Content uploaded by Miltiadis Iatrou
Author content
All content in this area was uploaded by Miltiadis Iatrou on Nov 14, 2016
Content may be subject to copyright.
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3622
REMOTE SENSING FOR THE MANAGEMENT OF
VERTICILLIUM WILT OF OLIVE
George Iatrou1*, Spiros Mourelatos1, Zois Zartaloudis2, Miltiadis Iatrou1,
Sandra Gewehr1, Stella Kalaitzopoulou1
1Ecodevelopment S.A., Thessaloniki, Greece
2Agroecosystem L.P., N. Moudania, Chalkidiki, Greece
ABSTRACT
Verticillium wilt (VW) is a major disease in
all the olive producing countries and leads to
extensive losses of production and olive trees.
There is not chemical treatment available currently
for this pathogen and only preventive cultivation
measures can be applied. In Greece, the most
sensitive cultivars to VW infection are Chondrolia
(Chalkidiki) and Amfissa (Fthiotida, Central
Greece). Previous studies showed that remote
sensing can be used to assess VW infection levels
using spectral indices, such as the Carotenoid
Reflectance Index 2 (CRI2) and Normalized
Difference Vegetation Index (NDVI), which depict
the non-visible and visible stages of the infection,
respectively, suggesting a timely and area wide
method to map the status and spread of the
pathogen in fields grown with Chondrolia cultivar
in Chalkidiki. In the same area, experimental and
also commercial applications of a new Plant
Growth Enhancer Formulation (PGEF) comprising
of natural minerals and biological control agents
provided clear visual field evidence that infected
trees can be brought back in full productivity. The
present work assessed the effect of the application
of the PGEF on the spectral properties of the trees.
The results showed that the application had a
significant effect on the recovery of the olive trees,
as expressed by the change in the NDVI rate of the
treated trees compared to the untreated in two
successive years. Furthermore, the use of CRI2 and
NDVI for the second most important table olive
cultivar in Greece (Amfissa) was validated for the
detection of the early and advanced infection
symptoms, respectively.
KEYWORDS:
Verticillium wilt, Unmanned Aerial Vehicle, remote
sensing, olive, carotenoid reflectance index, NDVI.
INTRODUCTION
The olive groves are quite stable ecosystems
compared to other field crops, behaving like forests.
This is because they enrich the underground water
resources, due to the large root system of the olive
trees, support high biodiversity and soil quality due
to the prevention of soil erosion and constitute an
important feature of the agricultural landscape for
the Mediterranean region. They also provide the
productivity advantages of a crop [1].
VW of olive (caused by the fungus
Verticillium dahliae) is a soilborne disease and
constitutes a major threat for the olive producing
countries [2-5]. It causes serious economic losses
and has environmental impact causing degradation
of the olive agroecosystems. Taking also into
account Xyllela fastidiosa, which has already led to
an olive orchard eradication program in Italy [6],
action should be undertaken in order to prevent a
severe damage to the olive agroecosystems.
Results presented by Zartaloudis et al. [7]
showed that a combination of natural minerals,
organic agents and beneficial microorganisms can
bring VW infected trees back in full productivity.
However, a cost effective application of this
treatment is based on the detection of trees infected
by VW, at the early stages of the infection to
minimize treatment costs [8-9].
An objective of this work is to validate the
CRI2 index for the early stages of the infection by
the disease for the second most important table
olive cultivar in Greece (cv. Amfisssa) in order to
provide an area wide early warning management
tool for the most important Greek table olive
cultivars. In addition, a second objective is to study
the spectral response of the trees to the above
mentioned treatment. Since the rehabilitation of the
olive trees is a process which could take months or
years depending on the stage of the infection,
remote sensing techniques can be used as an early
proof of their recovery process.
