Previous PageTable Of ContentsNext Page

Use of reflectance measurements to clearly identify water stress in wheat (Triticum aestivum L.)

Simone Graeff1, Zhongxue Fan and Wilhelm Claupein1

1Institute of Crop Farming and Grassland Research (340), University of Hohenheim, Fruwirthstr. 23, 70593 Stuttgart, Germany, Email graeff@uni-hohenheim.de

Abstract

Availability of water is one of the most limiting factors in crop production. Current technologies for measuring plant water status are limited. Considering plant and irrigation management it is essential to discriminate between water stress and various other possible biotic and abiotic stress factors.

A greenhouse study was conducted to determine specific reflectance wavelength ranges responsive to water stress in wheat. Pot experiments were carried out in Mitscherlich pots on a sandy loam over 10 weeks. Reflectance of wheat plants grown under five different water treatments ranging from 65 % field capacity to 26 % field capacity was determined once a week from the beginning of the 4th leaf stage until the 6th leaf stage. Reflectance measurements were performed at the 4th leaf of wheat plants with a digital camera under controlled light conditions. Reflectance was measured in different wavelength ranges in the visible and infrared spectra using various long-pass filters. Reflectance of wheat leaves changed significantly 9 d after induction of water stress at a leaf water content < 75 %. Reflectance patterns of 510780 nm, 5161300 nm, 5401300 nm were found most suitable to identify water deficiency regardless of the sampling date or growth stage. Reflectance pattern of water deficient plants were significantly different from those of other plant stresses. The results indicated that reflectance measurements may serve as a rapid, non-destructive approach to discriminate water stress from other biotic and abiotic stresses.

Media summary

A new reflectance measurement technique with the ability to discriminate between water stress and multiple other plant stresses, may enhance site-specific farming.

Key Words

Water potential, site-specific farming, irrigation strategies, L*a*b* color system, remote sensing techniques, reflectance.

Introduction

Availability of water is one of the most limiting factors in crop production. Over the past decade, the increased use of irrigation and concern over groundwater resources has brought about an awareness of efficiently utilizing water resources. So far direct plant based measurements are limited to leaf water potential by pressure chamber, stomatal conductance by porometry, and canopy temperature by infrared thermometry. These measurements are time-consuming and require a number of observations to characterize a whole field (Jackson, 1982). Because of limitations to the above methods, it would be beneficial to use remote sensing techniques to help managers to determine when and/or where a water stress exists and additionally to predict possible yield losses. Early detection of a water stress could trigger irrigation before yield loss occurs.

Several studies have examined technologies involving remote sensing to quantify water stress (Bowman, 1989; Peñuelas et al. 1993). Moran et al. (1989) investigated the effect of water stress on canopy architecture in alfalfa and the sequential effect on canopy temperature. They found water stressed canopies to have a lower spectral reflectance in the NIR and red wavebands when compared with unstressed canopies. Other studies estimated leaf water status by measuring reflectance spectra. Carlson et al. (1971), Gausman et al. (1971), and Hunt et al. (1987) analyzed the relationship between reflectance spectra and leaf water status in numerous plant species, and pointed out a possibility to estimate relative leaf water content by reflectance at specific wavelengths in the range of the near-infrared. Work by Bowman (1989) with cotton showed that the reflectance in the infrared spectra (810, 1665, and 2210 nm) increased as relative water content decreased.

So far a lot of investigations focused on the development of spectral indices for the detection of water stress. The results indicated that the relationships e.g. NIR/red and NDVI may be useful for estimating the subsequent need for irrigation. Jackson et al. (1983) used several ratios and wavelength bands and determined that water stress could not be detected until after there was a stress-reduced retardation in growth. The ability of these ratios to detect water stress depends on plant growth stage, soil background, and atmospheric changes. Further on, these reflective indices might not differ from those of other stresses (Tarpley et al., 2000) thus indicating a lack of selectivity and consequently a decrease of accuracy in predicting the water status of plants.

In the situation of employing environmental remote sensing, it is necessary to develop universal methods which can be used for the evaluation of the water status of plants. Above this the developed methods should be able to discriminate water stress from other stresses. Studies of Osborne et al. (2002) and Graeff et al. (2001, 2003) have shown that nutrient deficiencies could be identified and quantified by means of reflectance measurements based on selected stress specific wavelength ranges. This study extends the work of Graeff et al. (2001) and aims to determine whether reflectance measurements can be effectively used to identify and to discriminate water stress from other plant stresses. Greenhouse studies were used to establish a calibration for the determination of leaf water content in wheat plants by rapid and non-destructive reflectance measurements and to increase the accuracy of irrigation recommendations by clearly discriminating plant stress factors.

