| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Inst. of Crop Sci. and Plant Breeding, Grass and Forage Sci./Organic Agric., Christian-Albrechts-Univ. Kiel, Olshausenstr. 40, D-24098 Kiel, Germany
* Corresponding author (aherrmann{at}email.uni-kiel.de)
Received for publication February 19, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CNC, critical nitrogen concentration at silage maturity • DM, dry matter • LUFA, Agricultural Analytical and Research Institute • Nc, nitrogen concentration • Ncth, threshold nitrogen concentration • Nmin, soil mineral nitrogen • NNI, nitrogen nutrition index • W, dry matter yield • Wrel, relative dry matter yield
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nonpoint N emissions show maximum values in regions where high livestock population density coincides with sites vulnerable to leaching. This situation applies especially to northwestern Germany, which is characterized by sandy soils, shallow groundwater tables, and by a concentration of dairy farming with intensive silage maize production. Previous and current efforts on the national and state level to reduce N emissions to the environment include various measures, such as the obligation under the Fertilizer Act to record nutrient balances at the farm gate or field level and enacted policies such as the large-scale monitoring of environmental conditions (Bundesregierung, 1996). In the future, the allocation of subsidies to producers will be linked to the respect of environmental, food safety, animal and plant health, and animal welfare standards as well as the requirement to keep all farmland in good agricultural and environmental condition ("cross-compliance") (Karnitschnig, 2002; European Commission, 2004). The successful implementation of such subsidy allocation policies requires the availability of efficient and reliable indicators of sound nutrient management. With respect to the N status of crops, a favorable indicator should be (i) sensitive to N deficiency as well as overfertilization, (ii) easy to determine and to update for in-season adjustments of N management at the farm level, and (iii) suitable for broad-scale monitoring and assessment programs.
Many studies have been conducted to explore the potential of plant tissue sampling methods as an indicator of N status. The chlorophyll meter technology (Piekielek et al., 1995; Fox et al., 2001) provides point measurements only, which limits its suitability for quantifying the N status of entire fields or for a broad-scale monitoring (Bausch and Duke, 1996). Moreover, this method is not capable of detecting luxury N uptake since maize plants achieve a maximum chlorophyll content irrespective of the level of overfertilization (Dwyer et al., 1995). This serious disadvantage applies also to remote-sensing techniques, which allow a large-area monitoring with the possibility of assessing spatial variability within a field (Bausch and Duke, 1996; Blackmer and Schepers, 1996; Osborne et al., 2002). The end-of-season stalk nitrate test has proven effective for the assessment of the N status of maize from one-fourth milkline growth stage to 3 wk after maturity (Binford et al., 1992; Hooker and Morris, 1999). The concept of a critical N concentration provides another indicator, which can be applied over a wider time frame. Plénet and Lemaire (1999) suggested restricting its validity to the period from emergence to silking plus 25 d. Herrmann and Taube (2004) recently showed that this indicator is valid until silage maturity. The underlying concept assumes the existence of a minimal N concentration necessary to achieve a maximum crop growth rate. This required minimal concentration is defined as the critical N concentration. The current N status at any time can now be expressed by the N nutrition index (NNI), defined as the ratio of actual N over the corresponding critical N concentration value (Lemaire and Gastal, 1997). The end-of-season stalk nitrate test as well as the critical N concentration are well suited to record the N status and to distinguish N deficiency from N excess situations. Both methods, however, require a considerable effort of sampling and processing, which for all practical purposes prohibits their routine use as a regional monitoring tool.
In Germany, most dairy farmers routinely make use of the services of an existing network of private and official consultancy agencies to conduct a forage quality analysis of their silages. The data on N concentrations of maize silages are therefore widely available to them. Provided that the changes of N concentration during ensiling are negligible, this information could be used by farmers to calculate their individual NNI, and based on this value, they could adjust their N management in the following growing season ("learning-by-doing" strategy). Furthermore, policy makers could include this extensive data set into their effort of large-scale monitoring of the overall N status of silage maize production. There is, however, one obstacle to this otherwise promising approach, namely the necessity of recording DM yield parallel to N concentration.
