Use of remote sensing to detect virus infected wheat

0 downloads 0 Views 6MB Size Report
aerial photography, and viral contents were quantitated by DAS-ELISA of randomly taken samples. Analysis of the results showed a direct relationship between ...
Vol. 31 Nos. 1-2 2003

Cereal Research Communications

Use of remote sensing to detect virus infected wheat plants in the field R. Gáboijányi1, L. Pásztor2, M. Papp3, J. Szabó2, Á. Mesterházy3, T. Németh2, T. Kőmíves4 1

Plant Protection Institute, Veszprém University, Georgikon Faculty of Agricultural Sciences, Keszthely, H- 8360. Fax: 36-83-315-105, email: [email protected] 2 Research Institute for Soil Science and Agrochemistry, Budapest, H-1525, P.O. Box, 3 Cereal Research Non-profit Company, Szeged, H-6701 P.O. Box 391. 4 Plant Protection Institute, Hungarian Academy of Sciences, Budapest, H-1525 P.O. Box 102.

ABSTRACT Applicability of remote sensing to detect virus infection of wheat was studied under field conditions. The experiments were carried in a wheat variety experimental field in Szeged, Hungary, continuously exposed to natural infection by viruses from sowing to harvest. Plants were evaluated for virus infections by visual inspection and aerial photography, and viral contents were quantitated by DAS-ELISA of randomly taken samples. Analysis of the results showed a direct relationship between the extent of yellowing of the leaves and the rate of Wheat dwarf bigemovirus (WDV) and Barley yellow dwarf luteovirus (BYDV) infection, but not with that of other wheat pathogenic viruses, such as Wheat streak mosaic tritimovirus (WSMV), Brome mosaic bromovirus (BMV), and Barley stripe mosaic hordeivirus (BSMV). Key words: remote sensing, virus infection, wheat, Wheat dwarf bigemovirus INTRODUCTION Remote sensing and image analysis in agriculture is gradually taking root in areas of Europe. Methods of site-specific, so-called precision agriculture enables growers to improve efficiency, reduce pesticide and fertilizer use, and optimize yields (Reisinger et al., 2001). Although remote sensing can be used in agriculture and forestry for surveillance of different environmental stresses (Nilsson, 1995; Szabó et al., 1998; Tóth et al., 1998), little is known about the usability of this technique in detecting virus infections. This paper reports on the results of a three-year study on virus infections of wheat plants by using remote sensing and image analysis. MATERIALS AND METHODS The experiments were carried in a wheat variety experimental field, called "virus nursery" in Szeged, Hungary, exposed to natural virus infection from sowing to harvest. Wheat seeds were sown in late September (two weeks earlier than usual) of 1999-2001 in a wide spaced design (10 m long rows, 50 cm row distance and 10 cm intervals between plants) by Wintersteiger plot spider. This arrangement helped the vectors (viruliferous aphids, leafhoppers, mites, and beetles) to transmit the viruses. Plots were evaluated by visual inspection, as well as by aerial photography (from a low-flying airplane and an 113

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

GÁBORJÁNYI et al.:

