Agronomy Journal 93:809-814 (2001)
© 2001 American Society of Agronomy
PRODUCTION PAPER
Critical Compositional Nutrient Indexes for Sweet Corn at Early Growth Stage
Lotfi Khiaria,
Léon-Étienne Parent*,a and
Nicolas Tremblayb
a Dep. of Soil Sci. and Agri-Food Eng., Laval Univ., Paul-Comtois Building, Sainte-Foy, QC, Canada G1K 7P4
b Hortic. Res. and Dev. Cent., 430 Gouin Blvd., St-Jean-sur-Richelieu, QC, Canada J3B 3E6
* Corresponding author (leon-etienne.parent{at}sga.ulaval.ca)
Received for publication August 10, 2000.
 |
ABSTRACT
|
|---|
It would be instrumental to define nutrient norms from small-size crop databases at the V4–V6 corn (Zea mays L.) growth stage for in-season N fertilizer recommendations. Our objective was to derive Compositional Nutrient Diagnosis (CND) and Diagnosis and Recommendation Integrated System (DRIS) nutrient index ranges from a sweet corn database and to relate nutrient concentration and indexes to ear yield. A sweet corn database of 240 observations on commercial yields and N, P, K, Ca, and Mg concentrations in 30-cm high corn seedlings were divided at random between survey (n = 200) and validation (n = 40) subpopulations. The proportion of low-yield specimens in the survey population was computed at inflection point of a cubic cumulative variance ratio function and was associated with a chi-square value (CND r2) of 3.9 that was confirmed in the validation subpopulation. Critical CND nutrient indexes were found to be symmetrical about zero as follows: -0.70 to +0.70 for N, -0.45 to +0.45 for P, -1.14 to +1.14 for K, -0.63 to +0.63 for Ca and Mg, and -1.05 to +1.05 for the residual filling value. Summing squared critical nutrient indexes also gave a CND r2 of 3.9, the minimum CND imbalance index for high-yield targets (>6.7 Mg ha-1). Nutrient concentration values were little to closely related to CND indexes (R2 = 0.34–0.87). The DRIS and CND indexes were highly related to each other (R2 = 0.91–0.99). For N at V4–V6 growth stage, the CND N index was the most closely related to ear yield.
Abbreviations: CND, Compositional Nutrient Diagnosis • CVA, Critical Value Approach • DRIS, Diagnosis and Recommendation Integrated System
|
INTRODUCTION
|
|---|
TISSUE NUTRIENT STATUS can be diagnosed by the Critical Value Approach (CVA) (Bates, 1971), the Diagnosis and Recommendation Integrated System (DRIS) (Walworth and Sumner, 1987), and Compositional Nutrient Diagnosis (CND) (Parent et al., 1994). Only DRIS and CND provide nutrient imbalance indexes although no threshold value has been validated yet for diagnostic purposes. The DRIS considers the numeric order of nutrient index values for diagnosing nutrient deficiencies (Walworth and Sumner, 1987). As a result, DRIS often diagnosed false nutrient deficiencies. Hallmark et al. (1987) proposed a DRIS dry matter index to separate limiting from nonlimiting nutrients. Savoy and Robinson (1990) corrected some defects in DRIS index interpretation by defining critical DRIS index ranges.
Khiari et al. (2001a) developed a mathematical procedure to separate low- and high-yield subpopulations in a crop survey. A yield cutoff value was obtained at an inflection point of a cubic cumulative variance ratio function of nutrient row-centered log ratios. A critical CND imbalance index was derived from the chi-square distribution function. A CND imbalance index was also formulated as the sum of squared CND indexes within the limits of a critical radius for the high-yield subpopulation. However, no critical CND index ranges were proposed. Because data acquisition is expensive, databases are often of limited size. A simple procedure is required to generate leaf nutrient norms from small-size survey databases if a large number of crops and cultivars are to be diagnosed at low cost.
Critical nutrient ranges could be established for V4–V6 stage for prescribing in-season fertilization because it would be too late for correcting a severe N deficiency at V8 stage (Varvel et al., 1997). Varvel et al. (1997) found that maximum corn yields were attained only when early season N levels (V8 stage) were adequate to maintain N sufficiency indexes between 90 and 100% as determined by chlorophyll meter readings.