MATERIALS AND METHODS
For the validation of the efficiency of the
CRI2 and NDVI indices to detect the early and the
advanced infection levels for cv. Amfissa,
respectively, airborne spectral data were acquired
by UAV from three selected fields (Field 1 Fth,
Field 2 Fth and Field 3 Fth) in Fthiotida on 3rd June
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3623
2015. Field 1 Fth, Field 2 Fth and Field 3 Fth were
planted with 151, 247 and 77 trees (Olea europea
L. cv. Amfissa) and 54, 28 and 26 trees were
visually observed for VW infection. The olive trees
were assessed using an eleven-point calibration
scale [10], where 0-10 infection levels respond to
the following symptoms: healthy trees with non-
visible symptoms of infection (0), light leaf green
discoloration (1), chlorotic canopy (2), flower
cluster or fruit or twig drying (3), branch drying (4),
wilting of the half portion of the entire canopy (5),
a small portion of the entire canopy is healthy (6),
75% of the tree is dead (7), a branch is alive (8), a
twig is alive (9) and dead tree (10). The planting
spacing was variable for the olive groves in
Fthiotida.
Multispectral data were acquired from
Fthiotida plots on 3rd June 2015 using a fixed wing
UAV (ebee, senseFly) and a four-band camera
(multispec 4C, airinov) (Figure 1). Camera's four
spectral bands were modified to detect Verticillium
wilt of olive according to Zartaloudis et al. [11] and
centered at 510, 660, 710 and 790 nm with a
spectral resolution of 10 to 20 nm. The pixel size of
the camera was equal to 4-5 cm on the ground and
the camera's lens-to-focus distance was equal to 13
mm. Olive tree canopies were manually digitized
from the aerial images in order to be separated from
soil pixels and the mean multispectral reflectance
was calculated for each digitized polygon. The
mean reflectance calculated from the 4 bands was
used to calculate the Carotenoid Reflectance Index
2 (CRI2) [12] and the Normalized Difference
Vegetation Index (NDVI) [13] according to the
following equations:
CRI2 = ((1/R510) - (1/R710)) × R790
NDVI = (R790 - R660)/(R790 + R660)
FIGURE 1
An image of the three Field parcels in Fthiotida
acquired by the UAV.
For the assessment of the PGEF application on
the spectral response of the VW infected trees, the
following approach was adopted:
Two commercial olive orchards (Field 1 Ch and
Field 2 Ch) located in Chalkidiki (northern Greece)
were selected. The olive orchards in 2014 had 228
and 92 trees (Olea europea L. cv. Chondrolia
Chalkidiki), respectively, at a spacing of 6 × 5 m
(Figure 2). Heavily infected olive trees were
removed by the grower at the end of the cropping
season in 2014. Thus, the olive orchards in 2015
had 198 and 77 trees, respectively. In May
FIGURE 2
Images of Field 1 Ch (A) and Field 2 Ch (B) acquired by UAV. The treated with PGEF trees are outlined.
A
B
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3624
2014, for the purposes of a previous study [11]
all the trees for Field 1 Ch and Field 2 Ch were
assessed for VW infection by visual observation
according to the 11-point calibration scale. The
trees had various levels of VW infection.
Verticillium dahliae infection was confirmed for the
two fields in Chalkidiki using laboratory
techniques. Infected vascular tissues were sent to a
plant pathology (Aristotle University of
Thessaloniki), were placed on culture mediums,
incubated for a week and then examined
microscopically. In Field 1 Ch and Field 2 Ch, 9
and 20 trees were treated with PGEF in November
2014 and March 2015, respectively. From the 9
treated trees in Field 1 Ch, three trees were healthy
in the assessment of 2014, three trees had infection
level 1 and three trees had infection level 2. In Field
2 Ch, 4, 9 and 7 treated trees had VW infection
levels 0, 1 and 2 in May 2014, respectively. All the
treated trees were assessed by visual observation in
November 2015 and were found to be healthy
(infection level 0).
