Table Of ContentsNext Page

Spectral response of nitrogen fertilization in cotton (Gossypium species)

Mohammad Shamim Ansari1, Raj Kumar Mahey2, Sukhjinder Singh Sidhu3 and Gulsan Singh Bahl4

1 Dookie College, LFR, the University of Melbourne, Dookie, Victoria - 3647, Australia. Email m.ansari@pgrad.unimelb.edu.au
2
Department of Agronomy and Agrometeorology, PAU, Ludhiana – 141 004, India. Email maheyrk@yahoo.com
3
Department of Mathematics, Statistics and Physics, PAU, Ludhiana -141 004, India. Email sidhu_ss@yahoo.com
4
Department of Soils, PAU, Ludhiana-141 004, India

Abstract

Remote sensing is an effective tool to monitor and assess nutrient stressed crop growth and conditions. A field experiment was conducted on American (Gossypium hirsutum) and Desi (G. arboreum) cotton at Ludhiana, India during Kharif 1997-98 to study spectral response and understand relationships between agronomic and spectral parameters of nitrogen fertilized crops. Five nitrogen (N) levels were applied to a Trans-Gangetic alluvial sandy loam soils. Reflectance’s data were collected fortnightly interval throughout growth period in red (R) (625 - 689 nm) and near infrared (NIR) (760-897 nm) wave band from a hand held multiband ground truth radiometer (GTR 12-0.4). In no N fertilization, highest and lowest reflectance percentages were observed in R and NIR, respectively and vise versa in highest N fertilized crop. Both spectral indices i.e., radiance ratio (RR) and normalized difference vegetation index (NDVI) were lowest under no N and highest in 160 Kg N. The maximum differences and peak in spectral reflectances and indices of different N level crop were observed during 95-125 days after sowing for both cotton species. This stage coincided with the initiation to 25% of boll opening in G. hirsutum and 50% boll formation to initiation of boll opening in G. arboreum. The plant height, total and partitioned dry biomass were highly correlated with NDVI than RR, and leaf area index (LAI) and chlorophyll content with RR than NDVI. Hence crop N conditions can be assessed from the percent reflectance and spectral indices (RR and NDVI) and the agronomic parameters can be estimated by the spectral indices.

Key words

reflectance, vegetation indices, nitrogen, cotton

Introduction

Detection of nutrient especially N of crops are basic pre-requisites for timely remedial measures to improve crop productivity. N deficiency in plant leads to reduction of LAI (Van Delden, 2001), chlorophyll content (Moorby and Besford 1983), dry mass and ultimately reduction in yield (Reeves et al 1993). The researchers reported the spectral discrimination of N treatments at different crops in many countries (Stanhill et al 1972 in Israel, Walburg et al 1982 and Hinzman et al 1986 in USA, Clevers 1986 in Netherlands, Kleman and Fagerlund 1987 in Sweden, Patel et al 1985, Mahey et al 1989 and Mahey et al 1990 in India, Boese 1994 in Germany, Fernandez et al 1994 in Spain, Shen et al 2001 in China, etc). As cotton is recognized a ‘white gold’ plays a dominant role in world agriculture and industrial economy. Few researches were investigated in cotton in respect of nitrogen assessment with remote sensing (Hattey et al 1994, Saranga et al 1998, Tarpley et al 2000 and Read et al 2002) and there were lack of studies to relate to spectral indices with agronomic parameters. Thus, there was a need to investigate the spectral response and understand the relationships between agronomic and spectral parameters in different levels of nitrogen fertilized cotton crops during crop growth period.

Methods

Site and treatments

A field experiment was conducted on cotton (Gossypium species) during Kharif season of 1997-98 at Punjab Agricultural University, Ludhiana, India situated at latitude of 30 54' N and longitude of 75 49' E at a height of 247 meters above the mean sea level. Crop was grown in sixty 10.0 m x 4.5 m test plots laid out in a split-plot design with two sowing dates (normal and late sowing) and two cotton species {American cotton; G. hirsutum (cv. F 846) and Desi cotton; G. arboreum (cv LD 327)} as the main treatments and five nitrogen fertilizer application rate (0, 40, 80, 120 and 160 Kg N /ha) as the sub-treatment. These treatments were applied to a Trans-Gangetic alluvial sandy loam soils having a neutral pH, low available N (140.8 Kg N/ha), medium phosphorus and high potassium content.

