bouldin and Selleck (1977) demonstrated that, in Long Island, maximum yields were obtained when fertilizer N applications were about 170 kg N ha–1, split with one-third of the amount applied at planting and the remaining two-thirds applied during the growing season, whereas farmers were currently applying about 225 kg N ha–1 at planting.
Hong et al. (2003) showed that split application of N did not increase tuber yield and N uptake. They suggested that N application should be matched with the real-time crop requirements by varying the time of split applications of N, or by developing a simulation model to predict N balance for N recommendation. This assumption had been previously formulated by Westermann et al. (1988), whose data indicated that a significant improvement in N fertilizer efficiency in the potato crop would result from split N fertilizer applications made according to crop growth needs.
The recently developed strategy summarized by MacKerron (2000) involves split fertilizer N applications, applying only part of a global N recommendation before emergence, and monitoring the crop after emergence to assess whether and when the crop needs supplemental N. Using the plant as an indicator appears to be more appropriate and convenient than the soil approach (sampling and analysis are time consuming). The plant analysis also offers the advantage of including the crop N requirement assessment. Such a procedure has been developed and described by Goffart and Olivier (2003). The global N recommendation is given according to the N balance-sheet method using the software AZOBIL, (INRA, Laon, France) developed by Machet et al. (1990). Only 70% of the N-recommendation level is applied at planting. Monitoring of the crop N status should then indicate the usefulness of supplemental N (i.e., the remaining 30% of the N recommendation), which depends on the intensity of the mineralization during the growing season and the real crop N demand.
To apply this strategy accurately, a tool is required to investigate the crop N status and to detect the need for supplemental N. The tool proposed is the HNT chlorophyll meter (Hydro Agri Europe, Dülmen, Germany). The determination of threshold values for the HNT in order to accurately make this decision was the main objective of this study. A good tool must be quick and easy to use under field conditions. It has to be sensitive to variations in plant N concentration, precise, and able to detect N deficiency early in the growing season. Another requirement for the tool is specificity to N. In addition, threshold values must be associated with the tool values to compare and support the decision to add N during the growing season.
Initially, invasive methods were proposed for use in the potato crop, based on the determination of the petiole NO3 concentration (PSC) either on dry matter (Gardner and Jones, 1975; Roberts et al., 1989) or in the sap (Williams and Maier, 1990; Vitosh et al., 1992). Threshold values of PSC to support the decision on the usefulness of supplemental N during the season were mainly set up by Van Loon and Houwing (1989) in the Netherlands and by Errebhi et al. (1998) in the USA. Such methods are interesting but have also been strongly criticized (MacKerron et al., 1995) and remain laborious. The first attempt to use noninvasive and quicker methods using the SPAD chlorophyll meter (Minolta, Osaka, Japan) on potato was proposed by Vos and Bom (1993). They showed that chlorophyll meter readings correlate well with analytical measurements of the chlorophyll content and with the N concentration in the leaves (r2 > 0.95).
Further investigations have been performed at our department on the use of the chlorophyll meter, in comparison with the petiole sap NO3 concentration approach. Conditions of use of both tools have been described in Olivier et al. (1999, 2001). The HNT is a recent version of the SPAD meter. This user-friendly measuring device is used directly on the foliage without invasive leaf sampling.
In studies recommending the use of a chlorophyll meter as a suitable tool for assessing the N status of a crop, practical interpretations of the values for managing N application are rarely proposed, especially in the case of potato. Factors linked to the site have significant effects on the HNT values so that raw HNT values are not specific to crop N status. Factors other than N have been identified as acting on HNT values, either on photosynthesis and chlorophyll concentration, or directly on the chlorophyll meter (Olivier et al., 1999, 2001). The effect of the cultivar is one factor, as leaf color differs among potato cultivars. The site factors can include soil type, climate, light, crop water status, and foliage diseases. Thus, the main drawback of the chlorophyll meter is its dependence on soil, climate, and crop aspects, i.e., its lack of specificity to the crop N status. An extensive review of these topics on the chlorophyll meter was recently given by Gianquinto et al. (2004). This drawback is not a distinctive feature of the HNT. Other tools for quick N status assessment, such as petiole sap NO3 concentration or crop reflectance, are also influenced by site factors. It must be noted that this disadvantage appears to be very important for the HNT. A decision support system based on this tool, to be accurate, has to bypass this disadvantage with a field-specific reference.