The objective of this paper is to derive nutrient index ranges from a small-size sweet corn database at V4–V6 stage and to compare CVA, DRIS, and CND nutrient expressions for predicting ear yield.
|
MATERIALS AND METHODS
|
|---|
We examined the diagnostic value of CVA, CND, and DRIS at V4–V6 growth stage for using survey data. Data acquisition and methods of analysis were described by Khiari et al. (2001a). Briefly, 240 observations of commercial yields and five nutrient determinations were used from field surveys conducted during the 1995–1997 period in the Montérégie region, south of Montreal, QC, Canada. Fertilizers had been applied before seeding according to local recommendations, and no additional fertilizer was applied thereafter. Composite aboveground plant samples (10–15 subsamples randomly taken in 5 by 10 m plots) were obtained 3 to 4 wk after emergence (approx. 30 cm high) (Benton Jones et al., 1991). The plant tissues were oven-dried, ground, and wet-digested (Isaac and Johnson, 1976). The N and P concentrations were determined colorimetrically (TRAACS 800), and K, Ca, and Mg concentrations by plasma emission spectroscopy (ICAP 9000 from Jarrel Ash). Ear yields were collected in two 2-m-long rows in the central part of the plots.
No fertilizer trials for making in-season N recommendations were considered in this study. Varvel et al. (1997) found that sufficient N must be present from either residual soil N or fertilizer N application for early season growth of the corn because in-season fertilized crops cannot recover completely from early N deficiency. If the N sufficiency index was <90%, maximum yields were not achieved even with in-season N fertilizer applications; although yield was increased with in-season N fertilization, yield potentials had been reduced irreversibly at V8 stage in N deficient plants. We thus hypothesized that despite spring N applications, N sufficiency levels, N balance, and high yields could not be attained in some plots. If early N deficiency occurs, the most sensitive nutrient expression for early N diagnosis (concentration, DRIS index, or CND index) would be the one most closely related to final ear yield. Therefore, we would obtain a useful diagnostic tool at an earlier growth stage than above (V4–V6 stage instead of V8) for making in-season N recommendations.
Statistical Analysis
The DRIS norms and indexes were computed according to Walworth and Sumner (1987), and the CND norms and indexes via the computation steps of Khiari et al. (2001a). The Cate–Nelson ANOVA procedure (Nelson and Anderson, 1977) was used to partition yield data between two groups by maximizing the between-groups sums of squares to determine the threshold values for CND indexes required to compute the critical CND r2 value. We used 200 observations for developing the nutrient norms and 40 independent observations selected randomly in the database for validating the norms. The concentration, DRIS, and CND nutrient expressions were related to one another by regression analysis. All computations were made using Excel software (Microsoft, 1997).
|
RESULTS AND DISCUSSION
|
|---|
The Compositional Nutrient Diagnosis Norms for Simplex S5
Step 1
The S5, i.e., six-dimensional (d + 1) sweet corn simplex comprised the five nutrients N, P, K, Ca, and Mg and the filling value R5. Nutrient concentrations were transformed into CND row-centered log ratios VN, VP, VK, VCa, VMg, and VR5. The cutoff yield between the low- and high-yield subpopulations was determined after examining the five cubic cumulative variance ratio functions FCi
, FCi
, FCi
, FCi
, FCi
, and FCi
(Fig. 1 and Table 1). The yields (Mg ha-1) at inflection points of the cubic functions, computed by setting the second derivative of FCi
to zero, were 4.48 Mg ha-1 for FCi, 5.56 Mg ha-1 for FCi, 6.67 Mg ha-1 for FCi, 6.34 Mg ha-1 for FCi, 5.87 Mg ha-1 for FCi, and 5.89 Mg ha-1 for FCi. Highest cutoff yield was obtained with FCi (Table 1). At FCi yield cutoff, 61 of the 200 specimens, or 30.5% of the population, exceeded 6.67 Mg ha-1.
View this table:
[in this window]
[in a new window]
|
Table 1. Commercial yield of sweet corn at inflection points of the cumulative variance functions  for row-centered log ratios in the survey population (n = 200 obs.).
|
|
Step 2
The critical chi-square value for five nutrients plus the filling value has 6 df. The CND r2 values were distributed like chi-square values (R2 > 0.999, P < 0.001; data not shown). As found in Step 1, 69.5%, i.e., 100 minus 30.5%, of the population was below yield cutoff, and the corresponding chi-square value with 6 df was 3.9. The critical chi-square value of 3.9 is the maximum chi-square value for qualifying a sample in the high-yield subpopulation.