The specific objectives of this study were a) to determine reflectance pattern of wheat leaves exposed to water stress b) to evaluate specific wavelengths ranges indicative to water stress, and c) to compare the reflectance patterns of water stress with those of various nutrient deficiencies obtained in earlier studies of Graeff et al. (2001).

Methods

Experimental design

A pot experiment was conducted with wheat (Triticum aestivum L.) cv. Thasos under controlled conditions in the greenhouse of the Institute of Crop Farming and Grassland Research, University of Hohenheim, Stuttgart, Germany from mid of July until end of September 2002. The soil type was a sandy loam with a field capacity of 23 %. The soil was fertilized with 250 mg N [NH4NO3], 125 mg P [NaH2PO4], 125 mg K [K2SO4], 120 mg Mg [MgSO4], 130 mg Ca [CaCl2], 10 mg Mn [MnSO4], 1.5 mg Zn [ZnSO4], 1.0 mg Cu [CuSO4], 0.3 mg B [H3BO3], and 0.02 mg Mo [Na2MoO4] prior to seeding. 15 seeds were sown in each pot and thinned after two weeks to 10 seedlings per pot. The plants were allowed to grow for 30 d under identical, well watered conditions before water stress was induced. Five different water stress levels were chosen based on the field capacity (FC) of the soil and designed as 65 % (control), 52 %, 39 %, 33 %, and 26 % FC with three replications. All pots were weighed and watered each day according to the designed water levels. The average temperature was 18 °C ± 3 °C with a minimum of 13 °C and a maximum of 24°C.

Reflectance measurements

Reflectance measurements were carried out at the 4th leaf, 9, 15, 21 and 24 d after starting of water stress. Measurements were performed with a digital, light-sensitive, high-spatial resolution camera (S1 PRO, Leica, Germany) in conjunction with a constant light source (HMI 21 W/D~10 Wm-2, Sachtler, Germany) of total daylight spectra. The spectra was split into various wavelength ranges using long-pass filters, active at wavelengths longer than 380 nm, 490 nm, 510 nm, 516 nm, 540 nm, and 600 nm. Scans were carried out with the software SILVERFAST V. 4.1.4 and analyzed with ADOBE® Photoshop 5.0 in the L*a*b*-color system (CIE, 1986) by splitting the scans into a* and b* parameters in different wavelength ranges. For each plant, scans were performed with the above-mentioned long-pass filters in conjunction with a LEICA daylight filter IRa E55 to cut all scans at 780 nm (wavelength ranges indicated with X780 nm). A second set of scans was taken without this daylight filter in order to scan in the near-infrared ranges, indicated with X1300 nm. The shoots of the measured plants were harvested and the fresh weight was determined at once. Afterwards, plant samples were dried at 60°C and total dry matter was determined. The water content of the whole plant was calculated using the difference between total fresh and total dry weight.

Statistical analysis

Analysis of variance (ANOVA) was carried out on all plant and reflectance data using the general procedures of the Tukey minimum significant difference (MSD) test at the 5 % significance level of the Statistical Analysis System (SAS) version 6.12.

Results

Figure 1 shows the results of parameter b* of the control plants (65 %) and the treatments 39 % and 26 % in the wavelength range 510780 nm, 9, 15, 21 and 24 d after starting of water stress. 9 d after the induction of water stress, the reflectance parameter b* increased in all investigated water stress treatments with increasing level of water stress. Regression analysis indicated that reflectance changes were not influenced by growth stage or measuring date. For all other examined wavelength ranges in the visible and the near-infrared spectra, the reflectance parameter b* was most responsive to water stress. The a* parameter did not show any response to water stress, neither in the examined visible nor in the assessed near-infrared wavelength ranges (data not shown). Figure 2 indicates that total dry matter of wheat plants was significantly reduced in the water stress treatments 9 d after the induction of water stress. Retardation in growth went along with reflectance changes. Growth was retarded to the same degree in both water stress treatments 39 % and 26 % when compared to the control.

Figure 1. Reflectance changes of wheat leaves with different water stress levels in the wavelength range 510780 nm, measured 9, 15, 21, 24 d after starting of water stress (mean, S.E. n = 3). Significant changes are indicated at the 0.05 probability level.

Figure 2. Total dry matter of wheat plants with different water stress levels, measured at 9, 15, 21 and 24 d after starting of water stress (mean, S.E. n = 3). Significant changes are indicated at the 0.05 probability level.

Quantification of water stress

Figure 3 shows the correlation between reflectance parameter b* and total water content of measured wheat plants in the wavelength range 510780nm. The relation of parameter b* and total plant water content was best described by a linear model in the visible and in the near-infrared wavelength ranges. The highest correlation coefficients were obtained in the wavelength ranges 510780 nm (r² = 0.96), 5161300 nm (r² = 0.81), and 5401300 nm (r² = 0.80). The obtained correlations seemed to be independent of plant age and measuring date.