Whereas Singh et al. (2002) proposed raw chlorophyll meter values to determine the right time for N topdressing in rice (Oryza sativa L.) and wheat (Triticum aestivum L.), other scientists working with foliage optical tools conclude, like us, that absolute values are not relevant. The first idea was to use an overfertilized plot as reference (control). The principle is the following: a supplemental N application is considered to be required when the difference between the HNT value in the overfertilized plot and the HNT value in the test N plot is higher than a defined threshold value. Based on the same trials as those in this study, Denuit et al. (2002) showed that the proposal was relevant for winter wheat but not for potato.
Varvel et al. (1997) determined an N sufficiency index in corn with the use of the chlorophyll meter and a well-fertilized plot to obtain the greatest N efficiency. Laurent and Lancelot (personal communication, 2000), like Varvel et al. (1997), used a chlorophyll meter and an overfertilized plot as a field-specific reference to decide when a supplemental N application was needed for a potato crop. Raun et al. (2002) and Mullen et al. (2003) tried to improve N efficiency in cereal grain production using vegetation optical sensing. With a control treatment (zero N) as reference, they determined an in-season a priori estimated grain yield and an in-season crop response index to modulate N application at 1-m2 spatial resolution. In our trial fields, the possibilities of using relative HNT values (the relation of HNT values at the test N rate to those of the zero-N control) were explored. The second idea was the use of a reference with no N fertilizer.
The trials conducted from 1997 to 2000 intended to test the usefulness of the HNT and to define threshold values for potato based on a specific reference system that is easy to apply at field scale.
When the intention is to characterize the response of a crop to N fertilization a posteriori, generally only the yield response to N is taken into account. A response index using harvest data that indicates the actual crop response to applied N within a given year was recently proposed (Johnson et al., 2000). This response index, RIharvest, is calculated by dividing the average yield of the highest yielding treatment receiving N by the average yield of a control treatment (no N). For wheat, Mullen et al. (2003) determined the relationship between RIharvest and the response index measured during the season and based on sensor readings. To our knowledge, the application of this method to potato has not been reported. For potato, the response to N fertilizer in terms of total yield is not sufficient to comprehensively describe the agronomic efficiency of the N applied. For this reason, two further harvest indices were added to take production quality and environmental impact into account. This permitted calibration of a relation between HNT values and the N status of a potato crop in terms of the responsiveness to supplemental N. Application of this relation, once properly calibrated, makes possible a priori determination of the relevance of supplemental N from HNT values.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Summary of the Strategy
The global strategy, called conditioned N splitting, includes four stages (Goffart and Olivier, 2003). First, total N fertilizer is recommended using the software AZOBIL (Machet et al., 1990). Because it is a predictive recommendation, only 70% of the recommended rate is applied at planting. During the growing season (25–55 DAE [days after emergence]), the crop N status is monitored by the chlorophyll meter and only when the usefulness of supplemental N is indicated by this tool is the remaining 30% of the N recommendation applied. Thus, in part of fields, the remaining 30% is not applied and the total N rate is 70% of the initial recommendation.
In both the research trials and the validation fields, the test N treatment (70% of the N recommendation) applied on cv. Bintje was between 80 and 125 kg N ha–1. The N fertilizer recommendation range suited to this system of reference is 110 to 180 kg N ha–1 (=100%), which corresponds to the N recommendation range given for potato (cv. Bintje) in the Walloon area of Belgium (by AZOBIL and other recommendation systems).