The PGEF treatment included an application
of a composition of natural minerals, organic agents
and beneficial microorganisms (Bioshell VERT,
Agroecosystem L.P.). Bioshell VERT is a
formulation based on micronized zeolite
(clinoptilolite 96%) at a rate of 64.5%, which
contains non-crystalline (amorphous) silica (24%)
of organic origin, soil beneficial microorganisms
(Mycorrhyzae, plant promoting bacteria such as
Bacillus subtilis and Streptomyces spp., strains of
saprophytic biocontrol fungi, such as Trichoderma
spp. (3.5%), organic acids in the form of humic and
fulvic acids (8%). This formulation acts as plant
growth enhancer and soil conditioner. The solution
prepared for watering the trees was composed of
500 g of Bioshell VERT per tree diluted in 10 liters
of water.
In order to validate the effect of the PGEF on
the olive trees' physiology, airborne multispectral
data were acquired for Field 1 Ch and Field 2 Ch on
19 May 2015 using the ebee and the multispec 4C
camera. The acquired data by the UAV were
compared with spaceborne data acquired from the
imagery archive of DigitalGlobe (WorldView 2
satellite) on 1 July 2014. For the calculation of
NDVI WV2 index the following equation was used:
NDVI WV2 = (RNear-IR1 - RRed)/(RNear-IR1 +
RRed) where Blue: 450-510nm, Red Edge: 705-745
nm, Near-IR1: 770-895 nm and Red: 630-690 nm.
The NDVI index was used for the computation of
the reflectance rate change, because the NDVI is a
normalized index, which means that there was not
much variation from year to year for the same field
and between the spaceborne sensor and the UAV
sensor.
The rate of NDVI change between the NDVI
acquired by the UAV and the NDVI WV2 was
calculated to assess the effect of the PGEF
application on the plant spectral reflectance, as
follows:
r = (log(NDVIuav/NDVIsat)/loge) × 100
The two flights with the UAV for Fthiotida
and Chalkidiki were conducted from 9:00 to 11:00
am (GMT) (11:00 a.m. to 13:00 p.m. local time) at
the same altitude (160 m above ground).
Differences in the reflectance rate change of
olive trees, as determined by the spaceborne and
UAV data in the Chalkidiki experiment, resulting
from the application of the PGEF were analyzed
using unbalanced ANOVA. The significance of
PGEF application × Field × infection level on the
reflectance rate change was tested, which indicated
whether the Field and the infection level factor
affected the influence of the PGEF application on
the reflectance rate change. Multiple regression
analysis with groups was used to study the
correlation of the two multispectral indices (CRI2
and NDVI) with the VW infection levels. Airborne,
spaceborne and infection levels were analyzed
quantitatively by using the unbalanced ANOVA
procedure at a probability level of 0.05. Data
analysis was carried out using the Genstat statistical
package (version 11, VSN International Ltd,
Oxford, UK).
RESULTS AND DISCUSSION
Studies conducted the last years by Calderon
et al. [4-5] and Zartaloudis et al. [8-11] have
pointed out the importance of detecting the infected
olive trees by VW. Verticillium dahliae is a
vascular pathogen that cannot be controlled by
conventional means (fungicides) and there is not
currently monitoring of the problem. There is an
urgent need today for reliable monitoring of the
disease on a tree level basis spatially and timely by
modern means of diagnosis, such as remote
sensing. It is known that, even at the early stages of
VW infection, discoloration of the periphery of the
olive leaf lamina can be observed [9-11]. The
vascular tissues of the trees become plugged due to
the mycelium growth, and thus the water movement
from the roots to the canopy is restricted. This
causes calcium deprivation of the infected tissue
and finally discoloration of the periphery of the leaf
lamina. At advanced stages of the infection the
newer leaves roll downwards or inwards and
progressively become greenish-grey to brown [9-
11]. According to the results of this study the early
infection levels (0 - 3) in a multiple regression
analysis were highly correlated with CRI2
(p<0.001), while the NDVI values were found not
to be significant in the regression model (Figure 3).