Step 3
The CND norms as means and standard deviations of VN, VP, VK, VCa, VMg, and VR5 values for the high-yield subpopulation (>6.67 Mg ha-1) are presented in Table 2. Nutrient indexes IN, IP, IK, ICa, IMg, and IR5 and CND r2 were computed as linearized, standardized variables (Khiari et al., 2001a). Using the Cate–Nelson partitioning procedure for the validation study, the relationship between CND r2 and ear yield for 40 independent specimens gave a yield cutoff value of 6.77 Mg ha-1 and a corresponding CND r2 value of 3.9 (Fig. 2). Hence, the independent population confirmed the yield cutoff value obtained at Step 1 and the critical CND r2 value obtained at Step 2.
View this table:
[in this window]
[in a new window]
|
Table 2. The Compositional Nutrient Diagnosis (CND) norms for k = 5 nutrients in a high-yield subpopulation producing more than 6.67 Mg ha-1 commercial yields of sweet corn.
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 2. Relationship between nutrient imbalance index Compositional Nutrient Diagnosis (CND) r2 and ear yield in S5.
|
|
Step 4. Nutrient Sufficiency Ranges
The DRIS norms for computing DRIS indexes are presented in Table 3. Nitrogen is the most common nutrient to be supplemented at early growth stage. Varvel et al. (1997) found that N deficiency at stage V8, as determined by chlorophyll meter readings, could not be fully corrected by later N fertilization. No earlier stage was examined. The relationship between raw tissue N concentrations and ear yield at V4–V6 stage showed no definite trend in the independent validation population (n = 40) (Fig. 3). However, the relationships between ear yield and DRIS or CND N indexes were significantly quartic (P < 0.05) (Fig. 3). As yield goal went higher, the critical N index range became narrower. The R2 value was larger for CND compared with DRIS. Hence, CND appeared to be the most sensitive diagnosis for early detection of N stress in sweet corn and could be instrumental in adjusting fertilization to crop needs after crop emergence.
View this table:
[in this window]
[in a new window]
|
Table 3. The Diagnosis and Recommendation Integrated System (DRIS) norms for dual ratios from five nutrients in a high-yield subpopulation producing more than 6.67 Mg ha-1 commercial yields of sweet corn.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3. Comparison between concentration, Diagnosis and Recommendation Integrated System (DRIS) and Compositional Nutrient Diagnosis (CND) relationships with ear yield for tissue N in the validation database.
|
|
Those results show that a N balance concept should be used for early diagnosis rather than a nutrient concentration approach. Presumably, the N status readings using the chlorophyll meter or other sensors of leaf N stress should be correlated with N leaf CND index rather than leaf N concentration alone because leaf greenness depends not only on N level, but also on levels of other nutrients.
To determine critical intervals, we partitioned ear yield into two groups using CND I2X for running the Cate–Nelson iterative procedure. The coefficients of determinations varied from 0.14 to 0.44 (Table 4). The squared critical CND indexes in the validation population (n = 40) were 0.50 for I2N, 0.20 for I2P, 1.30 for I2K, 0.40 for I2Ca, 0.40 for I2Mg, and 1.10 for I2R5. The sum of those additive unit-normal variables distributed like a chi-square value (Ross, 1987) was 3.9, the critical chi-square or CND r2 obtained at Steps 2 and 3.
View this table:
[in this window]
[in a new window]
|
Table 4. Critical Compositional Nutrient Diagnosis (CND) indexes and yield cutoff for k = 5 nutrients using the Cate–Nelson partitioning procedure.
|
|
Because squared critical CND indexes are additive to make a critical CND r2, the squared critical CND indexes could be used to define critical index ranges symmetrical about zero. The squared individual nutrient indexes must add to the critical CND r2 from theoretical considerations (Khiari et al., 2001a). The critical CND r2 value thus provides a bounded-sum constraint to the combination of nutrient ranges as defined by the Cate–Nelson procedure and can be used as a control procedure. This could not be obtained with DRIS.