VIRUS INFECTION IN WHEAT

aerial platform crane-truck). According to the field reference classes, two sampling strata were distinguished based on the brightness (intensity) of the digital image scanned from the panchromatic picture of the aerial photograph. One stratum represents the green the other the yellow(ish) crop mass. Color aerial photographs were spectrally decomposed according to traditional/typical RGB additive filters. R(ed): 600-700 nm, G(green): 500600 nm, B(lue): 400-500 nm. Viral contents of samples randomly taken at the time of visual inspection and photography were quantitated by DAS-ELISA. Serological assays were carried out by ELISA (DAS-ELISA) tests, according to Clark and Adams (1977). Antisera of wheatpathogenic viruses were purchased from Loewe Biochemicals Ltd. (Germany). Absorbance values were measured at 405 nm using a Labsystem Multiscan MS spectrophotometer. RESULTS AND DISCUSSION Preliminary studies on virus infections in the experimental field in Szeged were carried out in 1994-1999. High abundance of and complex infections by different wheat pathogenic viruses were detected (Papp et al., 1996; Mesterhazy et al. 2002). The most important viruses of the region are WDV and BYDV (Bisztray et al. 1989; Pocsai et al., 1995). These viruses cause dwarfing, yellowing and reddening of the plants. Papp et al. (1996) and Mesterhazy et al. (2002) demonstrated that importance of BYDV was previously overestimated and* WDV has a higher significance. Especially high infection rates with WDV were found in year 1998, while other viruses were not so abundant and did not influence the growth and the appearance of the plants. These findings prompted us to study the applicability of remote sensing in the detection of virus infection in wheat, and its use in the evaluation of wheat cultivars to virus infection tests. Before sowing of the wheat seeds aerial photographs were taken of the experimental fields to detect sites of soil inhomogeneity. The soil was found to be sufficiently homogenous: in areas of altered color (indicating lower humus content) growth of the wheat plants and spread of the viruses were not significantly different (data not shown). A representative segment of the experimental field (about one fifth of the total area) together with the sites of randomly taken samples are shown on Figure 1. We examined the intensity distribution of the image in each band within the two strata. Homogeneity analysis was carried out for the comparison of the pixel value distribution within the distinguished strata. Significant (p > 99.9%) differences were experienced in each bands both using chi-squre and Kolgomorov-Smirnov tests. In year 2001 photographs were taken from an aerial platform crane-truck in order to distinguish virus susceptibility of different varieties, and even individual plants. A general survey of the virus infections showed that the abundance of the viruses varied highly during the 2000-2001 period (Table 1). The weather in year 2000 was favorable for aphids and leafhoppers, which caused a high abundance of BYDV and WDV, and, as a result, numerous cases of complex infections were detected (Mesterhazy et al. 2002). The summer and the autumn of year 2000 were extremely dry, which also supported the spreading of these aphid- and leafhopper-born viruses. 114

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

GÁBORJÁNYI et al.:

VIRUS INFECTION IN WHEAT

Figure 1. Photograph of the experimental field and sites of randomly taken samples. The most abundant viruses (BYDV and WDV) caused yellow symptoms. Others (WSMV, BMV, BrStV and BSMV) occurred only in lower percent and did not induce yellowing of the infected plants. BrStV infection was studied only in 2001, but not in 2000.

Table 1. General abundance of wheat pathogenic viruses in years 2000 and 2001 (detected by serological assays of 207 samples) Years 2000 2001

BYDV* 32 16

WDV* 39 21

WSMV* 22 3,5

BMV* 37 3,5

BStV* -

1

BSMV* 22 1,5

* Barley yellow dwarf luteovirus - BYDV; Wheat dwarf bigemovirus - WDV; Wheat streak mosaic tritivirus - WSMV; Brome mosaic bromovirus - BMV; Barley streak mosaic tritivirus - BStV; Barley stripe mosaic hordeivirus - BSMV

The study also aimed at the reliability of the results of visual inspections and remote sensing in the function of virus content. For this reason in year 2000 more than eighty visually characterized "yellow" (supposedly diseased) and healthy looking "green" individual samples were collected and analyzed by serological tests. Rates of yellowing 115

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

GÁBORJÁNYI et al.:

VIRUS INFECTION IN WHEAT

of the leaves (evaluated by visual inspection) linearly correlated with rates of WDV infections (Tables 3-6). BYDV also caused yellowing of the leaf (indistinguishable from that caused by WDV) but its incidence was low, therefore it did not influence the correlation. Our findings also show the limitations of remote sensing to detect virusinfected plants: green plants can be infected with several viruses, which do not cause yellowing symptoms (Tables 3-6). Table 2. Wheat dwarf bigemovirus (WDV) content of "y e H° w " and "green" samples Symptom/ No. of WDV Reliability Infection infected plants % Yellow and 2 8

Symptom/ Infection Green and -

Yellow and +*

6

25

Green but +

1

9

Yellow and 2+ Yellow and 3+ Total"+" Total No.

5

21

Green but 2+

0

0

11

46

Green but 3+

4

36

22 24

92 100

Total "+" Total No.