The main objectives of the present study therefore were (i) to circumvent the necessity of yield recording by quantifying the relationship between relative DM yield and N concentration of the whole crop at silage maturity, based on data sets collected in nine experiments on sandy soils in northern Germany; (ii) to derive a CNC value, defined as the critical N concentration at silage maturity, i.e., the minimal N concentration at maximum relative yield; (iii) to monitor and assess the status of N management in silage maize production in northern Germany by applying the derived CNC value to silage quality data provided by several Agricultural Analytical and Research Institutes (LUFA); and (iv) to discuss the suitability of this new approach.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Fertilizer treatments included applications of mineral N fertilizer or combinations of mineral N and slurry (cattle or pig). With respect to mineral N application, fixed treatments were used in Exp. 1 to 3, regardless of the Nmin content. In contrast, fertilization of Exp. 4 to 9 was aligned with a fixed-target N fertilization level, taking into account the soil N release, i.e., Nmin in spring. With Nmin values ranging from 4 to 81 kg N ha–1, actual application rates varied substantially. Slurry was generally applied before sowing and mineral N fertilizer given in split doses at key growth stages, i.e., at the one- and six-leaf stage if not specified differently in Table 1. Phosphorus and K fertilization levels and further crop management measures were applied according to the common agricultural practice to allow for potential production, i.e., no other factor was limiting except N. In field experiments that included slurry treatments, P and K application were adjusted for all treatments to align with the highest slurry application rate to eliminate potential effects of P and K supply on maize yield and quality. Cultivars used in the experiments covered the range of early to midearly silage maturity. With only a few exceptions, the field trials did not receive any irrigation. Experiments were generally conducted in a randomized complete block design or a split-plot design with three to four replications.
Data Sampling
In Exp. 1, maize yield and quality data were recorded every 2 wk throughout the growing season. The present study, however, used only the data collected at silage maturity, which was assessed on the basis of DM content of the whole crop, optimally ranging between 300 and 350 g kg–1, depending on genotype. In Exp. 2 to 9, data on maize yield and forage quality were recorded at silage maturity stage only. In all experiments, aboveground biomass sampling was performed using a plot chopper, and representative subsamples were subsequently dried to constant weight at 65°C. All samples were fine-ground using a Cyclotec mill, which was fitted with a 1-mm screen. Nitrogen content of the samples was determined using the near infrared spectroscopy technology (NIRSystems 5000 monochromator, Foss-NIRSystems, Silver Spring, MD, USA), using software from Infrasoft International (ISI, Port Matilda, PA, USA). Calibration and validation subsets were analyzed for N with the Kjeldahl method according to the procedure of Naumann and Bassler (1993).
Selection of Data Sets
Our study intended to first derive a relationship between the standardized, relative DM yield and the N concentration of forage maize at silage maturity. This should then allow determination of the CNC value, i.e., the minimum N concentration required for achieving maximum relative DM yield. Only site-year combinations covering the whole range of N supply from N deficiency conditions up to luxury N consumption are suitable for determining the relationship between relative DM yield and N concentration. We therefore, in a first step, separated the informative from the noninformative data sets. This was achieved by fitting to each site-year combination a quadratic-plateau function, which described the response of DM yield to N concentration.
 | [1] |
While the quadratic part represents those situations where further increase in N supply is paralleled by an increase in DM yield, the plateau part of the function describes the non-N-limiting growth conditions, i.e., additional N supply results in an increased N concentration but no increase of biomass. Estimates of the function parameters were calculated for each site-year combination using PROC NLIN of SAS (version 8.2). With respect to our parameter estimation results, three cases may be distinguished: (i) parameters could not be estimated for the quadratic and plateau parts; (ii) the quadratic part could be fitted, but no DM yield maximum (i.e., the plateau part) was attained; and (iii) the complete quadratic-plateau function with all parameters, namely Ncth, a0, a1, and a2, and wmax, could be estimated. For our purposes, all data sets (site-year combinations) belonging to the first or second group are noninformative since for our subsequent analysis, we need to know the DM yield-to-N-concentration relation in the responsive part of that relationship and at the same time we need to know the threshold value Ncth, where this responsive part ends. All noninformative data sets were therefore discarded from further analysis.
