Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published in Agron J 100:543-550 (2008)
DOI: 10.2134/agronj2006.0153
© 2008 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhang, J.
Right arrow Articles by Koehler, K. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Zhang, J.
Right arrow Articles by Koehler, K. J.
Agricola
Right arrow Articles by Zhang, J.
Right arrow Articles by Koehler, K. J.
Related Collections
Right arrow Nitrogen
Right arrow Field-Scale Studies
Right arrow Maize
Right arrow Field evaluation techniques
Right arrow Nutrient Management
Right arrow Site-Specific Analysis
Right arrow Soil Fertility and Productivity
Right arrow Statistics

NITROGEN MANAGEMENT

Sensitivity of Chlorophyll Meters for Diagnosing Nitrogen Deficiencies of Corn in Production Agriculture

Jun Zhanga,*, Alfred M. Blackmerb, Jason W. Ellsworthc and Kenneth J. Koehlerd

a Statistical Consulting Center, Wright State Univ., 130 MM Bldg., 3640 Colonel Glenn Hwy., Dayton, OH 45435
b in memory, Dep. of Agronomy, Iowa State Univ., Ames, IA 50011
c Wilbur Ellis Company, 150 Burlington St., Pasco, WA 99301
d Dep. of Statistics, Iowa State Univ., Ames, IA 50011

* Corresponding author (jun.zhang{at}wright.edu).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Chlorophyll meters have been widely used to diagnose N deficiencies in research trials and to help refine N fertilizer recommendations in production agriculture. The diagnoses are based on the assumption that chlorophyll meters can detect a wide range of N deficiencies under field conditions according to relationships between chlorophyll meter readings (CMRs) and yield responses to N fertilization that are established from small-plot trials. We conducted a field-scale study to evaluate the sensitivity of chlorophyll meters when they are used to detect N deficiencies in corn (Zea mays L.) production. The multiple year–location data were gathered in three cornfields where N fertilizer was applied at different rates in strips that crossed several soil map units. Grain yields were related to CMRs taken on corn leaves in June, July, and August of 1998–1999. Results showed that the chlorophyll meters could detect severe N deficiencies early in the season, but small and moderate deficiencies of N could not be diagnosed with reasonable certainty until it was too late to make in-season fertilization. Two types of discontinuity are discussed in this paper as likely causes for the limited sensitivity of the chlorophyll meters. Our study suggests that problems associated with the use of chlorophyll meters to diagnose deficiencies of N in production agriculture are much greater than their use to help interpret results of well-controlled research trials where yields are also measured.

Abbreviations: CMR, chlorophyll meter reading


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Received for publication May 16, 2006.
   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
CHLOROPHYLL METERS are relatively recent innovations that can be used to monitor deficiency symptoms of N during the growth of crops (Piekielek and Fox, 1992; Schepers et al., 1992ab; Varvel et al., 1997; Zhang and Blackmer, 1999; Fox et al., 2001). The hand-held chlorophyll meters clamp onto leaves and measure light transmittance through a small test area of 2x3 mm2 on leaves. This device generates two specific wavelengths (650 nm for maximum absorption by chlorophyll and 940 nm for maximum transmittance) that are selected for determination of chlorophyll contents and the output is recorded as a CMR, a dimensionless unit. Effective use of the chlorophyll meters draws on many of the basic principles used to diagnose N deficiencies by analyzing plant tissues, but the chlorophyll meters offer the advantage of in-field nondestructive and quick diagnoses (Ortuzar-Iragorri et al., 2005). The basic principles of using plant analyses to diagnose nutrient deficiencies are discussed by Macy (1936), Tyner and Webb (1946), and Black (1992). In this case, nutrient deficiency is usually defined in terms of the plant's response to the added nutrient. A plant is deficient in a certain element if the supplying of that element to the plant in a suitable form causes an increase in yield (Goodall and Gregory, 1947).