Spectral and biometrical data

The reflectance data were taken between 0930 and 1300 hours from red (R) (625 - 689 nm) and near infrared (NIR) (760-897 nm) band from a hand held multiband ground truth radiometer (GTR 12-0.4) manufactured by Optomech Engineers Private Limited, Hyderabad, India. Six to eight measurements were made on each plot measuring 10.0 m x 4.5 m with radiometer elevated 1 m vertically above the top of crop canopy surface, throughout the growth cycle of the crop at fortnightly interval. Immediately after each spectral measurement on a given plot - a solar intensity calibration reading was taken from a white standard with BaSO4 coated reference panel board. The data were collected in direct sunlight in cloudless sky. The per cent reflectances (% R and % NIR) and spectral indices ( radiance ratio, RR and normalized difference vegetation index, NDVI) were calculated from the canopy reflectance and BaSO4 board data. The per cent R reflectance (% R) values were calculated by dividing the canopy reflectance values at R band with that of the corresponding standard values taken from a BaSO4 board. Similarly, the per cent NIR reflectance (% NIR) values were calculated by dividing the canopy reflectance values at NIR band with that of corresponding standard values taken from a BaSO4 board. The RR was calculated by dividing the canopy reflectance of NIR band with that of R band i.e., NIR/R. The NDVI was calculated as the difference between reflectance of NIR band and R band divided by the sum of these two bands viz., (NIR-R)/(NIR+R). The biometrical observations were recorded concurrently (except with chlorophyll content) with all dates spectral data.

Results and discussions

The highest and lowest reflectance percentages are observed in R and NIR band, respectively in no N fertilization and vice- versa in highest N fertilized crop (Fig.1). These results were supported by the work of Stanhill et al (1972). RR and NDVI are lowest under no N and highest at 160 Kg N fertilizations (Fig.2). Rao et al (1997) also reported similar results. The maximum differences and peak in spectral indices are observed during 95-125 days after sowing (Fig 2). This stage coincided with the initiation to 25% of boll opening in G. hirsutum and 50% boll formation to initiation of boll opening in G. arboreum (Ansari 1999).

Fig.1. Effect of nitrogen levels on (a) red and (b) near infrared reflectance

Fig.2. Effect of nitrogen levels on (a) radiance ratio and (b) NDVI

The LAI, plant height, total, stem and leaf dry biomasses are found highly significantly correlated with both RR and NDVI to all N levels for both cotton species (Table 1). Reproductive dry biomass and chlorophyll are not highly correlated with all N levels. N availabilities were an important determinant of crop productivity (Van Keulen et al 1989). Moreover, N stress can also cause qualitative depreciation of crops (Akiyama et al 1996). The regression analysis showed that in general, NDVI was more correlated than RR with agronomic parameters like plant height; total, stem, leaf and reproductive dry matter in both cotton species and N fertilization levels whereas RR was more correlated than NDVI with LAI and chlorophyll content. The chlorophyll content is highly significantly negatively correlated with RR than NDVI at 0 and 40 kg N/ha treatments in G. hirsutum but not with G. arboreum cotton. That implies that chlorophyll concentration can be estimated in N deficient crop.

Table 1. The correlation (r) between crop parameters and spectral indices affected by nitrogen fertilization (significant at α = 0.01)

N levels
(kg/ha)

LAI

Plant height

Total dry wt.

Stem dry wt.

Leaf dry wt.

Reproduc-
tive dry wt.

Chlorophyll content

RR

NDVI

RR

NDVI

RR

NDVI

RR

NDVI

RR

NDVI

RR

NDVI

RR

NDVI

American cotton (G. hirsutum)

0

.73

.77

.40*

.68

.47*

.63

.40*

.59

.68

.66

.15a

.38*

-.66

-.64*

40

.83

.76

.51

.66

.49*

.64

.44*

.59

.62

.74

.23a

.36a

-.69

-.67

80

.82

.74

.51

.65

.48*

.62

.48*

.60

.57

.69

.26a

.39*

-.47*

-.38a

120

.83

.74

.53

.66

.54

.68

.50

.62

.64

.76

.23a

.35a

-.60*

-.59*

160

.81

.72

.51

.68

.58

.73

.51

.65

.64

.76

.18a

.34a

-.40a

-.47*

mean

.82

.75

.50

.67

.54

.64

.51

.63

.64

.73

.18a

.34a

.20ab

.20ab

Desi cotton (G. arboreum)

0

.68

.63

.77

.81

.72

.79

.64

.71

.73

.75

.42*

.46*

-.33a

-.28a

40

.69

.66

.75

.83

.67

.78

.64

.74

.58

.76

.51

.50

-.11a

-.04a

80

.76

.68

.81

.84

.75

.81

.69

.76

.68

.72

.26a

.27a

-.01a

-.05a

120

.78

.68

.74

.81

.74

.83

.65

.77

.66

.73

.52

.52

-.18a

.01a

160

.78

.69

.67

.76

.74

.83

.66

.77

.69

.74

.46*

.46*

-.21a

-.14a

mean

.77

.68

.72

.80

.73

.81

.68

.76

.68

.74

.25a

.27a

.16ab

.24ab

American+Desi cotton

mean

.69

.66

.66

.75

.61

.73

.59

.69

.57

.68

.22a

.27a

.30ac

.34ac

* denotes significant at α = 0.05
a denotes not significant at either α = 0.05 or α = 0.01
n = 22 (n for mean is 110 for American and desi cotton, 220 for both as together)
n = 12
n at b = 60 and n at c = 120

Conclusions

Thus, the crop N conditions can be assessed from the per cent reflectance (%R and %NIR) and spectral indices (RR and NDVI) during the crop growth periods. The strong relationship between agronomic parameters and vegetation indices shows that the agronomic parameters can be estimated by the spectral indices with respect of nitrogen treatment.