5 11

45 100

No. of WDV Reliability infected plants % 6 55

*Reactions = + low, ++ medium and +++ high virus concentrations according to the extinction's at 405 nm. Positive reaction is three times higher than negative control.

Table 3. Barley yellow dwarf luteovirus (BYDV) content of "yellow" and "green" samples Symptom/ No. of BYDV Infection infected plants 19 Yellow and

Reliability % 79

Symptom/ No. of BYDV Reliability % Infection infected plants 91 10 Green and -

Yellow and +*

2

8

Green but +

0

0

Yellow and 2+ Yellow and 3+ Total"+" Total No.

0

0

1

9

3

13

0

0

5 24

21 100

Green but 2+ Green but 3+ Total "+" Total No.

1 11

9 100

*Reactions = + low, ++ medium and +++ high virus concentrations according to the extinction's at 405 nm. Positive reaction is three times higher than negative control.

116

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

GÁBORJÁNYI et al.:

VIRUS INFECTION IN WHEAT

Table 4 Brome mosaic bromovirus (BMV) content of "yellow" and "green" samples Symptom/ No. of BMV Infection infected plants Green and 21 Yellow but 1 +* Yellow but 2+ Yellow but 3+ Total"+" Total No.

Reliability % 88 4

2

8

0

0

3 24

13 100

Symptom/ No. of BMV Reliability Infection infected plants % Green and 8 73 Green but + 2 18 Green but 2+ Green but 3+ Total "+H Total No.

1

9

0

0

3 11

27 100

*Reactions = + low, ++ medium and +++ high virus concentrations according to the extinction's at 405 nm. Positive reaction is three times higher than negative control.

Table 5. Barley stripe mosaic hordeivirus (BSMV) content of the "yellow" and "green" samples Symptom/ No. of BSMV Infection infected plants Green and 13 9 Yellow but +* Yellow but 2+ Yellow but 3+ Total H+M Total No.

Reliability % 54 38

1

4

1

4

11 24

46 100

Symptom/ No of BSMV Reliability Infection infected plants % Green and 64 7 36 Green but + 4 Green but 2+ Green but 3+ Total"+" Total No.

0

0

0

0

4 11

36 100

*Reactions - + low, ++ medium and +++ high virus concentrations according to the absorbcmces at 405 nm. Positive reaction is three times higher than negative control.

117

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

VIRUS INFECTION IN WHEAT

GÁBORJÁNYI et al.:

Table 6. Wheat streak mosaic tritimovirus (WSMV) content of"yellow" and "green" samples. Symptom/ No. of WSMV Infection infected plants Green and 21 Yellow but 3 +* Yellow but 2+ Yellow but 3+ Total "+" Total No.

Reliability %

Symptom/ No of WSMV Reliability Infection infected plants % Green and 11 100 Green but + 0 0

88 13

0

0

0

0

3 24

13 100

Green but 2+ Green but 3+ Total "+" Total No.

0

0

0

0

0 11

0 100

*Reactions = + low, ++ me(hum and +++ high virus concentrations according to the absorbances at 405 nm. Positive reaction is three times higher than negative control.

Table 7. shows a repetition test made in 2001. This year samples were distinguished in three groups. By the so-called green and yellow samples, visually infected "streak" symptoms were also checked to evaluate the liability of visual infection. Table 7. Reliability of visual evaluation and virus infestation in year 2001. Symptoms/ Viruses

BYDV

WDV

Yellow Green Streak

15 4 12,5

65 4 12,5

Visually correct % 78 92 6

Visually false % 21 8 94

These results show that yellow symptoms can be characteristic for both BYDV and WDV, and liability of detection of infection based on symptoms is relatively high, if we do not want to differentiate between these two viruses. In leaf samples with "streak" symptoms BYDV and WDV dominated. WSMV was detected only in a single occasion. It is clear that the streak symptoms were due not only to WDV or to BYDV infection, but by other pathogens, mainly by WSMV could cause these symptoms. According to our previous studies, WSMV was a characteristic pathogen in South Hungary, the virus was described first in Hungary especially from this area (Pocsai and Barabas, 1985; Gaborjanyi and Nagy 1988). On the other hand, 92% of the "green" samples were free of WDV and BYDV virus, and infected by WDV and by BYDV by 4% each. These false results could be due to recent infections by these two pathogens, when the typical yellows symptoms have not appeared yet. Because of the drought during the fall of year 2000 seeds of certain winter wheat varieties germinated poorly, leaving empty plots in the experimental field. These 118