Field Trials Conducted between 1997 and 2000 (Calibration Trials)
Twelve field trials with cv. Bintje were conducted from 1997 to 2000 on loam or sandy loam soils. The characteristics of the experimental sites and the trials are described in Table 1. For each experimental site, the balance-sheet method developed in AZOBIL (Machet et al., 1990) was used to give a field-specific total N rate recommendation at planting time (Table 2). The N treatments applied at planting in the trials represented 0, 75, 100, and 125% of the N-rate recommendation for the years 1997 and 1998, and 0, 70, 100, and 130% for the years 1999 and 2000. The N fertilizer was broadcast as NH4NO3 (solid granules, 27% N). The treatment with 70 or 75% is referred to as the test N treatment (Table 2). The trials were organized in a complete randomized block design with four replicates per treatment. A single experimental plot was composed of eight rows, 8 to 20 m long, depending on the available space at the different locations. Dates and planting densities of each trial are given in Table 1, which also gives the emergence and harvest dates, and rainfall during the growing season.
|
|
The initial fertilizer N and the remaining supplemental N were applied by the farmer either as NH4- or NaNO3 solid granules, or as a liquid foliar application in the form of diluted urea or N solution. All the strips were located in a homogeneous field area according to soil texture and crop practices. The zero-N plot was used as reference and was needed to calculate the HNT slope threshold. The other parts of the field trial were used to compare the yield and other parameters of the plot receiving our method of split applications (Treatment 4) with those receiving the other N treatments in order to evaluate the performance and validate the method proposed.
The Hydro N Tester Chlorophyll Meter
The HNT is a hand-held instrument measuring the light transmittance of a leaf at two specific wavelengths, based on the spectral absorption features of chlorophyll. The first wavelength is located in the red range (650 nm) specifically absorbed by chlorophyll, while the second, located in the infrared range (940 nm), is absorbed by cell walls and water but very slightly by chlorophyll and living vegetation. An internal algorithm using the ratio between both transmittance values determines a chlorophyll index related to the leaf chlorophyll content.
The measuring head of the HNT is composed of two parts, between which the leaf is inserted. The upper part emits two beams of light sequentially, while the receptor on the under part measures the transmittance of both wavelengths and records the data as a single value of the chlorophyll index. A mean index value is displayed on the LCD screen only after 30 single recorded readings not deviating by >3% from the mean value. In the trials, the single HNT readings were conducted as described by Olivier et al. (2001). The tip leaflet in the first fully developed leaf at the top was used for the reading. The HNT index (dimensionless) for plant leaves ranges from 300 to 800. The average value of 30 leaf readings is obtained in a few minutes, depending on the route in the field (the size of the field plot and the distance between successive plants measured).
Hydro N Tester Measurements in the Trials
The plant N status of each plot was assessed by chlorophyll meter measurements beginning 10 d after emergence at the earliest, and repeated four to six times during the season at intervals of 7 to 15 d. A single HNT measurement (30 readings on 30 individual leaves, one or two leaves being measured per plant) was made per plot on the same plant rows specifically dedicated to assessments during the growing season. Through time, the leaves measured were not the same because each time the first fully developed one at the top of the stem was used. The HNT value from plots with treatments of 70 or 75% of the N recommendation (the test N treatment) was called the test HNT value.
Harvest Evaluations in the Calibration Set of Trials
The potato plots were harvested mechanically after a growing period of 140 to 190 d after planting (Table 1). Two central rows per plot (at least 8 m long) were harvested to assess the effect of the N treatments on tuber yield and quality parameters in the four replications. The total tuber yield (in t ha–1) and the weight percentage of large tubers (>50 mm in diameter) were assessed. After harvest, the soil mineral N content was measured in each plot by taking samples on a 60-cm depth profile divided into four layers of 15 cm, according to the method of Guiot et al., 1992.
Harvest Evaluation in the Validation Set of Trials
In each of the four N treatments within the validation fields, four samples of tubers were harvested by hand, each one corresponding to a row 5 m long in 2001 and 4 m long in 2002. Total tuber yield and the weight percentage of large tubers were assessed for the entire sample. Dry matter content (as a percentage) was assessed on 3-kg subsamples of tubers by gravimetry in a water bath (weighing machine Robbe 9306, Torhout, Belgium). After harvest, the soil mineral N content was measured in each plot at a depth of 60 cm as described above. For the validation set, the statistical variable used was the deviation between the split N treatment and each other treatment (in the cases where a supplement was advised by the proposed method) or the deviation between the 70% N treatment and the 100% N treatment (in the case where no supplement was advised; Table 3).
|
The response indices are calculated with the following relations:
 |
 |
 |
The N treatment giving the highest response was determined from a comparison of N treatments (70, 100, or 130% N, with 100% N corresponding to the application rate recommended by AZOBIL). Indices were calculated from ratios of the highest response to the response in the control (zero-N treatment) for each trial and replication separately and the average of those ratios was taken. The highest data (yield, percentage in large tubers, and soil residual mineral N after harvest) give an idea about the potential use of N by the crop.