The remaining informative site-year combinations constituted the data pool for the determination of the relative DM yield-to-N-concentration relationship and the derivation of a CNC value. Before pooling these data sets, however, the DM yields had to be standardized since the yield potential may differ substantially between sites or years. For each site-year combination, the estimated DM yield wmax = W(Ncth) of the corresponding plateau part was assumed to be the maximum yield. Each DM yield value W(Nc) of a data set therefore was standardized with respect to the corresponding wmax value, i.e., it was converted into a relative yield Wrel(Nc) = 100·W(Nc)/wmax.
Derivation of the Critical Nitrogen Concentration Value
After standardization of the DM yields in all informative data sets, these data were pooled to quantify the relationship between relative DM yield and N concentration. For the purpose of finding that relationship, however, only those data points belonging to the quadratic (responsive) parts are essential. Hence, all data points with N concentrations above their corresponding threshold value Ncth were discarded before the pooling of the data.
For fixed conditions of soil and climate, one may express the relationship between relative DM yield and N concentration as an exponential function of the form
 | [2] |
To include soil and climatic conditions into the model, we regarded site and year as good representations of soil and climatic conditions. Rather than entering them as two separate variables, we only distinguished different site-year combinations. This was necessary because the available experimental data were not cross-classified, i.e., trials on different sites were not conducted over all years; also, the plots at some sites changed over years. With the site-year effects random, the resulting model equation was
 | [3] |
+ ai) + (ß + bi)·Wrelij + eij. Using PROC MIXED in SAS, we obtained estimates for all parameters involved. The desired CNC value, defined as the N concentration Nc(Wrel) at maximum relative DM yield, is then obtained by setting Wrel = 100 in the exponential function of Eq. [2]. |
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
For the 15 (out of 29) remaining informative data sets, the parameters a0, a1, and a2, and wmax of the quadratic as well as the plateau part, could be estimated according to the response function W(Nc), as introduced in Eq. [1]. Figure 2 displays the calculated threshold values Ncth and the corresponding estimated DM yield values wmax = W(Ncth) for all 15 site-year combinations. Estimated maximum yields (plateau values), which ranged between 10.6 and 18.3 t ha–1, indicate that the yield potential differed considerably among the 15 site-year combinations. They were, however, in accordance with findings of Schröder (1999) and Ogola et al. (2002) for comparable soils and climatic conditions.
Nitrogen Concentration as a Function of Dry Matter Yield
The data of all informative sets contained a total of 147 pairs of (Wrel, Nc), 45 of these had Nc values greater than their corresponding Ncth. Discarding them left us a pooled set of 102 remaining points for conducting a meta-analysis using a random coefficient model (Eq. [3]).
A statistically meaningful relationship between relative DM yield and N concentration at silage maturity could be estimated for the remaining 102 data points (Fig. 3). While the model fits satisfactorily in the upper yield range, a systematic deviation between observed and predicted values is apparent in the lower range, i.e., for relative DM yield values below 70%. This is mainly due to the imbalanced data structure. In Berkhof, Dasselsbruch, and Bremervoerde, the DM yields of the control treatments (no N application) were consistently high compared with the corresponding maximum yield (plateau value), exceeding 73% of the latter. This effect is most likely caused by specific environmental conditions of these sites, e.g., soil type, mineralization capacity, manuring history, crop rotation, and climate. In contrast, for the Schuby, Ostenfeld, and Karkendamm experiments, DM yield of the control treatments in most site-year combinations fell substantially below 70% of the corresponding plateau values.