Diagnoses of N deficiencies made by using the chlorophyll meters may have several applications (Blackmer and Schepers, 1995). For example, they can help interpret the results of N-response trials because measurements of leaf chlorophyll content and responses of yield to applied N provide independent assessments of the sufficiency (supply relative to demand) of N for plant growth (Piekielek and Fox, 1992). The diagnoses of N deficiencies can help to make in-season predictions of yield losses and to estimate amounts of fertilizer immediately needed to avoid these losses in production agriculture (Piekielek and Fox, 1992; Blackmer and Schepers, 1995; Scharf et al., 2002). Such predictions can be made because symptoms of N deficiencies are expressed much earlier in leaf chlorophyll content than in grain yield (Peterson et al., 1993). The use of the chlorophyll meters in this manner requires that diagnoses be based on previously established relationships between CMRs and yield responses to fertilizer N. The diagnoses of N deficiencies by using chlorophyll meters can also involve assessing N sufficiency levels in production agriculture to learn how recommendations of N fertilizer can be refined in future years (Piekielek et al., 1995; Varvel et al., 1997; Fox et al., 2001; Scharf et al., 2006; Zhang et al., 2008a). This application essentially uses the chlorophyll meters as an alternative to yield response trials. The chlorophyll meters offer an attractive alternative because it is much easier to assess N sufficiency by taking a few readings than by establishing plots, applying fertilizer treatments, and measuring yield responses to the fertilizer. Little attention has been given to estimating the sensitivity of diagnoses for N deficiencies, that is, the small deficiency of N that can be detected with reasonable certainty when chlorophyll meters are used in production agriculture. Such a concept of sensitivity is comparable to the sensitivity analysis which deals with how overall uncertainties in a model output (e.g., yield response) can be quantitatively apportioned to variation of model inputs (Saltelli et al., 2004). The inputs were chosen as CMRs in our study. Poor sensitivity of chlorophyll meter measurements can be expected because CMRs are closely linked to the concentrations of N in corn leaves before tassel emergence (Schepers et al., 1992a; Wood et al., 1992, 1993; Markwell et al., 1995; Waskom et al., 1996) and because tissue tests based on leaf N concentration have poor sensitivity for diagnosing N deficiencies (Cerrato and Blackmer, 1991). The underlying problem is the difficulty to determine "the critical concentration", that is, the concentration that distinguishes plants having deficient N from plants having adequate N. Such a critical concentration should be clearly defined and relatively constant across sites and years. This paper describes field-scale studies to evaluate the sensitivity of chlorophyll meters for diagnosing N deficiencies in production agriculture. Specific objectives include (i) illustrating symptoms of limited sensitivity of chlorophyll meters when data from N-response trials are presented, and (ii) enumerating some likely causes of the limited sensitivity.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experimental Design
Experiments were established in 1998–1999 at Sites 1 and 3 in Ogden, Boone County and at Site 2 in Grand Junction, Greene County, Iowa. Major soil associations include Canisteo (fine-loamy, mixed, superactive, calcareous, mesic Typic Endoaquolls), Clarion (fine-loamy, mixed, superactive, mesic Typic Hapludolls), Webster (fine-loamy, mixed, superactive, mesic Typic Endoaquolls), and Nicollet (fine-loamy, mixed, superactive, mesic Aquic Hapludolls). The fields were planted to corn following soybean (Glycine max L. Merr.), a popular cropping system used in production agriculture of the Corn Belt. Except for N fertilizer treatments, all fields were managed by farmers using their own practices. The sites were selected to provide a wide range in responses to fertilizer N treatments. No fertilizer N was applied in the previous fall or before planting corn at Sites 1 and 2, while liquid swine manure was applied uniformly across Site 3 in the fall before planting. The manure was broadcast at a rate of 37,400 L ha–1 (approximately 168 kg N ha–1) and was not incorporated into the soil. Corn hybrids were Pioneer 34R06 at Site 1, Dekalb 595 High Oil at Site 2, and Pioneer 33A14 at Site 3. Different corn hybrids were selected to represent the common genotypes available in this region. The variation in chlorophyll measurements and yields caused by differences in corn hybrids was calibrated by converting their absolute values to relative ones.