Validation of the Threshold Value (Validation Trials)
In the validation trials with cv. Bintje, the strategy to manage N was applying 70% of the N fertilizer recommendation at planting (test N rate) and, from 25 to 55 DAE, to link the decision to apply the remaining 30% to indications of crop N status given by the HNT measurements. The values were compared against a HNT threshold value. In the case of a positive decision, the remaining supplemental N was applied by the farmer either as NH4- or NaNO3 solid granules, or as a liquid foliar application in the form of diluted urea or N solution. Four parameters influenced by N were considered at harvest to assess the effect of supplemental N on the potato crop: total yield, percentage in large size tubers (>50 mm), residual mineral N in the soil after harvest, and tuber dry matter content. The effects of these parameters are summarized in Table 3. The validation of the decision to apply or not apply the remaining supplemental 30% N was based on the combination of all these effects.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
This method allows a distinction between sites responding strongly and sites responding weakly to N, the crops being considered as highly responding only if each of these three indices matches the chosen norms. Only at sites responding strongly to N was the application of supplemental N relevant.
Combining these results and comparing them with the fixed norms, seven sites were considered as presenting a globally relevant response to N fertilization (Table 4). According to this selection, two groups of sites were determined: a group of five sites with a weak crop response to N, including Sites 1, 5, 6, 7, and 11; and a group of seven sites with a strong crop response to N, including Sites 2, 3, 4, 8, 9, 10, and 12.
Table 4 also illustrates the usefulness of the conditioned N splitting approach. For example, for Sites 5 and 8, both cropped in 1999 in loam soil after an autumnal application of poultry manure, AZOBIL N recommendations that take these applications into account were respectively 150 and 125 kg N ha–1 based on specific information about the fields. But results at harvest showed that the Site 5 response to N was very poor and that, conversely, the Site 8 response was strong especially in terms of tuber size increase. We concluded that, at Site 5, N was not the limiting factor and that the optimum N rate was lower than the N recommendation. It was not the same at Site 8, where the optimum N rate seemed higher than the N recommendation. The soil N mineralization, being more uncertain with manure, was probably different between the two sites, which would explain a part of these results. To be effective, the decision support system based on chlorophyll meter values has to provide a distinction between Sites 5 and 8.
Absolute Values as a Decision Support System (Measurements in the Calibration Set)
For a single site and N treatment, the chlorophyll meter HNT values vary with the growth of the crop. For the test N treatment, the time courses were quite different from site to site. As sites were identified by the response of the crop to N and specified as sites with a weak crop response (Group 1) or a strong response (Group 2), Table 5 shows that the HNT values for the first group are not higher than the second group in the same year or across years. Consequently, the link between the time course of the raw HNT values and the crop N status at a site is not clear. Other factors than N status can affect the HNT values, as indicated above. The chlorophyll meter is not specific to N. These results show that raw HNT values do not really indicate the crop need for supplemental N.
|
|
Based on the HNT measurements taken from 20 to 60 DAE within each of the 12 trials of the calibration set, the comparison leads in several sites to a "significant" difference between control and test HNT values (Table 6). At most sites (2, 3, 4, 6, 7, 8, 9, and 11), only the control HNT values were significantly lower than the HNT values obtained within the fertilized plots. At Sites 1 and 5, there was no significant difference between N treatments. At Sites 10 and 12, mean HNT values in the zero-N treatment, test N treatment, and overfertilized treatment were significantly different. An example given in Fig. 1 illustrates the relation between HNT values and N rates applied at two sites: Site 5 belongs to the group of sites with a weak nonsignificant crop response to N, and Site 8 belongs to the group of sites with a strong crop response. The difference in HNT values between control and test N plots appears to be a good base to discriminate between sites where the potato crop responds strongly to N and sites where it does not; it should also serve as a base to indicate the need for supplemental N.