|
With decreasing N input, i.e., in the lower yield level range, the N content of the crop is assumed to approximate a minimum value. The study by Plénet and Cruz (1997) specified a minimum N concentration of 7 g N kg–1 DM for maize, which is in agreement with the value of 8 g N kg–1 DM suggested by Lemaire and Gastal (1997) for structural plant N concentration. In the present study, the minimum N contents estimated (5.7 g N kg–1 DM) and observed (6.4 g N kg–1 DM) for a relative DM yield of 30% were lower (Fig. 3). CERES-Maize (Jones and Kiniry, 1986) specifies a minimum N concentration of 4.5 g N kg–1 DM from silking until physiological maturity, which however, refers to the stover. This value is in agreement with the stover data of the Karkendamm experiment, ranging between 2.1 and 4.7 g N kg–1 DM for the zero-fertilization treatment.
Critical Nitrogen Concentration at Silage Maturity
For our primary objective of quantifying the CNC constant, i.e., the CNC, only the quality of the model fit in the upper range of relative DM yield is of importance. While lower relative DM yield levels showed considerable deviations between observed and predicted N concentrations, our model performed sufficiently well in the crucial range of relative DM yield values above 70% (Fig. 3). It was therefore possible to calculate a CNC value for silage maize. Setting Wrel = 100, we obtain a CNC constant of 10.5 g N kg–1 DM.
We compared this result with Plénet and Lemaire (1999). Their study provides two regression functions describing the relationship between DM yield and critical N concentration. The first function covers the period from the 10-leaf stage up to silking plus 25 d (Table 2). Extrapolating this regression line to silage maturity by assuming a DM yield of 24 to 25 t DM ha–1, one obtains a CNC value of 10.4 to 10.5 g N kg–1 DM, which is practically identical to our estimations. The second regression line in Plénet and Lemaire (1999) spans from 10-leaf stage up to silage maturity (Table 2), but the authors doubt the validity for later grain-filling stages. For this regression, the CNC was 10.2 to 10.4 g N kg–1 DM, which again coincides quite well with our analysis.
|
Some nonresponsive sites, e.g., Markhausen, Bramsted, and Celle, showed high DM yields and N contents even without N fertilization and were excluded from further analysis. Most such sites have a long history of high N input and manuring, which is likely to cause a high mineralization potential, especially in combination with humous soils. These characteristics are often encountered in maize-growing regions of northwestern Germany. Applying the CNC constant to nonresponsive data sets indicated an overfertilization of the crop (Fig. 2), which seems to be a quite realistic assessment of the actual N status.
The development of the CNC indicator was based on a Kjeldahl method for N determination (Naumann and Bassler, 1993). We chose this method despite its limitations (N in N–O linkages is only partially determined) because it is the standard method for N determination of forage crops used by German LUFA institutes. Applying this method throughout the whole study should provide consistent results. A CNC determination based on total N content may give a slightly different CNC constant.
Assessment of Critical Nitrogen Concentration as Indicator of Nitrogen Status
The CNC allows, on the one hand, for detection of luxury N consumption if the N concentration exceeds the calculated critical threshold. On the other hand, the derived functional relationship enables, in case of insufficient N supply, the quantification of relative yield losses. The CNC value thus provides a diagnostic tool to assess the N status of the maize crop that can guide maize growers in adjusting their N fertilization for the following growing seasons in terms of a "learning-by-doing" strategy. The results of this study are in good agreement with those of Plénet and Lemaire (1999), which were based on different genotypes and environmental conditions. The derived relationship may therefore be generalized to other regions with climatic and soil conditions similar to northwestern Europe, as long as crop management with respect to tillage and row spacing is compatible. It remains, however, to be clarified to what extent the relationship between N concentration and DM yield is affected by other management practices, such as tillage or understory crops, since findings in literature are ambiguous (Cusicanqui and Lauer, 1999; Mehdi et al., 1999; Cox and Cherney, 2001 and 2002; Widdicombe and Thelen, 2002; Nevens and Reheul, 2003).
