spectral reflectance indices as a rapid and nondestructive...

1

http://journals.tubitak.gov.tr/agriculture/

Turkish Journal of Agriculture and Forestry Turk J Agric For(2015) 39: © TÜBİTAKdoi:10.3906/tar-1406-164

Spectral reflectance indices as a rapid and nondestructive phenotyping tool for estimating different morphophysiological traits of contrasting spring wheat germplasms

under arid conditions

Salah EL-HENDAWY1,2,*, Nasser AL-SUHAIBANI1, Abd El-Azeem SALEM1,3,Shafiq UR REHMAN4, Urs SCHMIDHALTER5

1Department of Plant Production, College of Food and Agriculture Sciences, King Saud University, Riyadh, Saudi Arabia2Department of Agronomy, Faculty of Agriculture, Suez Canal University, Ismailia, Egypt

3Department of Field Crops Research, National Research Center, Giza, Egypt4Department of Botany, Kohat University of Science & Technology, Kohat, Pakistan

5Department of Plant Sciences, Technische Universität München, Freising-Weihenstephan, Germany

* Correspondence: [email protected]

1. IntroductionArid and semiarid regions are seriously lacking in fresh water. Water shortages in these regions have become the basic norm rather than the exception. Most importantly, the situation of water shortage is growing worse due to abrupt climatic changes and continuous population growth. All of these factors will decrease the amount of water allocated to the agricultural sector, which consumes about 75% of the available water supply. Therefore, as the water supply for agronomic purposes becomes insufficient, development of new germplasms with higher yield potential becomes more imperative and it will be one of the major adaptation strategies to sustain crop productions under arid and semiarid conditions. In

order to improve the germplasms for these conditions in breeding programs, it is essential to evaluate a large number of germplasms using multiple selection criteria. Several morphophysiological traits, particularly those related to crop processes, yield characteristics, and drought-tolerance mechanisms such as the water status of plants, photosynthetic efficiency, stomatal conductance, canopy temperature, accumulation of dry matter, stay-green character of leaves, harvest index, moisture stress index, and water use efficiency, are usually effective as useful complementary selection criteria for screening germplasms under different environmental conditions (Blum, 2005; Royo et al., 2005; Hura et al., 2007; Elsayed et al., 2011; Chen et al., 2012; Bürling et al., 2013). However,

Abstract: The aim of this study was to investigate whether spectral reflectance indices could be used to estimate different destructive morphophysiological traits of a wide and diverse range of spring wheat germplasm in a rapid and nondestructive manner. A total of 90 spring wheats germoplasms were evaluated under water shortage by applying only three irrigations during the growing cycle of germplasm with the amount of water totaling 2550 m3 ha–1. Ten selected spectral reflectance indices were related to the green leaf number per plant, green leaf area per plant, total dry weight per plant (TDW), grain yield per hectare (GY), leaf water content (LWC), leaf area index (LAI), and canopy temperature (CT). Significant genotypic variability was shown for all morphophysiological traits and the ten selected spectral reflectance indices. The broad-sense heritability of the normalized water index (NWI)-3, NWI-4, water band index (WBI), and normalized difference vegetation index (NDVI) was high to medium as reflected in the morphophysiological traits, while for other spectral reflectance indices it was low. The indices NWI-3 and NWI-4 proved to be better predictors for LWC, GY, and LAI than NWI-1 and NWI-2. Spectral indices based on combine visible and near-infrared wavelengths such as the NDVI, the ratio of WBI/NDVI, and the R940/R960/NDVI were viable options to estimate TDW, GY, and LAI, whereas the WBI and R1000/R1100 had the best fit to LWC. The R940/R960 index failed to capture the genotypic variability in any morphophysiological traits. The LAI was more correlated to and had more direct effects on all agronomic traits than CT. The overall results indicated that it is indeed possible to apply spectral reflectance tools in wheat breeding programs to estimate the destructive morphophysiological traits and assess genotypic variability of a large number of germplasms in a rapid and nondestructive manner.

