PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jzusbLink to Publisher's site
 
J Zhejiang Univ Sci B. 2010 June; 11(6): 465–470.
PMCID: PMC2880361

Detection of nitrogen-overfertilized rice plants with leaf positional difference in hyperspectral vegetation index*

Abstract

The main objective of this work was to compare the applicability of the single leaf (the uppermost leaf L1 and the third uppermost leaf L3) modified simple ratio (mSR705 index) and the leaf positional difference in the vegetation index between L1 and L3 (mSR705L1−mSR705L3) in detecting nitrogen (N)-overfertilized rice plants. A field experiment consisting of three rice genotypes and five N fertilization levels (0, 75, 180, 285, and 390 kg N/ha) was conducted at Xiaoshan, Hangzhou, Zhejiang Province, China in 2008. The hyperspectral reflectance (350–2500 nm) and the chlorophyll concentration (ChlC) of L1 and L3 were measured at different stages. The mSR705L1 and mSR705L3 indices appeared not to be highly sensitive to the N rates, especially when the N rate was high (above 180 kg N/ha). The mean mSR705L1−mSR705L3 across the genotypes increased significantly (P<0.05) or considerably from 180 to 285 kg N/ha treatment and from 285 to 390 kg N/ha treatment at all the stages. Also, use of the difference (mSR705L1−mSR705L3) greatly reduced the influence of the stages and genotypes in assessing the N status with reflectance data. The results of this study show that the N-overfertilized rice plants can be effectively detected with the leaf positional difference in the mSR705 index.

Keywords: Rice, Nitrogen (N), Overfertilization, Leaf position, Hyperspectral reflectance

1. Introduction

Rice is the main food staple for more than 50% of the world’s population (Fageria and Baligar, 2003). Nitrogen (N) fertilizer is one of the most important inputs for rice production in the world. Increasing the N fertilizer rate for rice plants, however, does not always increase grain yield, due to diminishing returns. The excessive use of N fertilizers poses potential adverse environmental and health concerns (Bohlool et al., 1992), and increases the incidence of foliar pathogens and plant lodging (Stroppiana et al., 2009). Farmers tend to use more N fertilizer than needed mainly because of its subsidized price and its immediate visible impact on plant growth and leaf color (Islam et al., 2007). Approximately 30% of the N used as fertilizer in the world is consumed in China, rice crop use accounting for about 37% use (Peng et al., 2002). Therefore, rapid and real time detection of N overfertilization during rice production could be very helpful for site-specific N management.

In recent years, spectral measurements have been used for rapid and non-destructive estimation of crop N status. The use of radiometric data for N estimation has been reported over a wide range of crops (Blackmer et al., 1996; Boegh et al., 2002; Hansen and Schjoerring, 2003). Controversy exists, however, in the use of radiometric data for rice (Shibayama and Akiyama, 1986; Takebe et al., 1990; Xue et al., 2004; Nguyen and Lee, 2006; Zhu et al., 2007; Yi et al., 2007; Stroppiana et al., 2009). Estimating plant N status by spectral data is sometimes difficult as values may vary according to plant age, location, cultivar selection, and soil characteristic (Spaner et al., 2005). Particularly, N estimation for crops of high N and chlorophyll (Chl) contents using reflectance data may be difficult because crop reflectance may be insensitive to crops of high N and chlorophyll contents due to the saturated light adsorption in the chlorophyll adsorption region (Thomas and Gausman, 1977).

N availability profoundly influences the distribution of leaf N, and thus leaf chlorophyll within a canopy profile as N is a highly variable element in the plant. Wang et al. (2006) showed that the leaf positional differences in N concentration in response to varied N rates could be a useful indication of plant N status in rice. It was implied that the leaf positional differences in N-sensitive or chlorophyll-sensitive hyperspectral indices might be useful for reducing the influences of some factors (e.g., locations and cultivar) on the critical hyperspectral index, thereby allowing for detection of plants with high N and chlorophyll contents by spectral data.

Previous research has paid little attention to detecting N overfertilization when assessing plant N status with the spectral approach. The main objectives of this work are (1) to test the applicability of the selected chlorophyll-sensitive hyperspectral index in detecting N-overfertilized rice plants and (2) to explore the feasibility of using the leaf positional differences in the chlorophyll-sensitive hyperspectral index for detecting N-overfertilized rice plants.