Apart from assessing the N status on the farm scale, one could exploit the CNC value on a large scale to monitor the "environmental performance" of farmers in a whole region. At least in Germany, the necessary data on N concentration of maize silage are routinely gathered and therefore available. This approach assumes that no substantial N losses occur during the ensiling process. Since maize is characterized by a relatively high DM content, a low buffering capacity, and adequate levels of water-soluble carbohydrates, we can exclude any extensive proteolysis during ensiling (McDonald et al., 1991). In theory, one might even argue that the N concentration increases due to DM losses caused by respiration and fermentation. Studies on changes of N concentration during ensiling of maize are scarce. Bergen et al. (1974) reported a slight increase of N concentration during ensiling of the whole maize plant. The study of Amos et al. (1996), however, found a negligible decrease of N content in maize silage compared with the unfermented forage.
Evaluation of the Nitrogen Status of Silage Maize Production in Northern Germany
Farmers in Germany routinely send silage samples for forage quality analysis to a nationwide network of institutes (LUFA institutes), which have a unified and coordinated approach toward analysis and advisory. Based on forage quality data provided by three of the LUFA institutes (Kiel, Hameln, and Kassel), the derived CNC value of 10.5 g N kg–1 DM was used to investigate the current state of N supply in Northern Germany's forage maize production. The LUFA institutes are located in different federal states of northern Germany, each having a large catchment area. The data thus can be assumed to be a representative sample on the N status of forage maize production in the northern part of Germany.
The data provided by the LUFA institutes show a clear tendency for all locations and years for the N concentrations to exceed the CNC value of 10.5 g N kg–1 DM in most samples by more than one standard error (1.03 g N kg–1 DM), indicating for that region a predominance for luxury crop N uptake (Table 3). In the region of the LUFA Kassel, where 4 yr of data were available, a steady decrease from 98 to 61% of samples with luxury N uptake could be observed. The pronounced decrease in the Kassel distribution is most likely due to a reduction of N input, resulting from strict measures of groundwater protection. In the region of LUFA Hameln, where only 2 yr were analyzed, we find the opposite development with an increasing percentage from 76 to 95%. The increase of N concentrations in the Hameln data may be attributed to several reasons. The lower mean DM content, probably a consequence of delayed crop development, may have led to higher N contents in 2000. Also, water shortage or possibly higher N input may have contributed. The Kiel data for the year 2000 showed a similar excess of N fertilization with 71% of the samples above the CNC value plus a standard error. Unfortunately, for the 1999 Kiel data, only frequencies were available with classes differing from those used in our analysis. But the data exhibit the same luxury N consumption.
|
Maize is quite dependent on temperature and sufficient irradiation. Since northern Germany is a marginal region, where the conditions for successful forage maize production are not always met, the stage of maturation at harvest varies strongly depending on the prevailing weather conditions. Unfavorable conditions in summer and early autumn may delay plant development and therefore lead to suboptimal results at harvest time with lower DM content and higher N concentrations. To determine to what extent weather conditions influenced the N concentration, we compared DM contents with corresponding N concentrations for each site-year combination and observed only a weak negative correlation, if any, in some site-year data sets (Fig. 4 and Table 3). Calculated over all years for a given site, the corresponding correlations were slightly negative (Hameln: r = –0.38) or practically absent (Kassel r = –0.23). The low r values can be regarded as negligible, indicating that N management, not weather, was the important factor for the crop's N content.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
s manual (version 3.0). Biol. Syst. Eng. Dep., Washington State Univ., Pullman.This article has been cited by other articles:
|
|
 |
|
  Z. Cui, F. Zhang, Z. Dou, M. Yuxin, Q. Sun, X. Chen, J. Li, Y. Ye, Z. Yang, Q. Zhang, et al. Regional Evaluation of Critical Nitrogen Concentrations in Winter Wheat Production of the North China Plain Agron. J., January 8, 2009; 101(1): 159 - 166. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