Key words: Canopy reflectance, canopy temperature, leaf area index, morphophysiological traits, phenomics

Received: 29.06.2014 Accepted: 16.01.2015 Published Online: 00.00.2015 Printed: 00.00.2015

Research Article

2

EL-HENDAWY et al. / Turk J Agric For

direct measurements of those traits by traditional methods are destructive and time-consuming and some of them are difficult to do when a large number of genotypes need to be evaluated across different environments. For example, although the water status of plant is a well-established phenotypic trait that can be used to prevent crop water deficit through irrigation and to select the best genotypes adapted to abiotic stress, the determination of this trait by destructive methods is impractical and unaffordable when a large number of genotypes have to be evaluated (i.e. ca. 20–30 samples per hour using a Scholander pressure chamber) (Jongdee et al., 2002; El-Hendawy et al., 2005; Munjal and Dhanda, 2005; Gutierrez et al., 2010). Thus, the main goal of recent innovative breeding strategies is to look for easy, rapid, inexpensive, and nondestructive tools to assess agronomic and physiological traits of a large number of germplasms in a relative short time.

The high-throughput phenotyping techniques possess great potential for indirect assessments of morphophysiological traits for large-scale evaluation of germplasms in a rapid and nondestructive manner (Cobb et al., 2013; Erdle et al., 2013a, 2013b; Araus and Cairns, 2014; Kipp et al., 2014a, 2014b). This technique is based mostly on spectral reflectance information from leaves and canopies in the visible (VIS) and near-infrared (NIR) wavelengths and uses this information to build a number of spectral reflectance indices by using simple mathematical formulas (e.g., ratios or differences in the reflectance at given wavelengths) (Araus et al., 2001). Fortunately, several spectral reflectance indices have successively been used to estimate different morphophysiological traits such as green biomass, green leaf area, photosynthetic efficiency, plant vigor, chlorophyll content, water status of plants, and grain yield in different crops (Royo et al., 2007; Berger et al., 2010; Mistele and Schmidhalter, 2010; Cabrera-Bosquet et al., 2011; Kipp et al., 2014a, 2014b). For example, significant relationships between the spectral reflectance index ((R940/R960)/NDVI) and the leaf water potential were detected in wheat and maize, with determination coefficient values ranging between 0.74 and 0.92 (Elsayed et al., 2011). The simple water index (WI, R970/R900) showed significant correlation with the relative water content under water-stressed conditions in Phaseolus vulgaris L. (Peñuelas et al., 1993). Gutierrez et al. (2010) reported that the normalized water index (NWI)-3 ((R970 – R880)/(R970 + R880)) was successfully used to detect genotypic differences in the water status at the canopy and soil levels in contrasting wheat genotypes. Under salinity conditions, the normalized difference vegetation index (NDVI) and the spectral reflectance ratio R780:R550 showed significant correlation with the fresh weight, the water content of the aboveground biomass, and the water potential of the youngest fully developed leaf in wheat

(Hackl et al., 2013). Furthermore, a number of optimized indices have been developed to better estimate various agronomic parameters such as plant biomass, plant vigor, biomass partitioning between spikes and vegetative plant parts, and final grain yield (Royo et al., 2004; Babar et al., 2006; Teal et al., 2006; Winterhalter et al., 2011; Erdle et al., 2013a, 2013b; Kipp et al., 2014b).

In addition, infrared thermography and plant canopy analyzers were successfully applied as an effective noncontact, high-throughput tool to estimate genotypic variability in water status at the canopy and soil levels (Araus et al., 2008; Gutierrez et al., 2010; Pask and Reynolds, 2013; Zia et al., 2013). These devices were developed to indirectly measure canopy temperature (CT) and leaf area index (LAI), respectively. In summary, the germplasms that exhibit lower CT will have the ability to extract water from deeper in the soil profile and exhibit a higher transpiration rate to increase adiabatic cooling (Jones, 1999; Feng et al., 2009; Mutava et al., 2011). Furthermore, the germplasms that have the ability to cover the soil early will have the ability to reduce the amount of water lost from the ground and thus improve the water status of the plants (Pask and Reynolds, 2013). Therefore, CT and LAI have been used as indirect indicators to evaluate a large number of germplasms for grain yield, capacity of roots to access available soil water, agronomic water use efficiency, and the water status of plants under field conditions and in many field crops.

The objective of this study was to test the ability of different spectral reflectance indices as a rapid and nondestructive high-throughput indirect tools to estimate various morphophysiological traits in a wide and diverse range of spring wheat germplasms growing under irrigation water shortage in the distinctive arid Arabian Desert conditions.