2. Materials and methods

2.1. Field experiment

A field experiment was conducted on a sandy loam soil from June to October in 2008 at the research farm of Xiaoshan Agricultural Sciences Research Institute, Xiaoshan, Hangzhou, Zhejiang Province, China (30°20′ N, 120°31′ E). Prior to planting, the original soil had 13.1 g/kg organic C, 5.6 mg/kg bicarbonate extractable P, 35.2 mg/kg exchangeable K, and 1.26 g/kg total N with pH 7.5 (soil:water 1:1 (w/v)). Three genotypes, Yongyou 8, Zhongzheyou 1, and Zhejing 22 of rice (Oryza sativa L.), were selected for study. A random block design was used with the three genotypes and five N rates (0, 75, 180, 285, and 390 kg N/ha). Forty percent of the N fertilizer (urea) was applied at pre-planting, 30% at the tillering stage, and 30% at initial heading. Each treatment was replicated three times. The plants were transplanted on July 8, 2008 and harvested on October 20, 2008.

2.2. Spectral measurement

One whole rice plant from each plot was collected at panicle initiation, heading and milk stages, and transported to the lab for leaf reflectance measurements. The reflectance of the single leaf (the uppermost leaf L1 and the third uppermost leaf L3) on a randomly selected main stem was measured with an integrating sphere (Model LI-1800, LiCor Inc., Lincoln, NE, USA) coupled to a FieldSpec® model spectral radiometer (Analytical Spectral Devices, Boulder, CO, USA) in the wavelength range of 350–2500 nm around the midpoint of each leaf, each measurement of a leaf being the average of ten scans. The white reference was taken before each spectral measurement.

2.3. Leaf chlorophyll concentration determination

After spectral measurements, leaves of 0.20 g from the middle part of each leaf were sampled for determination of leaf chlorophyll concentration (ChlC). The Chl a and Chl b concentrations per unit mass were measured spectrophotometrically using 80% (v/v) acetone as a solvent, employing the equations of Lichtenhaler and Wellburn (1983).

2.4. Data analysis

Modified simple ratio [mSR705=(R 750R 445)/(R 705R 445), where R 750, R 445, and R 705 are the reflectance at 750, 445, and 705 nm, respectively (Sims and Gamon, 2002)] was used as the vegetation index (VI) for retrieving leaf ChlC. This VI was selected because it can reduce the effect of differences in leaf surface reflectance and is relatively insensitive to leaf structure (Sims and Gamon, 2002).

Statistical analysis was done with SPSS 16.0 (Chicago, IL, USA). An analysis of variance (ANOVA) was done to determine the significance of the effects of genotypes, stages, N level, and leaf position on the spectral index and ChlC. Factorial analyses of variance were conducted to determine the significance of the effects of genotypes, stages, and N levels on the VI and ChlC differences between L1 and L3. A linear contrast test was used to test the significance of the differences among the five N levels.

For the convenience of illustration, 0, 75, 180, 285, and 390 kg N/ha are designated as N0, N1, N2, N3, and N4, respectively, VIs of L1 and L3 are expressed as VIL1 and VIL3, respectively, and the leaf ChlCs in L1 and L3 are expressed as ChlCL1 and ChlCL3, respectively. The leaf positional difference in the VI between L1 and L3 is expressed as mSR705L1−mSR705L3 or VIL1−VIL3, and the ChlC difference between L1 and L3 is expressed as ChlCL1−ChlCL3.

3. Results

3.1. Response of leaf chlorophyll to N rate

The leaf ChlC ranged from 0.20 to 6.18 mg/g, indicating the leaf ChlC range was wide due to the genotypes, N treatments, stages, and leaf positions. As shown in Table Table1,1, the genotype, the stage, and the N rate significantly affected (P<0.05) the leaf ChlC. A significant (P<0.05) variation due to stage×genotype, stage×N rate, N rate×leaf position, N rate×genotype, stage×genotype×N rate, and genotype×N rate×leaf position interactions was also observed, although the effect of the leaf position on the leaf ChlC was insignificant (P>0.05). The N rate also affected significantly (P<0.05) ChlCL1−ChlCL3 (Table (Table2).2). As presented in Fig. Fig.11 and Table Table3,3, both the mean ChlCL1 and the mean ChlCL3 across the three stages and the three genotypes and across the three genotypes at each stage tended to increase with the N rate, although the N2 treatment appeared to have the highest ChlCL1 among the N treatments at the milk stage. In particular, the N4 treatment had the highest ChlCL3 and the lowest ChlCL1−ChlCL3 among the five N levels at all the stages (Table (Table33).