At growth stages of V3 to V4 as described by Ritchie et al. (1993), four rates of urea–ammonium–nitrate solution (56, 112, 168, and 224 kg N ha–1 at Site 1 and Site 2, while the rate 224 kg N ha–1 changed to zero at Site 3) were injected midway between every other row to a depth of 15 cm. These rates were chosen to represent a normal range of N fertilizer application without manure (Sites 1 and 2) and with manure (Site 3) for cornfields in this region. The six-row (76 cm spacing) strips extended to the full lengths of the fields (500–820 m). Each treatment was replicated at least four times depending on the width and complexity of spatial variation in each field. The blocks were selected to include the range in soil characteristics of the fields and to minimize variation in soil characteristics and planting density within a block. Each block extended 12 m along the rows and was wide enough to include each N treatment. The block was further divided into plots that corresponded to the six-row strips having different N treatments. Each block included a complete set of treatments in an increasing or decreasing order of N rates (with a stratified block design) to reduce the border effect. Plots were marked by flags and located by using a differential global positioning system.

Leaf Chlorophyll and Grain Yield Measurements
Leaf chlorophyll concentrations were measured using a Minolta SPAD-502 meter (Spectrum Technologies, Inc., Plainfield, IL). The uppermost mature leaf was used for measurement until tassel emergence; thereafter the ear-leaf was measured. Three measurements were taken at growth stages V6 to V9 in June, V12 to R1 in July, and R2 to R5 in August at each of the three sites. The measurements were taken halfway between the stalk and the leaf tip on 30 plants randomly selected from the center four rows of each plot. Relative CMRs were obtained by expressing CMRs for plots being considered as a percentage of CMR for the plot receiving the highest N rate (i.e., 224 kg N ha–1 for Sites 1 and 2; 168 kg N ha–1 for Site 3) within the same block.

Grain yields were measured by using a combine equipped with yield monitors and global positioning system receivers. Each strip was harvested as a single combine swath. Yields were recorded at 1-s intervals by an AgLeader Yield Monitor 2000 (AgLeader Technologies, Inc., Ames, IA), and positions were recorded using a Trimble AgGPS 132 receiver (Trimble Navigation Ltd., Sunnyvale, CA) when the combine moved continuously across all test plots at a speed of 6.5 km h–1 and in the same direction to avoid errors associated with lag calibration (Arslan and Colvin, 2002). A mean yield for each 12-m plot included five to six measurements exclusive of outliers that exceed two standard deviations of the mean. Relative yields were computed in the way similar to relative CMRs. Linear regression analyses and significance tests were performed by using Statistical Analysis Systems (v. 9.1, SAS Institute, Cary, NC).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Yield Responses and Measurement Errors
Analyses presented in Table 1 indicate that our ability to detect yield responses by using yield monitors was as good as normally observed in small-plot N response trials. The great difference (up to 21%) in attainable yields at the highest N rate was mainly due to the difference in corn hybrids and field conditions across the three sites. Yield responses to N fertilizer were observed in a decreasing order from Site 1 to Site 2 to Site 3. The responses to incremental increases in rates of N fertilization at all three sites followed a diminishing trend as yields asymptotically approached a maximum (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Corn yield responses to increases in rates of N fertilization and levels of statistical significance test.

 
The importance of errors associated with yield measurements in our study (standard errors <5%) deserves attention because it has been reported that yield monitors may not be reliable when they are used on small areas (Blackmer and White, 1998). Arslan and Colvin (2002) reported an average error of 3.4% in estimating yields at a constant speed and of 5.2% in estimating yields at a varying speed of the combine. They proved the reliability of individual yield monitor data that were averaged over lengths of 10 to 25 m or about 6-s intervals with an acceptable standard error about 3%.