The effect of the experimental field plots was
statistically significant and the Field 2 Fth and Field
3 Fth were found to be different compared to Field
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3625
1 Fth (p=0.015 and p=0.012, respectively). The
regression model, which included the CRI2, NDVI,
the early infection levels and the Field factor,
accounted for 39.1% of the total variance. When all
the infection levels were taken into account in the
model (0-9), the regression model accounted for
61.3% of the total variance. Both CRI2 and NDVI
were significantly correlated with the infection
levels (p=0.008 and p=0.006, respectively). The
Field 3 Fth was found to be significantly different
compared to Field 1 Fth and Field 2 Fth. Thus, data
taken by the UAV in Fthiotida (cv. Amfissa) show
that the early non-visible symptoms of the disease
were better detected by the CRI2 index, while the
NDVI accounts for the advanced stages. This is in
accordance with data presented by Zartaloudis et al.
[11] for Chondrolia variety. Considering that
Chondrolia and Amfissa varieties are the two
prevailing varieties of table olive in Greece, the use
of the CRI2 index for the identification of the early
stress by VW is very important for the sustainable
table olive production in Greece.
Gitelson [14-15] introduced two new indices,
the CRI1 (Carotenoid Reflectance Index 1) and
CRI2, for the determination of carotenoid content in
tree species (e.g. beech, chestnut and maple).
Zartaloudis et al. [11] showed that the CRI2 is the
most appropriate spectral index for determining the
early stress of olive trees by VW. The CRI2 is an
index that estimates the carotenoid content using
the mathematical difference of reciprocal
reflectance in the red edge and the reciprocal
reflectance in a band centered at 510 nm. This
difference is also multiplied by reflectance at NIR.
The scope of this index is to remove the effect of
chlorophyll on the estimation of carotenoids, as
carotenoids and chlorophylls absorb in the blue
(400 to 500 nm) and thus it is difficult to estimate
carotenoid concentration independently from
chlorophyll concentration using nondestructive
techniques [14]. According to Gitelson [14] the
band in the red edge is sensitive for determining the
chlorophyll content of the plants, because the depth
of the light penetration is found to be higher in the
red edge compared to blue and red bands (the light
absorption by the chlorophyll is reduced).
For the assessment of the PGEF application
the NDVI of the treated and untreated trees in 2014
and 2015 was calculated. As shown in Figure 4A
the NDVI of the infected trees for Field 2 Ch in
2014 was lower 7.2% (p=0.016) compared to the
healthy trees. For Field 1 Ch the difference hardly
failed to be significant. However, Figure 4B shows
that the NDVI of the treated plants was only 3.7%
(p=0.002) lower compared to the NDVI of the
untreated plants in 2015. The same trend holds for
Field 1 Ch. From these two graphs becomes also
evident that the differences between the NDVI of
the healthy and infected trees was reduced when
treated with PGEF in 2015. To validate this
evidence, the NDVI reflectance rate change
was calculated to determine the plant health
recovery process of the treated plants. It was found
using unbalanced ANOVA that the NDVI
reflectance rate change of the treated with PGEF
was statistically significantly higher compared to
the untreated trees (p=0.005) for the Fields 1 Ch
and 2 Ch. The results also showed that there were
significant effects of the Field (p=0.005) and the
infection level (p<0.001) factors on the NDVI
reflectance rate change. There were not significant
interactions between the factors. Thus, despite that
the NDVI of the treated plants with PGEF was
lower compared to the untreated, obviously because
the treated plants were infected, the NDVI
reflectance rate change of the treated plants was
higher indicating that these plants were in the
recovery process. The increase of the NDVI
reflectance rate of the treated trees with the PGEF
compared to the untreated can probably attributed
to increased mycorrhizal colonization in the olive
tree root system, which is known to provide root
disease resistance [16]. The addition of the
formulation through root watering probably
provided favorable conditions in the soil for the
establishment of mychorrhizae in the plant roots
due to increased soil water holding capacity and ion
exchange reactions around the root system of the
treated plants [17-21]. According to the results of
the present study the PGEF resulted in an increase
of the NDVI values, which is a measure of the
general health and vigorousness of the trees. As a
result, the present study shows that it is possible by
using PGEF to aid the plants recover from VW
infection. Considering also that the trees were
assessed and found to be healthy in November
2015, it can be concluded that the PGEF is a
promising methodology for managing VW of olive,
especially taking into account that there is not
fungicide to control this soilborne disease. A
previous work also by Zartaloudis et al [7] showed
that trees with early symptoms of VW can be
brought back in full productivity quicker compared
to trees with advanced stages of symptoms when
treated with PGEF. The detection of olive trees
with early VW infection stages using the CRI2
index can provide the advantage of the targeted
application of the PGEF. Thus, an economic and
efficient management of VW of olive can be
achieved in an integrated management strategy
using remote sensing and PGEF [22].