Fig. 1

Mean leaf chlorophyll concentrations (ChlCs) in response to the five N rates for the three genotypes at Xiaoshan, Hangzhou, Zhejiang Province, China in 2008

Table 1

Analysis of variance of the effects of stage (S), genotype (G), N fertilization rate (N), and leaf position (LP) on mSR705 index and leaf chlorophyll concentration (ChlC) of three rice genotypes grown under five N rates at Xiaoshan, Hangzhou, Zhejiang ...

Table 2

Analysis of variance of the effects of stage (S), genotype (G), and N fertilization rate (N) on the leaf positional difference values in mSR705 and leaf chlorophyll concentration (ChlC) between the uppermost leaf (L1) and the third uppermost leaf (L3) ...

Table 3

Mean mSR705 and leaf chlorophyll concentration (ChlC) values of the uppermost leaf (L1) and the third uppermost leaf (L3) and the difference values between L1 and L3 in response to the five N rates for the three genotypes at different stages at Xiaoshan, ...

3.2. Response of the VI to the N rate

As shown in Table Table1,1, the genotype, stage, and N rate affected significantly (P<0.05) the mSR705 index. A significant (P<0.05) variation due to stage×genotype, stage×N rate, N rate×leaf position, N rate×genotype, and stage×genotype×N rate interactions was also observed for the VI, although the effect of the leaf position on the VI was insignificant (P>0.05). The results indicated that the VI varied greatly with the four factors.

The N rate had a significant (P<0.05) effect on mSR705L1−mSR705L3, but the effects of stage and genotype on mSR705L1−mSR705L3 became insignificant (P>0.05) (Table (Table2).2). The results indicated that the influences of stage and genotype could be greatly reduced when mSR705L1−mSR705L3 instead of mSR705L1 or mSR705L3 was used in assessing the N status by reflectance data.

As shown in Table Table3,3, the values of mSR705 tended to increased with the N level from N0 to N2 at all the stages for both L1 and L3. The differences in the VI were considerable, although not always significant, between N0 and N1 treatments. The mean values of the VI were not significantly (P>0.05) different among N2, N3, and N4 except that the N4 treatment had significantly (P<0.05) lower IV than N2 or N3 treatment at milk stage, indicating that the VI was not sensitive to the high N levels (above N2).

The mean mSR705L1−mSR705L3 across the genotypes decreased from N2 treatment to N4 treatment at all the stages, and was significantly (P<0.05) higher at heading and milk in N2 and N3 treatments than in N4 treatment. The difference in the mean mSR705L1−mSR705L3 was considerable, although not always significant, between N2 and N3 treatments at all the stages. The mSR705L1−mSR705L3 appeared, however, to be poor in separating N0 and N1 treatments. The results indicated that mSR705L1−mSR705L3 was sensitive to the high N levels, although it was not sensitive to changes at low N levels (N0 and N1). The N-overfertilized rice plants (N4 treatment) were characterized with the lowest and most negative values of mSR705L1−mSR705L3.

The ChlC-mSR705 relationships were positive and significant (P<0.05), but not robust as the coefficient was 0.49 and 0.64 (n=135), respectively, in L1 and L3, implying that the single leaf VI was not a robust estimator of the leaf ChlC across the genotypes and stages. Fig. Fig.22 shows that the mSR705 index tends to increase with the leaf ChlC, but becomes insensitive to changes in high leaf ChlC (above 3.5 mg/g).

Fig. 2Fig. 2

Relationship between mSR705 and leaf chlorophyll concentration (ChlC)

4. Discussion

The chlorophylls contain a large proportion of total leaf N. Therefore measurements of ChlC can provide an accurate indirect assessment of plant nutrient status (Moran et al., 2000). The results of this study further confirm that leaf ChlC can be used to assess rice N status. At the milk stage, N2 treatment appeared to have the highest ChlC in the flag leaves, but N3 and N4 treatments still had a remarkably higher ChlC than N2 treatment in L3, which might be attributed to a feedback mechanism in response to N remobilization from the leaves to the grain.