2. Materials and methods 2.1. Plant material A total of 90 spring wheat germplasms, comprising 22, 34, and 30 F4:6 and F4:7 recombinant inbred lines from the crosses Sids 1/Sakha 61, Sids 1/Sakha 93, and Sakha 93/Sakha 61, respectively, and three parents and one drought-sensitive cultivar (Yecora Rojo) were grown under irrigation water shortage in the years 2012/2013 (F6) and 2013/2014 (F7). The three parents used in the crosses, Sakha 93, Sakha 61, and Sids 1, were characterized as tolerant, moderately tolerant, and sensitive to moisture stress, respectively (Abd El-Kareem and Saidy, 2011). The 90 wheat germplasms were selected to represent a range of genetic diversity.2.2. Experimental conditionsField experiments were conducted at the Agricultural Research Station of King Saud University (Dierab, near

3


Riyadh; 24°25′N, 46°34′E, 400 m a.s.l.). The study area is mostly sunny and dry during the growing cycle of wheat crops. The soil texture at the experimental station is a loamy sand (82.4% sand, 9.5% silt, and 8.1% clay), low in organic matter, with a plant-available water retention capacity of about 120 mm m–1 and slightly alkaline (pH 7.9) in nature. The daily values of mean temperature, mean relative humidity, and evapotranspiration rate at the experimental station during the two growing seasons are summarized in Figure 1.

The field experiments were designed as a randomized block design with three replications. Each germplasm was sown at 300 seeds per square meter in a six-row plot. The

plot size was 4 m in length and 1.2 m in width. All essential nutrients, including N, P, and K, were adequately applied to avoid any nutrient deficit. Weeds and diseases were controlled throughout the growing season.

Only three irrigations with the base irrigation were applied during the growing cycle of the germplasms. The base, first, second, and third irrigations were applied during the seedling (ZS 15), tillering (ZS 25), heading (ZS 59), and complete emergence of florescence (ZS 69) growth stages (Zadoks et al., 1974), with the amount of water totaling 2550 m3 ha–1. Irrigation was provided via the furrow method. The irrigation system had one water-emitting tube for each plot to deliver constant and equal amounts of water to each plot. The amount of water was monitored with a discharge gauge and regulated through manually operated control valves.2.3. Spectral reflectance measurementsCanopy reflectance was measured under clear sky conditions between 1000 and 1400 hours using a portable FieldSpec spectroradiometer (Analytical Spectral Devices, Boulder, CO, USA). This instrument was capable of detecting reflectance light from 350 to 2500 nm wavelengths with a sampling interval of 1.0 nm of the spectrum. Thus, 2151 continuous bands were obtained at each measuring. Reflectance measurements were taken at a height of 50 cm above the canopy in nadir position with 25° field-of-view fiber optics at a vertical position after the spectroradiometer was calibrated using a white reference panel of barium sulfate (BaSO4). Four spectral measurements were taken from four different places in each plot and the mean of four readings was used to calculate the spectral indices of each individual plot. Readings were taken at heading growth stage (ZS 59). Because there was a difference of 2 to 8 days among the tested germplasms in reaching ZS 73, the readings were taken at the middle of this range to ensure minimal influence on the readings. All possible spectral reflectance indices were evaluated within this study using different combinations of VIS/NIR wavebands as ratio and/or normalized indices, but only the best performing ones are presented in Table 1. 2.4. Morphophysiological measurementsFive plants from each replicated germplasm set were taken almost synchronously with the spectral measurements to determine the number of green leaves per plant (GLN), the leaf area of green leaves per plant (GLA), and the total dry weight per plant (TDW). The GLN included all leaves of a plant, excluding leaves that were completely brown. The GLA was measured as the maximum width × maximum length × 0.75 for green leaves (Francis et al., 1969). The TDW was obtained from the weight of oven-dried (80 °C for 72 h) plant material.

The LAI, CT, and relative water content were also measured immediately following the spectral

C

ETo

(mm

day

–1)

0123456789

101112131415

2012–20132013–2014

Days after sowing

1 11 21 31 41 51 61 71 81 91 101 111 121 131 141

B

Dai

ly m

ean

rela

tive h

umid

ity (

%)

20

30

40

50

60

70

80

90

10A

Dai

ly m

ean

tem

pera

ture

(°C

)

0

5

10

15

20

25

30

35

Base irrigationFirst irrigation

Measurement datesSecond irrigation

Third irrigationHarvest

Figure 1. Daily values of mean temperature (A), mean relative humidity (B), and evapotranspiration rate ETo (C) at the experimental station during the growing periods of wheat in 2012/13 and 2013/14.