Relationships between Chlorophyll Meter Readings and Yields
Nature of the Relationships
Linear relationships between CMRs and grain yields observed in this study (Fig. 1 ) are similar to those observed in production agriculture fields (Blackmer and Schepers, 1995). The variables used to establish the relationships are actually two different symptoms of N deficiencies (i.e., reduced CMRs during growth and reduced yields at harvest) over various ranges in degree of N deficiencies. The underlying cause of the relationship is that both symptoms were essentially caused by the same shortages of N or deficiencies of N. Hence, the relationships should not be confused with the relationships indicating dependency of variables, or cause and effect.


Figure 1
View larger version (29K):
[in this window]
[in a new window]

  		Fig. 1. Relationships between chlorophyll meter readings (CMRs) and corn yields in test plots within each of three field-scale trials where various rates of N were applied. Solid dots indicate values from plots having the highest rate of N.

 
The fact that the CMR data were collected substantially before the grain yield data makes it tempting to assume that their relationships indicate cause and effect. This temptation is enhanced by knowledge that shortages of N should be expected to reduce rates of chlorophyll production, which in turn should limit the potential for grain production. It is necessary to recognize, however, that each relationship is based on information that was not available until the completion of the growing season. Results shown in Fig. 1 are post-season analyses of the data collected at individual sites and different times and do not involve any prediction of yields.

Slope
The CMRs taken at different times at three sites tended to be directly related to the range in yield levels measured (Fig. 1). Relationships observed at Site 1 showing relatively large yield responses to the added fertilizer were associated with a high degree of linearity (Fig. 1a, 1b, 1c). The slope in each component of Fig. 1 indicates the units of yield increase associated with a unit of CMR increase with respect to fertilizer-induced effects on the corn plant (Zhang et al., 2008a). Each slope is established by regression analysis of numerous observations after outliers have been eliminated by commonly acceptable techniques as described in the section of Materials and Methods. The determined values for slopes, therefore, are not greatly influenced by random errors in measuring individual CMRs or yields. All linear regression equations in Fig. 1 are statistically significant (P < 0.05). Increases in the number of observations, especially within the range of interest, should be considered an effective way to reduce uncertainty in the values determined for slopes (Kyveryga et al., 2007).

Linearity
The linear relationships between CMRs and yields at Sites 1, 2, and 3 all tested statistically significant on three measurement dates (P < 0.05). This suggests that the sensitivity of CMRs with respect to changes in yields was constant over a wide range in rates of fertilization. This is especially true at Sites 1 and 2 (the most responsive sites) where no manure fertilizer was applied and large amounts of fertilizer N were needed to maximize yields (Fig. 1a-1f. Note the points associated with the highest N rate are concentrated at the upper-right corner). The finding that the two symptoms of N deficiency show linear relationships over wide ranges of N deficiencies is important in situations where extreme deficiencies of N are common and need to be detected. This property, however, should not be confused with the ability of the chlorophyll meters to detect small N deficiencies in situations where extreme deficiencies of N usually are not observed.

The linearity occurs in most cases shown in Fig. 1 because CMRs show poverty adjustment, that is, increases in CMR followed by increases in yield as similarly defined by Macy (1936) for plant tissue tests. There is no physiological reason, however, to expect that the relationship should always be linear rather than curved (Markwell et al., 1995). Indeed, breaks in linearity should be expected at the lower portion of yields when the critical CMRs are reached and at the upper portion of yields when luxury production of chlorophyll occurs according to Macy's concept (Zhang et al., 2008b).

Coefficient of Determination (r2)
Values of r2 shown in Fig. 1 represent coefficients of determination or proportions of the sample variation around the mean yield that is explainable by a linear relationship between CMRs and yields at each site. Because the range in yield levels indicates the magnitude of yield response observed to added fertilizer, this range determines the sensitivity of chlorophyll meters for diagnosing N deficiency at a site. Regression analyses indicate that all the r2 values tested significant at P < 0.05.