|
 |
|
|
|
 | [1] |
In this relation, the test HNT value is the average of at least two HNT values in the field plots with 70% of the recommended N rate (used to test the rates of N applications). Table 3 shows the evaluation a posteriori of the effect of a supplemental application of N beyond an application of 70% of the rate recommended by AZOBIL. The main effects are observed on the total yield and the proportion of large tubers (>50 mm). As these factors increase with increasing N, they are combined in Table 3 in one parameter, the yield of large tubers. For Sites 13 through 19 and 22, the parameters are compared between the split-N treatment (basal treatment corresponding to 70% of recommendation plus complement according to the HNT slope method) and the other N treatments and, for Sites 20 and 21, between the 100% N treatment and the treatments with lower N fertilization. The supplemental N affected the percentage in tuber dry matter only at two sites and does not increase the residual soil N after harvest. At Sites 14 and 16, the tuber dry matter percentage was significantly lower in the split treatment vs. the 70% treatment but that result does not change the evaluation of the decision because the values in the split treatment are good: 21.45 and 19.95%, respectively.
Among the 10 trials of the validation set with cv. Bintje, the method of determining a priori the relevance of supplemental N (is test HNT value – control HNT value > 0.5test N?) gave an affirmative answer in eight cases, i.e., Sites 13, 14, 15, 16, 17, 18, 19, and 22. In seven cases, the a posteriori evaluation (Table 3) indicated that the diagnosis with our method appeared to be correct, in Sites 13, 14, 15, 16, 18, 19, and 22 where the split-N treatment was more productive than either the 70 or 100% N treatments. A wrong decision regarding the need to apply supplemental N occurred only for Site 17, where the crop was irrigated. None of the other trials were irrigated. In the other two trials (Sites 20 and 21), the method had indicated no need for supplemental N application. In these trials, the evaluation a posteriori of the effect of no application of supplemental N indicated no negative effects.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A predictive model like the AZOBIL software is a good first approach to assess the mineral N fertilizer requirements of the crop, but it is a forward-looking and static method (Goffart and Olivier, 2003). In case of a soil with manure applied the previous fall, the soil mineralization during the growing season is uncertain and the N recommendation accuracy is reduced in comparison with a situation without manure. The second part of the strategy with the use of the chlorophyll meter is intended as an improvement that makes the method more dynamic and makes it possible to adapt the N fertilization to the crucial period of intensive N uptake by the crop (June–July).
This method tries to match N fertilizer supply and N crop demand throughout the growing season. Bijay-Singh et al. (2002) reported that, in rice and wheat, when N fertilizer application is not synchronized with crop demand, N losses from the soil–plant system can be large. Their need-based N management approach uses the chlorophyll meter to determine the right time for N topdressing of both crops. Chlorophyll-meter-based N management saves 12.5 to 25% of the fertilizer N recommendation, with no loss in yield. Peng and Cassman (1998) demonstrated that apparent recovery of topdressed N during the panicle initiation stage of rice could be as high as 78% and suggested that high N fertilizer recovery can be achieved with large rates of N applied when growth conditions are favorable.
In regard to potato, the results of this study show that the improvement of N efficiency should be also achieved by splitting N fertilizer applications and by monitoring the crop N needs to match crop N requirements and mineral N supply throughout the growing season.
The advantages of the HNT are its ease of use and its sensitivity and accuracy linked with the rapidity of measurement without destructive sampling.
A decision support system based on this tool, to be accurate, has to bypass the lack of specificity to the crop N status, as HNT readings are sensitive to other factors than N. In this context, the HNT slope was introduced as a new parameter, to characterize the difference between the chlorophyll meter HNT values in the fertilized field and in a zero-N subplot used as a reference in the field.
According to our results, the zero-N plot reference seems adequate to assess the N status of the potato crop. It leads to relative HNT threshold values useful for closely matching crop N demand and split N fertilizer applications. There are two other advantages to using a zero-N subplot as a reference plot within the field: first, it provides indirect information on the contribution of the soil mineral N supply to the crop N need during the growing season, and second, it gives the producer confidence in his soil potential to supply N when no or little difference appears between parts of the field with and without N application.
Meier et al. (2001) also proposed using this reference in vegetable crop fields, but with a simpler method without a diagnostic tool. The amount of topdressing is determined by the time it takes for the appearance and the intensity of N deficiency symptoms to become visible in the control plot compared with the rest of the field.