CONCLUSION
This study presents preliminary results for the
management of a dangerous disease for the olive
growing. There is currently not much information
in the literature about the management and the
detection of VW of olive using remote sensing.
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3626
FIGURE 3
The effect of the infection levels of olive trees by VW on CRI2 (A, B, C) acquired by UAV for the three
fields in Fthiotida. Error bars display s.e.d.
FIGURE 4
NDVI of healthy and infected trees acquired by spaceborne WV2 data in 2014 (A) and NDVI of healthy
and infected trees treated with PGEF acquired by UAV data in 2015 (B). Error bars display s.e.d.
However, there is urgently a need for developing
methods for managing this disease and as a result
further studies need to be conducted to address
these issues. The results of the present study show
that the CRI2 and NDVI indices are proven to be
appropriate for the detection of early and advanced
stages of VW infection in two of the major table
olive cultivars in Greece. The successful
application of the PGEF on infected trees and the
verification of their recovery process through
spectral reflectance data can provide a management
tool for the table olive growers to cope with VW.
REFERENCES
[1] Loumou, A. and Giourga, C. (2003) Olive
orchards: The life and identity of the
Mediterranean. Agriculture and Human Values,
20: 87–95.
[2] Jiménez-Díaz, R. M., Cirulli, M., Bubici, G.,
del Mar Jiménez-Gasco, M., Antoniou, P. P.
and Tjamos, E. C. (2011) Verticillium wilt, a
major threat to olive production: current status
and future prospects for its management. Plant
Disease, 96(3), 304–329.
A
C
B
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3627
[3] Calderón, R., Navas-Cortés, J. A., Lucena, C.
and Zarco-Tejada, P. J. (2013) High-resolution
airborne hyperspectral and thermal imagery for
early detection of Verticillium wilt of olive
using fluorescence, temperature and narrow-
band spectral indices. Remote Sensing of
Environment, 139, 231–245.
[4] Calderón, R., Navas-Cortés, J. A. and Zarco-
Tejada, P. J. (2015) Early detection and
quantification of Verticillium wilt in olive
using hyperspectral and thermal imagery over
large areas. Remote Sensing, 7(5), 5584.
[5] López-Escudero, F. and Mercado-Blanco, J.
(2011) Verticillium wilt of olive: a case study
to implement an integrated strategy to control a
soil-borne pathogen. Plant and Soil, 344(1-2),
1–50.
[6] European Commision (2015) Member States
endorse reinforced measures at EU level to
prevent the spread of Xylella Fastidiosa.
http://europa.eu/rapid/press-release_IP-15-
4887_el.htm
[7] Zartaloudis, Z. D., Ioannidis, F. and Mokiou, S.
(2010) Integrated management of Verticillium
dahliae in olive growing for Municipality of
Chalkidiki, 15th PanHellenic Plant Pathology
Conference, Corfu, 5-8 October (in Greek).
[8] Zartaloudis, Ζ. D., Savvidis, G., Savvidis, K.,
Iatrou, Μ., Glavenas, D., Theodoridou, S.,
Kalogeropoulos, K. and Kiparissi, S. (2015)
Innovative services for diagnosis and
management of agroecosystem pathology
agents, 2(1), 1–9.
[9] Iatrou, M., Zartaloudis, Z.D., Savvidis, G.,
Savvidis, K., Glavenas, D., Theodoridou, S.,
Kalogeropoulos, K. and Kyparissi, S. (2014)
Relationship between calcium concentration in
leaves and early stress in olive trees by
Verticillium wilt infection, verified by remote
sensing data. 17th PanHellenic Soil Science
Conference, Patra, 27-28 November.