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

VIRUS INFECTION IN WHEAT

GÁBORJÁNYI et al.:

conditions were very favorable for studying virus infections of individual plants in the field. Photographs taken from an aerial platform crane-truck allowed clear distinction between green and yellow plants. A study of virus content showed the abundance of WDV (Table 1), and supported the observations of the previous year. CONCLUSIONS We conclude that remote sensing is a useful technique to detect infections of wheat with two wheat pathogenic viruses that cause serious yield losses in Hungary (WDV and BYDV). Panoramic features of space imagery can be useful for the delineation of infected regions and determination of the affected area. Actual spatial distribution of infection could be determined, forwardness of infection could be monitored, actual regional extension could be estimated. The extreme warm and dry falls and winters of years 1999-2001 supported the migration and overwintering of virus vectors (primarily aphids and leafhoppers). Therefore, in order to prevent yield losses it is very important to detect virus infections in wheat fields and to determine the susceptibility of different wheat varieties to virus infections. Other remote sensing methods were suitable to detect WDV and BYDV infected plants, but the observation from the aerial platform cranetruck was found to be the most convenient and the least expensive. Acknowledgements We acknowledge with thanks the help of Dr. Zoltán Kertész of the Cereal Research Non-profit Co., Szeged, Hungary, for making the institute's experimental field available for this study. This work was supported by grants for OTKA (T-029121) and OM (NKFP-4/008 and NKFP-4/037). REFERENCES Bisztray, G.- Gáboijányi, R.- Vacke, J. 1989: Isolation and characterisation of wheat dwarf virus found first time in Hungary. Z. Pflanzenkrankheiten und Pflanzenschutz 96: 449. 454. Gáboijányi, R. and Nagy, P.D. 1988: Wheat streak disease in Hungary. Növénytermelés 37: 391-395. Mesterházy, Á.- Gáboijányi, R.-Papp, M.- Fónad. P. 2002: Multiple virus infection of wheat in South Hungary. Cereal Res. Comm. 30: 329-334. Papp, M.- Mesterházy, Á.- Vasdinnyei, R.- Gáboijányi, R. 1996: Mixed virus infection of wheat in South East Hungary in 1994 and 1995. Cereal Res. Comm. 24: 179-192. Pocsai, E. and Barabás, Z. 1985. Identification of wheat streak mosaic virus in Hungary. Növényvédelem 9: 411. (Abstract in Hungarian)

119

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions

VIRUS INFECTION IN WHEAT

GÁBORJÁNYI et al.:

Pocsai, E. - Kovács G. - Murányi, I. - Orosz, Á. - Papp, M. - Szunics, L. 1995: Differentiation of barley yellow dwarf luteovirus serotypes infecting cereals and maize in Hungary. Agronomie 15: 401-408. Nilsson, H.-E. 1955: Remote sensing and image analysis in plant pathology. Annu. Rev. Phytopathol. 33: 489-527 Reisinger, P. - Kőmíves, T. - Lajos, M. - Lajos, K. - Nagy S. 2001. Visualization of spreading of noxious weeds within an agricultural field by using GPS-based maps (in Hungarian). Hungarian Weed Research and Technology. 2: 25-32. Szabó J., Pásztor L., Suba Zs., Várallyay Gy., (1998) Integration of remote sensing and GIS techniques in land degradation mapping. Agrokémia és Talajtan 47: 63-75. Tóth, T., Pásztor L., Kertész ML, Zágoni B., Bakacsi Zs. 1998: Allocation of soil reclaiming material based on digital processing of aerial photograph. In. International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, pp 178-181.

Received 22 May, 2002, accepted November 8, 2002

120

This content downloaded from 193.224.158.254 on Wed, 06 May 2015 07:55:47 UTC All use subject to JSTOR Terms and Conditions