As computed from the critical CND I2N value of 0.50 obtained at cutoff yield of 6.50 Mg ha-1 (Fig. 4), the critical CND IN range was -0.70 to 0.70. Similarly, the Cate–Nelson procedure indicated a critical value of 0.20 for I2p; thus, the critical Ip range was between -0.45 and 0.45 (Table 4). Other critical CND index ranges were -1.14 to 1.14 for IK, -0.63 to 0.63 for both ICa and IMg, and -1.05 to 1.05 for IR5 (Table 4). As shown in Table 4, the CND index sufficiency ranges differed among nutrients. For example, a CND index of -0.5 would be sufficient for N, K, Ca, and Mg but deficient for P. This contrasts with the DRIS concept that the more negative an index is, the more deficient the nutrient (Walworth and Sumner, 1987). Thus, a diagnosis based on CND or DRIS sufficiency ranges would be preferable to a classification of the nutrients in the order of their deficiency. Savoy and Robinson (1990) came to the same conclusion with DRIS diagnosis of white clover (Trifolium repens L.) using various rates of P and K in fertilizer experiments. Nevertheless, DRIS nutrient index ranges could not be defined from survey data using a generic model like the chi-square distribution function, as shown above for the CND nutrient index ranges.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 4. Relationship between the squared value of Compositional Nutrient Diagnosis (CND) N index and ear yield in S5.
|
|
A further validation step using fertilizer trials that encompass nutrient deficiency and excess is required to ascertain the symmetry and the validity of critical CND index ranges derived from a survey database. Such data were not available for sweet corn. A case study is presented for potato (Solanum tuberosum L. cv. Superior) in a companion paper (Khiari et al., 2001b).
Correlation between Diagnostic Approaches
The relationships between CND and either nutrient concentrations or DRIS indexes are presented in Table 5. Nutrient concentrations were not as closely related to CND indexes as were DRIS indexes. Large deviations from one of the coefficients of determination (R2) indicated that nutrient concentrations were distorted as other compositional data having a bounded-sum constraint to 100%. Those data required dual ratioing or multiratioing for constraint adjustment (Aitchison, 1986). Hence, the DRIS and the CND indexes were adjusted indexes closely related to each other (Table 5).
View this table:
[in this window]
[in a new window]
|
Table 5. Equations relating Y as CND index to X as nutrient concentration or Diagnosis and Recommendation Integrated System (DRIS) index including or not including the filling value R5.
|
|
Small deviations from one of R2 values indicated that DRIS indexes were still somewhat distorted. A possible explanation was that an expression for the dry matter content was not included. We thus computed DRIS norms including R5 rather than the dry matter content (Hallmark et al., 1987) because the dry matter content includes, by definition, nutrient concentrations already defined in the tissue simplex (Table 3). The coefficient of determination was improved for N, P, and K expressions and for the nutrient imbalance index as well (Table 5). The DRIS nutrient imbalance indexes, computed by summing up DRIS nutrient indexes irrespective of sign, showed a quadratic relationship with CND r2 due to two highly unbalanced specimens (Fig. 5). The great similarity between CND and DRIS including R5 supports the view of Savoy and Robinson (1990) that DRIS index ranges must be symmetrical about the mean. However, Sumner (1979) did not obtain symmetrical DRIS index ranges for corn.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 5. Relationship between Compositional Nutrient Diagnosis (CND) r2 and Diagnosis and Recommendation Integrated System (DRIS) nutrient imbalance indexes (NII).
|
|
Due to its unique characteristic as a chi-square variable, the CND r2 imbalance index is amenable to CND index ranges defined on a theoretical basis. Additive critical CND I2X values mean that critical IX ranges between -IX and +IX could be obtained simply by computing the square root of I2X. Should CND index ranges be ascertained by a limited number of fertilizer trials, CND would provide a simple and cost-effective tool for diagnosing nutrient imbalance in a large number of specific soil–plant systems with small databases.
|
CONCLUSION
|
|---|
The mathematical approach of Khiari et al. (2001a) was expanded to five nutrients to generate CND nutrient norms for sweet corn at early growth stage. In keeping with the chi-square distribution, the critical CND r2 value was 3.9. Except for Ca, the leaf nutrient concentrations were not as closely correlated with CND nutrient indexes as DRIS indexes. Including the filling value R5 as a DRIS variable, the relationships between DRIS and CND indexes were improved for N, P, K, and the nutrient imbalance index. Because squared values of CND indexes are additive to make CND r2, the CND indexes must show nutrient ranges symmetrical about zero. Critical CND index ranges were found to be ±0.70 for IN, ±0.45 for IP, ±1.14 for IK, ±0.63 for ICa, ±0.63 for IMg, and ±1.05 for IR5. Such ranges could not be obtained with DRIS because DRIS indexes are not compatible with the chi-square distribution function. A further validation step using fertilizer trials that encompass nutrient deficiency and excess is required to ascertain the symmetry of critical CND index ranges.