4


measurements. The LAI was measured with an LAI-2000 Plant Canopy Analyzer (LI-COR Inc., Lincoln, NE, USA) using the method of Hicks and Lascano (1995). The CT was measured using thermal infrared imaging (Therma CAM SC 3000 infrared camera, FLIR Systems, Wilsonville, OR, USA) with a wavelength range of 8–9 µm, 320 × 240 pixels, 45° × 34° lens of field view, and built-in 20° lens. The camera was held at a height approximately 1.0 m above the top of the ear and the CT analysis of the infrared images was performed using FLIR Quick Report 1.2 SP1 (FLIR Systems). The relative water content was determined on flag leaves by taking 7–10 cm2 from four flag leaves and immediately determining the fresh weight (FW). The leaf samples were then rehydrated in deionized water at 25 °C until they were fully turgid and then blotted dry, and the turgid weight (TW) was determined. Finally, the leaf samples were dried at 80 °C in an oven until no further change in dry weight (DW) was observed. The relative water content was calculated using the following equation:

RWC (%) = ([FW – DW] / [TW – DW]) × 100.After physiological maturity, the total grain yield per

hectare was determined by harvesting and threshing an area of 3 internal rows, each 3 m in length (1.8 m2 in total area), from each plot. The grain yield was adjusted to a water content of 15.5%.2.5. Data analysisData were analyzed using SPSS 21.0 (SPSS Inc., 2012). Broad sense heritability (H2) was estimated for each trait individually using means of each germplasm across two years as H2 = σ2

g/σ2

p, where σ2g and σ2

p indicate genotypic and phenotypic variance, respectively.

Genetic correlations were estimated by combining two years as the ratio of the genetic covariance between two traits and the square root of the product between genetic variance of the two traits. The formula used to estimate genetic correlation was:

rg = (Covxy)/√(VarxVary),where Cov and Var indicate components of covariance and variance between two traits, respectively.

The genetic covariance was estimated using the statistical property of the sum of two random variables using the following equation:

σ2(x + y) = σ2

x + σ2y + 2σ2

xy, where covariance is expressed by σxy.

Path analysis was used to measure both the direct and indirect effects of traits. Path coefficient analysis was performed based on logical relationships between morphophysiological traits and spectral indices and taking morphophysiological traits as a dependent characteristic and the spectral indices as causal. Multiple linear regression analyses were conducted with SPSS 21.0.

3. Results3.1. Genotypic variability and response of traits to water shortage ANOVA analysis revealed significant variation among wheat germplasms for all morphophysiological traits and spectral reflectance indices except NWI-1, NWI-2, and the R1000/R1100 ratio (Tables 2 and 3). Genotypic variation was significant for traits at P = 0.05 or higher (Table 3). The results also display a wide range between the minimum and maximum values for all traits. For instance, the values of GLN ranged from 4.9 to 18.1, GLA from 73.8 to 433.8 cm2, TDW from 4.1 to 11.1 g, grain yield per hectare (GY) from 2.14 to 7.56 t ha–1, leaf water content (LWC) from 61.9% to 80.6%; LAI from 0.99 to 3.57, CT from 27.1 to 35.7 °C, NWI-1 from –0.038 to –0.005, NWI-2 from –0.040 to 0.000, NWI-3 from –0.045 to –0.005, NWI-4 from –0.039 to –0.013, R1000/R1100 from 0.938 to 1.024, R940/R960 from 1.002 to 1.089, water band index (WBI) from 1.002 to 1.117, NDVI from 0.406 to 0.838, R940/R960:NDVI from 1.233 to 2.473, and WBI:NDVI from 1.263 to 2.475.

Table 1. Description of the spectral reflectance indices examined in this study.

Spectral reflectance indices Formula Reference

Normalized water index 1 (NWI-1) (R970 – R900)/(R970 + R900) Babar et al. (2006); Prasad et al. (2007)




Ratio of reflectance between 1000 and 1100 nm R1000/R1100 Elsayed et al. (2011)

Ratio of reflectance between 940 and 960 nm R940/R960 Elsayed et al. (2011)

Water band index (WBI) R900/R970 Peñuelas et al. (1993); Claudio et al. (2006)

Normalized difference vegetation index (NDVI) (R800 – R680)/(R800 + R680) Claudio et al. (2006); Mistele and Schmidhalter (2008)

Ratio of reflectance between 940/960 nm and NDVI (R940/R960)/NDVI Elsayed et al. (2011)

Ratio of reflectance between WBI and NDVI WBI/NDVI Peñuelas et al. (1997); Claudio et al. (2006)

5


Tabl

e 2.