The relationships observed are profoundly influenced by the number of observations showing severe N deficiencies and by the overall responsiveness of the sites included in this study (Fig. 1). The finding that r2 values are largely determined by the magnitude of responses poses a problem when trying to use these values to determine the minimum deficiency of N that can be detected. This problem can be resolved, however, if the r2 values are used to describe relationships in the near-optimal range of CMRs (associated with relative yields >95%, for example).

The results presented in Fig. 1 illustrate the dilemma that too few observations fall within the near-optimal range when trials are conducted at responsive sites, or statistically significant r2 values are not observed if trials are conducted at sites that show too little response. This suggests that N deficiencies in nonirrigated corn are too small to detect in situations where r2 values for an observed relationship between CMRs and yields are not statistically significant at P < 0.10 (Fox et al., 2001; Scharf et al., 2006). Pooling data from many sites may offer a potential way to resolve this dilemma.

Stability and Seasonal Changes
Figure 2 shows the relationships between CMRs and yields when data from three sites are pooled. The CMRs taken in June were more poorly related to grain yields than the CMRs taken in July and August. The r2 values indicate that only 26% of the sample variability in yields can be attributed to the CMRs taken in June, 51% to the CMRs taken in July, and 40% to the CMRs taken in August. Regression analyses further suggest that the relationships would become statistically insignificant (P > 0.10) if the observations with yields lower than 8.5 Mg ha–1 (corresponding to approximately the lower one-third of CMRs taken in June, July, and August) were omitted from the data set analyzed (data not shown). It must be concluded, therefore, that there existed a poor relationship between CMRs and yields at near-optimal supplies of N despite data from different sites being pooled.


Figure 2
View larger version (20K):
[in this window]
[in a new window]

  		Fig. 2. Relationships between chlorophyll meter readings (CMRs) and corn yields for pooled data from three field-scale trials where various rates of N were applied. Solid dots indicate values from the reference plots receiving the highest rate of N.

 
The finding that data from different sites cannot be pooled into a single useful relationship in the near-optimal range of N supplies suggests that the relationship observed at one site cannot be used to predict the relationship that would be expected at another site for the near-optimal range of N supplies. As noted by Schepers et al. (1992b) as well as Cerrato and Blackmer (1990, 1991) and Zhang et al. (2006), tools for diagnosing N deficiencies have little value in production agriculture unless observations from response trials conducted under a wide variety of conditions can be pooled into a common relationship with reasonable predictability within the range of practical interest. There is little evidence in our data to conclude that CMRs collected during vegetative growth can be successfully used to predict grain yields across the three sites. The underlying issue is the stability of the relationship between CMRs and yields across different conditions.

Some degree of limitation on the sensitivity of chlorophyll meters to measure symptoms of N deficiency early in the season is indicated by greater variability of CMRs taken in June at the highest rate of N fertilization (shown as solid dots in Fig. 1 and Fig. 2). The highest rate of N fertilization applied at early growth stages (V3–V4) was believed to be great enough to eliminate N deficiencies that would normally be found in production agriculture (Zhang et al., 2007a). The limitation on the sensitivity of chlorophyll meters is probably caused by the spatial variability in temporary deficiencies of N or by the effect of factors other than N deficiencies. The limitation of measuring symptoms of N deficiency early in the season is well indicated by observations that some CMRs from plots receiving the intermediate rates of N exceeded the CMRs from plots receiving the highest rate of N (Fig. 2). Nevertheless, the finding that data points were not located in clusters corresponding to N treatments provides additional evidence of the limited ability of CMRs when they are used to estimate the N deficiency in the near-optimal range of N supplies.

The observation that CMRs made during the growing season have limited ability to predict yields at the end of the season is not surprising because it is well established that both CMRs and yield levels are influenced by many factors (e.g., genetics, density or age of plant, moisture status of soil, and supply of other nutrients) other than deficiencies of N alone. In recognition of this problem, a sufficiency concept is usually applied where CMRs and yields are expressed as percentages of those observed on reference plants growing with ample supplies of N (Varvel et al., 1997).