Figure 3. Correlation of plant water content and reflectance parameter b* in the wavelength range 510780nm with r² = 0.96, nine days after the induction of water stress.

Discrimination of various plant stresses

Reflectance spectra of water stressed wheat plants were compared with reflectance patterns of nutrient deficient plants obtained in earlier studies of Graeff et al. (2001). In comparison to other plant stresses water stressed leaves showed a clear increase of reflectance parameter b* in the visible and near-infrared spectra. In contrast to other plant stresses, no response of parameter a* could be determined over all tested wavelength ranges and water stress levels. In addition, wavelength ranges responsive to water stress were different from those evaluated for the identification of nutrient deficiencies. These results go along with earlier findings of Graeff et al. (2001) and Osborne et al. (2002). These authors have shown that different stress factors could be discriminated by reflectance measurements using appropriate wavelength ranges. Both studies found strong evidence supporting the use of visible and near-infrared wavelength bands in the elucidation of plant stress factors. The increased sensitivity in stress discrimination using appropriate wavelength bands possibly arises from the differential effects plant stress factors may have on each of the classes of compounds responsible for the multiple bands of reflectance. These approaches indicated that various stress factors could be clearly identified by real-time reflectance measurements, thus enhancing a better plant management by the integration of remote sensing techniques.

In order to finally determine the potential of the presented non-destructive method to monitor plant water status and to discriminate between the spectral properties of leaves with various other stress factors this method will have to be tested in further studies. Further studies also will have to cover the aspect, if water stress could be detected by reflectance measurements before a significant inhibition of growth occurs.

Conclusion

The results of this study have shown that reflectance measurements in the 5161300 nm, 5401300 nm, and 510780 nm region indicate plant water stress. Reflectance changes were closely correlated with total plant water content and were significantly different from those of other biotic and abiotic stresses. Thus, leaf reflectance measurements may serve as a diagnostic tool to assess and quantify water status of wheat plants and may be used to trigger irrigation. The method contributes towards the integration of remote sensing techniques into crop management. The transfer of these greenhouse calibrations to the field would offer the potential to discriminate between various stress factors and increase the accuracy in irrigation recommendations on a site-specific scale. Further investigations in this area are needed.

References

Bowman, WD (1989). The relationship between leaf water status, gas exchange, and spectral reflectance in cotton leaves. Remote Sensing of Environment 30, 249-255.

Carlson, RE, Yarger, DN and Shaw, RH (1971). Factors affecting the spectral properties of leaves with special emphasis on leaf water status. Agronomy Journal 63, 486-489.

CIE (1986). Colorimetry. 2nd edn., Publication CIE No. 15.2, Central Bureau of the Commission Internationale de L’Eclairage, Vienna.

Gausmann HW, Allen, WA, Escobar, DE, Richardson, AJ and Cardenas, R (1971). Age effects of cotton leaves on light reflectance, transmittance, and absorption and on water content and thickness. Agronomy Journal 43, 465-469.

Graeff, S, Steffens, D and Schubert, S (2001). Use of reflectance measurements for the early detection of N, P, Mg, and Fe Deficiencies in Zea mays L. Journal of Plant Nutrition and Soil Science 164, 445-450.

Graeff, S and Claupein, W (2003). Quantifying nitrogen status of corn (Zea mays L.) in the field by reflectance measurements. European Journal of Agronomy, 19, 611-618.

Hunt, ER (Jr), Rock, BN and Nobel, RS (1987). Measurement of leaf relative water content by infrared reflectance. Remote Sensing of Environment 22, 429-435.

Jackson, RD (1982). Canopy temperature and crop water stress. In ‘Advances in Irrigation’. (Ed. DI. Hille) Vol. 1, pp. 43-85. Academic Press.

Jackson, RD, Slater, PN and Pinter, PJ (1983). Discrimination of growth and water stress in wheat by various vegetation indices through clear and turbid atmospheres. Remote Sensing of the Environment 15, 187-208.

Moran, MS, Pinter, PJ (Jr), Clothier, BE and Allen, SG (1989). Effect of water stress on the canopy architecture and spectral indices of irrigated alfalfa. Remote sensing of environment, 29, 251-261.

Osborne, SL, Schepers, JS, Francis, DD and Schlemmer, MR (2002). Detection of phosphorous and nitrogen deficiencies in corn using spectral radiance measurements. Agronomy Journal 94, 1215-1221.

Penuelas, J, Filella, I, Biel, C, Serrano, L and Savé, R (1993). The reflectance at the 950-970 nm region as an indicator of plant water stress. International Journal Remote Sensing 14, 1887-1905.

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

Previous PageTop Of PageNext Page