Ana

lysis

of v

aria

nce

(mea

n sq

uare

), m

inim

um, m

axim

um, a

nd m

eans

of d

iffer

ent m

orph

ophy

siolo

gica

l tra

its a

nd s

pect

ral r

eflec

tanc

e in

dice

s of

90

whe

at g

erm

plas

ms

grow

n un

der w

ater

shor

tage

cond

ition

s. Va

lues

wer

e ca

lcul

ated

usin

g m

eans

of e

ach

line

acro

ss tw

o ye

ars (

each

n =

270

).

SOV

dfG

LN G

LA T

DW

GY

LWC

LA

I C

T

Repl

icat

ion

22.

4614

3.08

0.08

14.

7936

.52

0.08

40.

56

Ger

mpl

asm

89

18.5

4***

9703

.24**

*5.

02**

*4.

74**

*59

.20**

*0.

624**

*6.

15**

Erro

r17

81.

1146

2.95

0.71

0.71

9.99

0.03

80.

859

Min

imum

4.

973

.84.

12.

1461

.90.

9927

.1

Max

imum

18.1

433.

811

.17.

5680

.63.

5735

.7

Mea

ns10

.319

1.69

6.57

4.93

71.4

41.

7431

.65

NW

I-1

NW

I-2

NW

I-3

NW

I-4

R100

0/R1

100

R940

/R96

0W

BIN

DV

IR9

40/R

960:

ND

VI

WBI

:ND

VI

Repl

icat

ion

20.

0012

0.00

180.

0000

160.

0000

940.

0012

0.00

056

0.00

130.

009

0.11

10.

0929

Ger

mpl

asm

89

0.00

018

0.00

036

0.00

021**

0.00

012**

0.00

068

0.00

025**

0.00

31**

*0.

033**

*0.

414**

*0.

378**

*

Erro

r17

80.

0001

60.

0003

20.

0000

140.

0000

190.

0005

30.

0001

30.

0003

70.

006

0.16

60.

173

Min

imum

–0

.038

–0.0

4–0

.045

–0.0

390.

938

1.00

21.

002

0.40

61.

233

1.26

3

Max

imum

–0.0

050.

000

–0.0

05–0

.013

1.02

41.

089

1.17

70.

838

2.47

32.

475

Mea

ns–0

.024

–0.0

20–0

.027

–0.0

260.

977

1.03

21.

054

0.65

71.

611

1.64

SOV,

sour

ce o

f var

ianc

e; g

reen

leaf

num

ber (

no.),

GLN

; gre

en le

af ar

ea (c

m2 ),

GLA

); to

tal d

ry w

eigh

t (g)

, TD

W; g

rain

yie

ld (t

ha-1

), G

Y; le

af w

ater

cont

ent (

%),

LWC

; lea

f are

a ind

ex,

LAI;

cano

py te

mpe

ratu

re (°

C),

CT;

nor

mal

ized

wat

er in

dex

1, N

WI-

1; n

orm

aliz

ed w

ater

inde

x 2,

NW

I-2;

nor

mal

ized

wat

er in

dex

3, N

WI-

3; n

orm

aliz

ed w

ater

inde

x 4,

NW

I-4;

w

ater

ban

d in

dex,

WBI

; nor

mal

ized

diff

eren

ce v

eget

atio

n in

dex,

ND

VI.

**,**

*: Si

gnifi

cant

at th

e 0.

01 a

nd 0

.001

pro

babi

lity

leve

ls, re

spec

tivel

y.

6


3.2. Broad-sense heritability of the investigated traitsThe broad-sense heritability for all traits was estimated by genetic and phenotypic variances using means of each germplasm across two years (Table 3). Morphophysiological traits showing high heritability included GLN, GLA, and LAI, ranging from 0.84 to 0.87. Moderate heritability was observed for TDW, GY, LWC, and CT, ranging from 0.62 to 0.67 (Table 3). The estimation of broad-sense heritability of NWI-3 (0.83) was high as indicated in the traits of GLN, GLA, and LAI, while that of NWI-4, WBI, and NDVI was moderate as reflected in the traits of TDW, GY, LWC, and CT (Table 3). The spectral reflectance indices showing low heritability included R940/R960, R940/R960:NDVI, and WBI:NDVI, while NWI-1, NWI-2, and R1000/R1100 showed no heritability (Table 3). 3.3. Genetic correlation analysis Genetic correlation analysis indicated that the spectral reflectance indices (WBI, NDVI, WBI/NDVI, and (R940/R960)/NDVI) provided higher genetic correlations with all morphophysiological traits. Regarding the NWIs, NWI-3 and NWI-4 showed better genetic correlations with morphophysiological traits compared to NWI-1 and NWI-2. LAI showed better genetic correlations with