The single leaf hyperspectral index appeared not to be highly sensitive to the N rates, especially under high N conditions. This might be partly explained by the fact that VI might not be sensitive to high ChlC, since a relatively low content of chlorophyll is sufficient to saturate absorption in the chlorophyll adsorption region (Thomas and Gausman, 1977).

The leaf position did not statistically influence mSR705 or leaf ChlC. The N rate×leaf position interaction, however, significantly (P<0.05) influenced both the mSR705 and leaf ChlC (Table (Table1),1), and the N rate was found to significantly (P<0.05) influence both mSR705L1−mSR705L3 and ChlCL1−ChlCL3 (Table (Table2).2). Therefore, the VIL1−VIL3 can be used to assess rice N status. The VIL1−VIL3 appeared to be useful, while both the VIL1 and the VIL3 performed poorly in separating N2, N3, and N4 treatments, possibly because the use of the VIL1−VIL3, instead of the single leaf hyperspectral index, can greatly reduce the influence of the genotypes and the stages, and relatively enlarge the VI differences among the high N treatments (Table (Table3).3). The VIL1−VIL3 appeared, however, not to be as effective as the single leaf VI in separating N0 and N1 treatments, corresponding to the insignificant (P>0.05) difference in the ChlCL1−ChlCL3 between the two treatments. The results suggest that combined analyses of the VIL1, VIL3, and VIL1−VIL3 are needed when assessing plant N status with hyperspectral reflectance data.

The ChlC-VI relationship was not robust in both L1 and L3. The relationship would be improved if the ChlC was expressed on the basis of area. The wide range of leaf ChlC might also influence the ChlC-VI relationship, since the VI was not sensitive to high leaf ChlC.

5. Conclusions

The single leaf hyperspectral index mSR705 appeared not to be highly sensitive to the N rates, especially when the N rate was high (above N2). The leaf positional differences in the VI appear to be sensitive to the high N rates at all the stages, which could be explained by the fact that the influence of stages and genotypes can be greatly reduced, and that the treatment differences among the high N level treatments are relatively enlarged by using the VIL1−VIL3 instead of the VIL1 or the VIL3 in assessing the N status via reflectance data. The N-overfertilized rice plant (N4 treatment) was characterized with the lowest and most negative value of the VIL1−VIL3 at all the stages. The results in this study suggest that the leaf positional differences in the VI can be used to effectively detect the rice plants with N overfertilization.

Footnotes

*Project (Nos. 2007AA10Z102 and 2006AA10Z201) supported by the National High-Tech R & D Program (863) of China