|
ACKNOWLEDGMENTS
|
|---|
We are grateful to the Fédération québécoise des producteurs de fruits et légumes de transformation, the Québec Food Processors Association, the Matching Investment Initiative Program of Agriculture and Agri-Food Canada, and the Natural Science and Engineering Research Council of Canada for financial support. The technical assistance of Yvon Perron and Marcel Tétreault is gratefully acknowledged.
|
REFERENCES
|
|---|
- Aitchison, J. 1986. Statistical analysis of compositional data. Chapman and Hall, New York, NY.
- Bates, T.E. 1971. Factors affecting critical nutrient concentrations in plant and their evaluation: A review. Soil Sci. 112:116–130.
- Benton Jones, Jr., J., B. Wolf, and H.A. Mills. 1991. Plant analysis handbook. A practical sampling, preparation, analysis, and interpretation guide. Micro-Macro Publ., Athens, GA.
- Hallmark, W.B., J.L. Walworth, M.E. Sumner, C.J. deMooy, J. Pesek, and K.P. Shao. 1987. Separating limiting from non-limiting nutrients. J. Plant Nutr. 10:1381–1390.
- Isaac, R., and W. Johnson. 1976. Determination of total nitrogen in plant tissue, using a block digestor. J. AOAC 59:98–100.
- Khiari, L., L.E. Parent, and N. Tremblay. 2001a. Selecting the high-yield subpopulation for diagnosing nutrient imbalance in crops. Agron. J. 93:802–808 (this issue).[Abstract/Free Full Text]
- Khiari, L., L.E. Parent, and N. Tremblay. 2001b. The phosphorus compositional nutrient diagnosis range for potato. Agron. J. 93:815–819 (this issue).[Abstract/Free Full Text]
- Microsoft. 1997. Microsoft Excel 97. Incline Village, NV.
- Nelson, L.A., and R.L. Anderson. 1977. Partitioning of soil test-crop response probability. p. 19–38. In M. Stelly (ed.) Soil testing: Correlating and interpreting the analytical results. ASA Spec. Publ. 29. ASA, Madison, WI.
- Parent, L.E., A.N. Cambouris, and A. Muhawenimana. 1994. Multivariate diagnosis of nutrient imbalance in potato crops. Soil Sci. Soc. Am. J. 58:1432–1438.[Abstract/Free Full Text]
- Ross, S.M. 1987. Introduction to probability and statistics for engineers and scientists. John Wiley & Sons, New York, NY.
- Savoy, H.J., Jr., and D.L. Robinson. 1990. Norm range size effects in calculating Diagnosis and Recommendation Integrated System indexes. Agron. J. 82:592–596.[Abstract/Free Full Text]
- Sumner, M.E. 1979. Interpretation of foliar analyses for diagnostic purposes. Agron. J. 71:348.[Abstract/Free Full Text]
- Varvel, G.E., J.S. Schepers, and D.D. Francis. 1997. Ability for in-season correction of nitrogen deficiency in corn using chlorophyll meters. Soil Sci. Soc. Am. J. 61:1233–1239.[Abstract/Free Full Text]
- Walworth, J.L., and M.E. Sumner. 1987. The Diagnosis and Recommendation Integrated System (DRIS). Adv. Soil Sci. 6:149–188.
This article has been cited by other articles:
|
|
|
|
 
M-C. Belanger, A. A. Viau, G. Samson, and M. Chamberland
Determination of a Multivariate Indicator of Nitrogen Imbalance (MINI) in Potato Using Reflectance and Fluorescence Spectroscopy
Agron. J.,
October 19, 2005;
97(6):
1515 - 1523.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Khiari, L.-E. Parent, and N. Tremblay
The Phosphorus Compositional Nutrient Diagnosis Range for Potato
Agron. J.,
July 1, 2001;
93(4):
815 - 819.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