morphophysiological traits than CT. The same spectral reflectance indices that showed significant genetic correlations with destructive morphophysiological traits also provided higher genetic correlations with LAI. R940/R960 failed to show a significant genetic correlation with the destructive morphophysiological traits, but its genetic correlation with LAI was significant (Table 4). 3.4. Direct and indirect effects of spectral reflectance indices on morphophysiological traits The simple correlation coefficients were partitioned into direct and indirect effects using path analysis. As can be seen from Table 5, (R940/R960)/NDVI and WBI/NDVI are the two spectral reflectance indices indicating the highest positive or negative direct effect on all morphophysiological traits with the exception of TDW. The NDVI was the single index most strongly influencing TDW, followed by WBI/NDVI and LAI. Though the direct effect of the four NWIs (NWI-1, NWI-2, NWI-3, and NWI-4) and R1000/R1100 on all morphophysiological traits was negligible or much weaker, they exerted a relatively strong indirect effect on all morphophysiological traits. LAI exerted a greater direct effect on all destructive morphophysiological traits than CT; the opposite held true for indirect effect (Table 5).

Table 3. Genotypic variance (σ2g), environmental variance (σ2

e), phenotypic variance (σ2p), broad-sense

heritability, least significance difference (LSD at P < 0.05), and coefficient of variation (CV) using means of each germplasm across two years.

Traits σ2g σ2

e σ2p Heritability LSD (5%) CV

GLN1 5.81 1.11 6.92 0.84 1.69*** 14.2GLA 3080.10 462.95 3543.05 0.87 34.37*** 12.7TDW 1.44 0.71 2.15 0.67 1.36*** 13.7GY 1.34 0.71 2.05 0.65 1.36*** 15.5LWC 16.40 9.99 26.39 0.62 5.09*** 6.2LAI 0.20 0.04 0.23 0.84 0.31*** 6.2CT 1.76 0.86 2.62 0.67 1.09** 6.6NWI-1 6.67E-06 1.60E-04 1.67E-04 0.04 ns 13.7NWI-2 1.33E-05 3.20E-04 3.33E-04 0.04 ns 13.1NWI-3 6.54E-05 1.37E-05 7.91E-05 0.83 0.006*** 13.3NWI-4 3.37E-05 1.88E-05 5.25E-05 0.64 0.007*** 14.6R1000/R1100 5.00E-05 5.30E-04 5.80E-04 0.09 ns 8.5R940/R960 4.00E-05 1.30E-04 1.70E-04 0.24 0.018* 10.2WBI 9.10E-04 3.70E-04 1.28E-03 0.71 0.031*** 13.0NDVI 9.00E-03 6.00E-03 1.50E-02 0.60 0.121*** 10.1R940/R960:NDVI 0.083 0.166 0.249 0.33 0.656*** 12.6WBI:NDVI 0.068 0.173 0.241 0.28 0.670*** 14.8

*: Significant at P < 0.05; **: significant at P < 0.01; ***: significant at P < 0.001; ns: nonsignificant.1The codes for the traits are presented in Table 2.

7


Tabl

e 4.

Gen

etic

cor

rela

tion

coeffi

cien

ts (l

ower

left)

and

thei

r sig

nific

ance

leve

ls (u

pper

righ

t) be

twee

n in

vest

igat

ed tr

aits

eva

luat

ed u

nder

wat

er sh

orta

ge c

ondi

tions

in 9

0 sp

ring

whe

at g

erm

plas

ms a

cros

s tw

o ye

ars.