Relationships between Relative Chlorophyll Meter Readings and Relative Yields
Importance of the Relationships
Changes in CMRs and yields caused by the effects of factors other than N deficiencies are reduced by expressing the two measurements of N deficiency on relative bases and establishing relationships as illustrated in Fig. 3 . This transformation is appropriate in our study because plots within each block were located on relatively homogenous soil. This technique reduces variability in both CMRs and yields at the highest rate of N fertilization. The reference plots are believed to have adequate N to avoid potential yield losses due to N deficiencies.


Figure 3
View larger version (35K):
[in this window]
[in a new window]

  		Fig. 3. Relationships between relative chlorophyll meter readings (CMRs) and relative yields of corn in test plots within each of three field-scale trials where various rates of N were applied.

 
The relationships between relative CMRs and relative yields are similar to those shown in Fig. 1, but some differences are noteworthy. Observations made on plots that received the highest N rate are not shown in Fig. 3 because of their unique value (i.e., 100% relative yields, 100% relative CMRs). The position of these points is identified by the intersection of the dashed lines in the figure. These points were not included in linear regression analyses to assess the ability of using measured CMRs to predict yield values. Any error of measurement made in the reference plots is essentially transferred to all other points during data transformation. When data from individual sites are analyzed, it is difficult to determine why some regression lines do not pass through the known values or the target for relative yields and relative CMRs at the highest rate of N fertilization.

Target of the Regression
Regression lines that do not pass through the target (100% relative yields, 100% relative CMRs) indicate an error when CMRs taken at other rates of N fertilization are used to predict the known yield values. The missing of target could be due to errors of measurements, which include improper or inconsistent use of chlorophyll meters or methods of measuring yields, inadequate numbers of observations to address variation among plants within an area, or incorrect assumption that different areas are identical in all growth-affecting factors except for supplies of N. All of these can be described as "experimental noise" when trying to assess differences between two measurements.

Errors in regression analyses could also be caused by false assumptions when selecting a model. For example, an error would occur if a straight line were used to describe a relationship that is actually curved or shows a discontinuity. The overall dilemma illustrated above is that analyses of relationships between the two measurements of N deficiencies in the near-optimal range must be based on extrapolations from outside the range of interest or relationships that are not statistically significant. Extrapolations from outside the range of interest, of course, cannot reveal possible discontinuities within the range of interest.

Discontinuity
Based on observations made elsewhere (Zhang et al., 2007b), two possible types of discontinuity in the relationships between relative CMRs and relative yields could be expected in the near-optimal range of N supplies. A type I discontinuity could be expected when CMRs are collected early in the season (Fig. 3a, 3d, 3g) because supplies of N can be depleted during plant growth and young plants may not show the deficiency symptoms that will appear later in the season. Fitting a straight line to data showing such a discontinuity would result in an error. For example, the regression analyses would underestimate relative yields by 7 to 9% at the relative CMR of 100% at the three sites (Fig. 3a, 3d, 3g).

A type II discontinuity could be expected if a phenomenon like luxury production of chlorophyll occurred, if plants recovered from a deficiency that produced symptoms as indicated by CMRs (Zhang et al., 2008b), or if plants receiving the highest rate of N showed negative effects on grain yields. The negative effects of incremental increase in N rate on yields would result in optimal relative CMRs < 100%. At Site 1, for example, the linear regression line would overestimate relative yields by 10% at the relative CMR of 100% (Fig. 3c). Obviously, straight line regression models can perform poorly in situations involving nonlinear relationships (Saltelli et al., 2004).

The discontinuity illustrated above becomes more obvious with the pooled data from three sites where both large and small yield responses were observed (Fig. 4 ). Although the regression line essentially passes through the intersecting point indicating 100% relative yields and 100% relative CMRs in Fig. 4a for the data collected in June, this trend changes in Fig. 4b, 4c and provides no assurance that reliable diagnoses can always be made in the near-optimal range of N supplies with the data collected in July and August.