References

1. Blackmer TM, Schepers JS, Varvel GE, Walter-Shea EA. Nitrogen deficiency detection using reflected shortwave radiation from irrigated corn canopies. Agronomy Journal. 1996;88(1):1–5.
2. Boegh E, Soegaard H, Broge N, Hasager CB, Jensen NO, Schelde K, Thomsen A. Airborne multispectral data for quantifying leaf area index, nitrogen concentration, and photosynthetic efficiency in agriculture. Remote Sensing of Environment. 2002;81(2-3):179–193. doi: 10.1016/S0034-4257(01)00342-X. [Cross Ref]
3. Bohlool BB, Ladha JK, Garrity DP, George T. Biological nitrogen fixation for sustainable agriculture: a perspective. Plant and Soil. 1992;141(1-2):1–11. doi: 10.1007/BF00011307. [Cross Ref]
4. Fageria NK, Baligar VC. Methodology for evaluation of lowland rice genotypes for nitrogen use efficiency. Journal of Plant Nutrition. 2003;26(6):1315–1333. doi: 10.1081/PLN-120020373. [Cross Ref]
5. Hansen PM, Schjoerring JK. Reflectance measurement of canopy biomass and nitrogen status in wheat crops using normalized difference vegetation indices and partial least squares regression. Remote Sensing of Environment. 2003;86(4):542–553. doi: 10.1016/S0034-4257(03)00131-7. [Cross Ref]
6. Islam Z, Bagch B, Hossain M. Adoption of leaf color chart for nitrogen use efficiency in rice: impact assessment of a farmer-participatory experiment in West Bengal, India. Field Crops Research. 2007;103(1):70–75. doi: 10.1016/j.fcr.2007.04.012. [Cross Ref]
7. Lichtenhaler HK, Wellburn AR. Determination of total carotenoids and chlorophylls a and b of leaf extracts in different solvents. Biochemical Society Transactions. 1983;11:591–592.
8. Moran JA, Mitchell AK, Goodmanson G, Stockburger KA. Differentiation among effects of nitrogen fertilization treatments on conifer seedlings by foliar reflectance: a comparison of methods. Tree Physiology. 2000;20(16):1113–1120. [PubMed]
9. Nguyen HT, Lee BW. Assessment of rice leaf growth and nitrogen status by hyperspectral canopy reflectance and partial least square regression. European Journal of Agronomy. 2006;24(4):349–356. doi: 10.1016/j.eja.2006.01.001. [Cross Ref]
10. Peng B, Huang JL, Zhong XY, Yang JC, Wang GH, Zou YB, Zhang FS, Zhu QS, Buresh B, Witt C. Research strategy in improving fertilizer-nitrogen use efficiency of irrigated rice in China. Scientia Agricultura Sinica. 2002;35(9):1095–1103.
11. Shibayama M, Akiyama TA. A sprectroradiometer for field use: VI. Radiometric estimation for chlorophyll index of rice canopy. Japanese Journal of Crop Science. 1986;55(5):433–438.
12. Sims DA, Gamon JA. Relationships between leaf pigment content and spectral reflectance across a wide range of species, leaf structures and developmental stages. Remote Sensing of Environment. 2002;81(2-3):337–354. doi: 10.1016/S0034-4257(02)00010-X. [Cross Ref]
13. Spaner D, Todd AG, Navabi A, McKenzie DB, Goonewardene LA. Can leaf chlorophyll measures at differing growth stages be used as an indicator of winter wheat and spring barley nitrogen requirements in Eastern Canada? Journal of Agronomy and Crop Science. 2005;191(5):393–399. doi: 10.1111/j.1439-037X.2005.00175.x. [Cross Ref]
14. Stroppiana D, Boschetti M, Brivio PA, Bocchi S. Plant nitrogen concentration in paddy rice from field canopy hyperspectral radiometry. Field Crops Research. 2009;111(1-2):119–129. doi: 10.1016/j.fcr.2008.11.004. [Cross Ref]
15. Takebe M, Yoneyama T, Inada K, Murakami T. Spectral reflectance ratio of rice canopy for estimating crop N status. Plant and Soil. 1990;122(2):295–297. doi: 10.1007/BF02851988. [Cross Ref]
16. Thomas JA, Gausman HW. Leaf reflectance vs. leaf chlorophyll and caroteniod concentrations for eight crops. Agronomy Journal. 1977;69(5):799–802.
17. Wang S, Zhu Y, Jiang H, Cao W. Positional differences in nitrogen and sugar concentrations of upper leaves relate to plant N status in rice under different N rates. Field Crops Research. 2006;96(2-3):224–234. doi: 10.1016/j.fcr.2005.07.008. [Cross Ref]
18. Xue L, Cao W, Luo W, Dai T, Zhu Y. Monitoring leaf nitrogen status in rice with canopy spectral reflectance. Agronomy Journal. 2004;96(1):135–142.
19. Yi QX, Huang JF, Wang FM, Wang XZ, Liu ZY. Monitoring rice nitrogen status using hyperspectral reflectance and artificial neural network. Environmental Science & Technology. 2007;41(19):6770–6775. doi: 10.1021/es070144e. [PubMed] [Cross Ref]
20. Zhu Y, Zhou D, Yao X, Tian Y, Cao W. Quantitative relationships of leaf nitrogen status to canopy spectral reflectance in rice. Australian Journal of Agricultural Research. 2007;58(11):1077–1085. doi: 10.1071/AR06413. [Cross Ref]

Articles from Journal of Zhejiang University. Science. B are provided here courtesy of Zhejiang University Press