12

34

56

78

910

1112

1314

1516

17

GLN

(1)1

***

***

***

**ns

ns*

***

ns**

****

****

*

GLA

(2)

0.82

***

***

**ns

ns*

***

ns**

****

****

*

TDW

(3)

0.85

0.81

***

***

nsns

****

**ns

****

***

***

***

**

GY

(4)

0.82

0.79

0.92

***

**

***

***

***

***

***

***

***

***

***

LWC

(5)

0.69

0.73

0.79

0.99

nsns

***

***

***

ns**

***

***

***

***

**

NW

I-1

(6)

–0.3

0–0

.35

–0.3

2–0

.44

–0.3

3*

**

**

***

**

nsns

NW

I-2

(7)

–0.3

7–0

.32

–0.3

5–0

.41

–0.3

70.

45*

**

ns*

**

**ns

*

NW

I-3

(8)

0.54

–0.5

6–0

.67

–0.8

5–0

.82

0.48

0.43

***

***

***

***

***

***

***

NW

I-4

(9)

–0.5

4–0

.53

–0.6

30.

83–0

.87

0.59

0.60

0.82

***

****

***

****

***

R 1000

/R11

00 (1

0)–0

.66

–0.6

9–0

.76

–0.9

5–0

.88

0.58

0.59

0.98

0.97

****

***

***

***

***

***

R 940/R

960 (1

1)0.

320.

330.

370.

400.

33–0

.50

–0.3

1–0

.60

–0.6

8–0

.73

***

nsns

*ns

WBI

(12)

0.64

0.69

0.77

0.88

0.89

–0.6

4–0

.58

–0.7

9–0

.82

–0.9

80.

65**

***

****

**

ND

VI (

13)

0.76

0.74

0.85

0.96

0.84

–0.5

9–0

.54

–0.8

1–0

.77

–0.8

90.

530.

81**

***

***

***

R 940/R

960:N

DV

(I 4

)–0

.76

–0.7

3–0

.91

–0.9

3–0

.74

0.52

0.57

0.83

0.77

0.90

–0.3

7–0

.76

0.99

***

***

**

WBI

:ND

VI (

15)

–0.7

6–0

.72

–0.9

0–0

.91

–0.8

10.

450.

720.

810.

750.

85–0

.38

–0.7

1–0

.98

0.99

***

**

LAI (

16)

0.75

0.68

0.85

0.98

0.89

–0.3

5–0

.35

0.78

–0.7

5–0

.89

0.60

0.84

0.87

–0.8

3–0

.80

**

CT

(17)

–0.5

1–0

.43

–0.4

7–0

.67

–0.6

00.

350.

430.

570.

570.

61–0

.32

–0.5

2–0

.68

0.70

0.70

–0.6

5

*: Si

gnifi

cant

at P

< 0

.05;

**: s

igni

fican

t at P

< 0

.01;

***:

signi

fican

t at P

< 0

.001

; ns:

nons

igni

fican

t.1 Th

e co

des f

or th

e tr

aits

are

pre

sent

ed in

Tab

le 2

.

8


Tabl

e 5. P

ath

coeffi

cien

t ana

lysis

show

ing d

irect

and

indi

rect

effec

ts o

f diff

eren

t spe

ctra

l refl

ecta

nce i

ndic

es o

n m

orph

ophy

siolo

gica

l tra

its ev

alua

ted

unde

r wat

er sh

orta

ge co

nditi

ons

in 9

0 sp

ring

whe

at g

erm

plas

ms a

cros

s tw

o ye

ars.

Trai

tsD

irect

effe

cts

Indi

rect

effe

cts

GLN

*(n

o.)

GLA

(cm

2 )TD

W(g

)G

Y(t

ha–1

)LW

C(%

)LA

IC

T(°

C)

GLN

(no.

)G

LA(c

m2 )

TDW

(g)

GY

(t h

a–1)

LWC

(%)

LAI

CT

(°C

)

NW

I-1

–0.0

81–0

.047

0.05

70.

049

–0.0

450.

041

0.00

0–0

.216

–0.2

95–0

.291

–0.3

90–0

.461

–0.3

760.

240

NW

I-2

0.02

7–0

.010

0.06

2–0

.054

–0.0

22–0

.080

0.03

4–0

.322

–0.3

17–0

.373

–0.4

08–0

.382

–0.3

320.

245

NW

I-3

0.15

20.

135

0.09

90.

015

–0.1

53–0

.122

0.10

0–0

.662

–0.6

65–0

.697

–0.7

71–0

.675

–0.6

080.

444

NW

I-4

0.04

40.

128

0.13

0–0

.014

–0.1

39–0

.049

0.18

4–0

.537

–0.6

15–0

.667

–0.6

86–0

.663

–0.6

360.

342

R 1000

/R11

000.