Figure 4
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 4. Relationships between relative chlorophyll meter readings (CMRs) and relative yields of corn for pooled data from three field-scale trials where various rates of N were applied.

 
Indeed, if the observations are compressed into the range of relative CMRs >85%, the r2 values for linear relationships between relative CMRs and relative yields are reduced to 0.02, 0.07, and 0.003 for June, July, and August, respectively. None of these values become statistically significant (P > 0.10). This is consistent with the findings of Fox et al. (2001) and Scharf et al. (2006). Similar results are reported by Ortuzar-Iragorri et al. (2005) with the conclusion that normalized CMRs cannot be considered a tool for grain yield prediction when only N rates >100 kg ha–1 are applied.

Obviously, relative CMRs and relative yields >100% indicate problems caused by measurement errors or by incorrect assumptions about the linearity of relationships. Errors in assumptions could relate to the importance of luxury production of chlorophyll in leaves (Zhang et al., 2008b), negative effects of excess N on yields, or nonlinearity of relationships in the below-optimal and above-optimal range of N supplies (Zhang et al., 2007b). Therefore, errors in observations cannot be distinguished from errors in assumptions without independent observations.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Reduced chlorophyll concentration of corn leaves and losses of grain yields are both symptoms of N deficiencies. The post-season analyses should not assume that the two measurements of symptoms of N deficiencies are cause and effect. The sensitivity of chlorophyll meters discussed in this paper is associated with fertilizer-induced yield responses, hence with deficiencies of N that would be usually observed in production agriculture.

This study suggests that CMRs measured early in the season cannot be successfully used to predict relative yields of grain with reasonable certainty when supplies of N are near optimal. Results of this study demonstrate several reasons why chlorophyll meter measurements have poorer sensitivity when they are used to diagnose deficiencies of N in production agriculture than they are used to help interpret results in a well-controlled experiment. First, diagnoses of N deficiencies in production agriculture must be based on relationships established in previous years under other conditions (i.e., use in production agriculture depends on "calibrations" established in research plots). Second, factors other than sufficiency of N have greater effects on CMRs and grain yields in field-scale trials than in well-controlled small-plot experiments. Third, the relationships between CMRs and yield responses may not be linear when supplies of N are increased from below-optimal to above-optimal rates of fertilization. All of these possibilities interact and create great uncertainty when chlorophyll meters are used to diagnose N deficiencies and estimate N fertilizer needs in production agriculture. Although these problems do not diminish the value of the chlorophyll meters for detecting severe deficiencies of N, they do limit the value of the test for refining estimates of fertilizer needs in production agriculture.


   ACKNOWLEDGMENTS
 
This paper was rewritten from part of the senior author's dissertation that was completed under close supervision of Drs. Alfred Blackmer, Kenneth Koehler, Antonio Mallarino, Irvin Anderson, and Cynthia Cambardella at Iowa State University. Dr. Randy Killorn kindly lent a chlorophyll meter for the field data collection. Maureen Schaber at the Pacific Agri-Food Research Centre, BC, Canada provided constructive criticism and comments. We were most encouraged by the spirit of the late professor Dr. Alfred Blackmer (1943–2006) to complete the submission of this paper. Financial support from the Case New Holland International (formerly Case IH) and the Iowa Soybean Association is most appreciated.

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhang, J.
Right arrow Articles by Koehler, K. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Zhang, J.
Right arrow Articles by Koehler, K. J.
Agricola
Right arrow Articles by Zhang, J.
Right arrow Articles by Koehler, K. J.
Related Collections
Right arrow Nitrogen
Right arrow Field-Scale Studies
Right arrow Maize
Right arrow Field evaluation techniques
Right arrow Nutrient Management
Right arrow Site-Specific Analysis
Right arrow Soil Fertility and Productivity
Right arrow Statistics

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