References

Akiyama T, Inoue Y, Shibayama M, Awaya Y, Tanaka N, Carter TR (1996). Monitoring and predicting crop growth and analysing agricultural eco systems by remote sensing and global climate change and agriculture in the North. Proc International Symposium on Agriculture and Food Science. Vol. 5, pp 367-376. Helsinki, Finland.

Ansari MS (1999). Spectral response in relation to growth and yield of cotton. MSc thesis. Punjab Agricultural University, Ludhiana, India. p 49.

Boese L (1994). Test and Comparison of Different Nitrogen Status Indicators of Winter-Wheat Crops. Zeitschrift Fur Pflanzenernahrung Und Bodenkunde 157(1): 53-60.

Clevers JGPW (1986). Application of remote sensing to agricultural field trials. Agricultural University of Wageningen Paper 86-4.

Fernandez S, Vidal D, Simon, E, and Solesugranes L (1994). Radiometric Characteristics of Triticum aestivum Cv Astral under Water and Nitrogen Stress. International Journal of Remote Sensing 15: 1867-1884.

Hattey JA, Sabbe WE, Baten GD and Blakeney AB (1994). Nitrogen and Starch Analysis of Cotton Leaves Using near-Infrared Reflectance Spectroscopy (NIRs). Communications in Soil Science and Plant Analysis 25: 1855-1863.

Hinzman LD, Bauer ME and Daughtry CST (1986). Effects of nitrogen fertilization on growth and reflectance characteristics of winter wheat. Remote Sensing of Environment 19: 47-61.

Kleman J and Fagerlund E (1987). Influence of different nitrogen and irrigation treatments on the spectral reflectance of barley. Remote Sensing of Environment 21: 1-14.

Mahey RK, Sidhu SS, Singh R and Narang RS (1989). Spectral assessment of agronomic parameters and grain yield of Gobhi Sarson (Brassica napus). Proc Nat Symp on Engineering Application on Remote Sensing and Recent Advances. pp 174-178.

Mahey RK, Singh R, Sidhu SS, and Narang RS (1990). Assessment of Spectral Response of Maize to Varying Nitrogen Levels. Indian Journal of Agronomy 35: 153-158.

Moorby J and Besford RT (1983). Mineral nutrition and growth. In: Lanchi A and Bieleski RL (ed.) Inorganic plant Encyclopedia of Plant Physiology New Series, Vol 15 B. pp 529. Verlag, Berlin.

Patel NK, Singh JP, Sahai B and Patel MS (1985). Spectral response of rice crop and its relations to yield and yield attributes. International Journal of Remote Sensing 6: 657-664.

Rao TV, Brahma BC and Singh RK (1997). Spectral response and canopy temperature of rainfed wheat in relation to pre-sown mulches and nitrogen levels. Annals of Agricultural Research 18: 127-134.

Read JJ, Tarpley L, McKinion JM and Reddy KR (2002). Narrow-waveband reflectance ratios for remote estimation of nitrogen status in cotton. Journal of Environmental Quality 31: 1442-1452.

Reeves DV, Mask PL, Wood CW and Delaney DP (1993). Determination of wheat nitrogen status with a hand held chlorophyll meter: Influence of management practices. Journal of Plant Nutrition 16: 769-781.

Saranga Y, Landa A, Shekel Y, Bosak A and Kafkafi U (1998). Near-infrared analysis of cotton leaves as a guide for nitrogen fertilization. Agronomy Journal 90: 16-21.

Shen YA, Lo JC, and Cheng SP (2001). Development of remote sensing techniques to identify N status of paddy rice. Chinese Journal of Agrometeorology 7: 23–31.

Stanhill G, Kafkafi U, Fuchs M and Kagan Y (1972). The effect of fertilizer application on solar reflectance from a wheat crop. Israel Journal of Agricultural Research 22: 109-118.

Tarpley L, Reddy KR and Sassenrath-Cole GF (2000). Reflectance indices with precision and accuracy in predicting cotton leaf nitrogen concentration. Crop Science 40: 1814-1819.

Van Delden A (2001). Yield and growth components of potato and wheat under organic nitrogen management. Agronomy Journal 93(6): 1370-1385.

Van Keulen H, Goudriaan J and Seligman NG (1989). Modelling the effects of nitrogen on canopy development and crop growth. In: Rusell G, Marshall B and Jorvus PG (ed) Plant Canopies: their growth form and function. pp 83-104. Cambridge University Press.

Walburg G, Bauer ME, Daughtry CST and Housley TL (1982). Effects of nitrogen nutrition on the growth, yield and reflectance characteristics of corn. Agronomy Journal 74: 677-683.

Top Of PageNext Page