Mullen et al. (2003) compared two approaches with control subplots in cereals: using a well-fertilized treatment as reference (Varvel et al., 1997), called the sufficiency concept, and the control treatment approach (Raun et al., 2002), which they named the response concept using a zero-N subplot. Both methods make reference directly or indirectly to the N status of leaf tissues.
Using the sufficiency concept, N fertilizer is applied in an attempt to match the tissue N concentration of plants in a well-fertilized control strip (assumed to be 100% sufficient) without recognizing yield potential (Mullen et al., 2003). The response concept, like the HNT slope method proposed here, recognizes the yield potential and then fertilizes based on the likelihood of obtaining a response (Raun et al., 2002). The response concept indicates the increase in yield that could be obtained via N fertilization but provides no information about the N rate to be applied, whereas the sufficiency concept is directly linked to an actual fertilizer N rate (Mullen et al., 2003).
In the method proposed here, the determination of the potential of N recovery by the crop is preferred against achieving a sufficient rate of N to apply. Better than the sufficiency concept, the response concept using a control plot without N fertilizer allows the grower to know if N is the main limiting factor and therefore to assess the utility of a N supplement.
Relevant threshold values of HNT slopes to detect the need for supplemental N were determined using a calibration set of trials. The method was afterward validated in commercial fields, which were different from those used for calibration. There was only one error of diagnosis (in an irrigated situation). It must be noted that an error can be due to a number of factors: a wrong AZOBIL recommendation, an error in HNT measurement, an improper threshold, or a faulty N application. The validation is not able to distinguish between the various factors but it allows assessment of the final results of the N fertilization. Thus, the results validate the proposed method (the HNT slope method, a threshold value of 0.5, and a zero-N control plot) in crops without irrigation, using the cv. Bintje, on loam and sandy loam soils under a temperate climate. This represents most situations encountered in Belgium, where 
70 to 75% of the total potato crop is planted with cv. Bintje while <5% the total crop is irrigated. For other conditions (different cultivars, irrigated fields, other soils and regions, etc.), the validity of the method must be confirmed.
For situations outside the range of the Walloon area, care must be taken in extrapolating N recommendations. Nevertheless, we may give some hints for cases where the recommended N rate is 
100 kg N ha–1. In those cases, the potato grower is advised to apply a rate not less than 80 kg N ha–1 and to check with the chlorophyll meter that this fertilization is not too low, as described above. For recommendations >180 kg N ha–1, the suitability of the threshold value of 0.5 needs to be confirmed.
Our method proposes to split the N recommendation given by AZOBIL into 70 and 30%, with the second application determined by the HNT values. This split is based on two observations. First, with a first N application >70 to 75%, the results in terms of yield and tuber size with or without the second application are rarely different—"the game is not worth the candle." With a first N application <65 to 70%, there is a risk of not being able to make up a possible deficiency of N observed in the second part of the season.
Further improvement of the method can be expected with a refinement of the decisions concerning the proportion of the total N recommendation to be given in the second split application (less or more than 30% of the total N application recommended, according to HNT results). Also, the fertilizer forms for this second split application could be clarified. In fact, the success of the second N fertilizer application depends on the use of N fertilizer types and application modes that allow quick uptake by the crop after application. The spraying of repeated doses of foliar urea (maximum 15 kg N ha–1 dose) was used successfully in our trials, avoiding sunny conditions while spraying.
The complete strategy, consisting of splitting the N application with the second application determined by the use of the chlorophyll meter, has spread, since 2002, to potato producers and other partners in the potato industry.
During the 6 yr of research and validation, from 1997 to 2002, a total of 22 potato fields were managed according to the proposed strategy and checked after harvest (calibration and validation trials pooled together). At 14 sites, the best N management was the split application of the recommended rate, with the supplemental N application (30% remaining) applied when the N efficiency was high. This means that, in 64% of the cases, it was possible to improve N recovery with the same N rate as advised a priori but choosing the best application schedule. The other eight fields, where N added during the season was ineffective, were also instructive because they showed that, in 36% of the cases, 70% of the advised dose was sufficient to express the potential of the crop without causing loss of productivity.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