[10] Zartaloudis, Z. D., Iatrou, M., Savvidis, G.,
Savvidis, K., Glavenas, D., Theodoridou, S.,
Kalogeropoulos, K. and Kyparissi, S. (2014) A
new 11-point calibration scale of Verticillium
wilt of olive verified by thermal remote sensing
and plant analysis, 17th PanHellenic Plant
Pathology Conference. Volos, 13-17 October
2014.
[11] Zartaloudis, Z. D., Iatrou, M. Savvidis, G.,
Savvidis, K., Glavenas, D., Kalogeropoulos, K.
and Kyparissi, S. (2015) Early and Timely
detection of Verticillium dahliae in olive
growing using remote sensing. El Aceite de
Oliva, Actas Simposio Expoliva 2015, Jaen,
Espana, 6-8 Mayo.
[12]Gitelson, A. A., Zur, Y., Chivkunova, O. B.
and Merzlyak, M. N. (2002) Assessing
carotenoid content in plant leaves with
reflectance spectroscopy.
Photochemistry and Photobiology, 75(3), 272–
281.
[13] Rouse, J.W., Jr Deering, D.W., Schell, J.A. and
Harlan, J.C. (1974) Monitoring the vernal
advancement and retrogradation (green wave
effect) of natural vegetation. Greenbelt, MD.
[14] Gitelson, A. A. (2012) Nondestructive
estimation of foliar pigment (chlorophylls,
carotenoids, and anthocyanins) contents:
evaluating a semianalalytical three-band
model. Pages 141-165 in: Hyperspectral
remote sensing of vegetation. P.S. Thenkabail,
G.J. Lyon, A. Huete, eds. Boca Raton, FL,
CRS Press.
[15] Penuelas, J., Baret, F. and Filella, I. (1995)
Semi-empirical indices to assess
carotenoids/chlorophyll a ratio from leaf
spectral reflectance. Photosynthetica, 31, 221-
230.
[16] Tarkalson, D. D., Jolley, V. D., Robbins, C. W.
and Terry, R. E. (1998) Mycorrhizal
colonization and nutrient uptake of dry bean in
manure and compost manure treated subsoil
and untreated topsoil and subsoil. Journal of
Plant Nutrition, 21(9), 1867–1878.
[17] Brechelt, A. (1990) Effect of different organic
manures on the efficiency of VA mycorrhiza.
Agriculture, Ecosystems & Environment,
29(1–4), 55–58.
[18] Weber M. A., Barbarick, K.A. and Westfall, D.
G. (1983) Ammonium adsorption by a zeolite
in a static and a dynamic system. Journal of
Environmental Quality, 12, 549-552.
[19] Rehakova, M., Cuvanova, S., Dzivak, M.,
Rimar, J. and Gaval'ova, Z. (2004) Agricultural
and agrochemical uses of natural zeolite of the
clinoptilolite type. Current Opinion in Solid
State & Materials Science, 8, 397-404.
[20] Yasuda, H., Takuma, K., Mizuta, N. and
Nishide, H. (1995) Water retension variety of
dune sand due to zeolite addition. Bulletin,
Faculty of Agriculture, Totorri University, 48,
27-34.
[21] Gul, A., Erogul, D. and Ongun, A. R. (2005)
Comparison of the use of zeolite and perlite as
substrate for crisp-head lettuce. Scientia
Horticulturae, 106, 464-471.
[22] Mercado-Blanco, J. (2012). Biological control
of Verticillium wilt of olive within an
integrated disease management framework.
IOBC/WPRS Bulletin, 79, 149-154.
© by PSP Volume 25 – No. 4/2016, pages xxx-xxx Fresenius Environmental Bulletin
3628
Received: 10.12.2015
Accepted: 13.06.2016
CORRESPONDING AUTHOR
George Iatrou
Ecodevelopment S.A. 57010, Filyro Thessaloniki,
Greece
E-mail: iatrou@ecodev.gr