049

–0.0

370.

079

–0.0

41–0

.168

–0.1

410.

032

–0.4

62–0

.428

–0.5

26–0

.604

–0.5

89–0

.498

0.38

1

R 940/R

960

–0.2

630.

182

0.02

60.

530

0.47

4–0

.166

–0.0

570.

509

0.12

20.

247

–0.2

01–0

.028

0.49

7–0

.131

WBI

0.65

50.

180

0.11

5–0

.945

-0.3

850.

479

0.23

9–0

.062

0.46

70.

582

1.71

01.

255

0.30

4–0

.725

ND

VI

–0.1

480.

026

1.09

10.

248

0.30

50.

934

–0.3

850.

850

0.65

5–0

.252

0.59

30.

485

–0.1

38–0

.240

R 940/R

960:N

DV

I2.

849

–1.2

22–0

.163

–5.5

17–4

.097

1.97

01.

015

–3.5

250.

566

–0.6

274.

710

3.34

8–2

.711

-0.3

97

WBI

:ND

VI

–3.0

990.

737

0.53

05.

074

3.81

8–1

.314

–0.8

492.

440

–1.3

67–1

.298

–5.8

49–4

.520

0.61

11.

454

LAI

0.47

80.

203

0.33

70.

698

0.01

2–

–0.2

400.

227

0.43

90.

446

0.24

50.

776

––0

.384

CT

(°C

)–0

.011

0.04

20.

021

–0.1

020.

019

–0.2

40–

–0.4

73–0

.460

–0.5

44–0

.582

–0.5

63–0

.384

–

*The

code

s for

the

trai

ts a

re p

rese

nted

in T

able

2.

9


3.5. Functional relationship between morphophysiological traits and spectral reflectance indices Based on R2, the best functional relationship between morphophysiological traits and the spectral reflectance indices was established through regression analysis (Figures 2A–2E and 3A–3E). NWI-1, NWI-2, and the 940/960 index were weakly related to all morphophysiological traits. The

best spectral reflectance indices for detecting the LWC were the WBI (R2 = 0.76), NWI-3 (R2 = 0.68), NWI-4 (R2 = 0.64), and NDVI (R2 = 0.62) (Figures 2E and 3E), while for differentiating TDW and GY they were the NDVI (R2 = 0.70 and 0.71), (940/960)/NDVI index (R2 = 0.74 and 0.72), and WBI:NDVI (R2 = 0.72 and 0.67), respectively (Figures 2C, 2D, 3C, and 3D). The NDVI showed always a

E

L eaf water content (%)60 62 64 66 68 70 72 74 76 78 80 82 84

NWI–1 R2 = 0.25*NWI–2 R2 = 0.17NWI–3 R2 = 0.68**NWI–4 R2 = 0.64**

B

Green leaf area (cm2 plant –1)50 10

015

020

025

030

035

040

045

050

0–0.056

–0.048

–0.040

–0.032

–0.024

–0.016

–0.008

0.000

0.008NWI–1 R2 = 0.11NWI–2 R2 = 0.11NWI–3 R2 = 0.28*NWI–4 R2 = 0.24*

Spec

tral

ref

lect

ance

ind

ices

D

Grain yield ( ha–1 )

3 4 5 6 7 8 9

NWI–1 R2 = 0.11NWI–2 R2 = 0.22NWI–3 R2 = 0.57**NWI–1 R2 = 0.49*

2

A

Green leaf number (no plant–1 )

4 6 8 10 12 14 16 18 20–0.056

–0.048

–0.040

–0.032

–0.024

–0.016

–0.008

0.000

0.008 NWI–1 R2 = 0.09NWI–2 R2 = 0.10NWI–3 R2 = 0.26 *NWI–4 R2 = 0.24*

Total dry weight (g plant–1)3 4 5 6 7 8 9 10 11 12

–0.056

–0.048

–0.040

–0.032

–0.024

–0.016

–0.008

0.000

0.008 NWI–1 R2 = 0.05NWI–2 R2 = 0.10NWI–3 R2 = 0.36*NWI–4 R2 = 0.29*

C

Figure 2. Functional relationship between morphophysiological traits and the four normalized water indices. *, **: Significant at the 0.05 and 0.01 probability level, respectively.

10


positive steep trend, while the WBI/NDVI and (940/960)/NDVI indices showed always negative steep trends with all morphophysiological traits (Figure 3). Although the relationships between the 940/960 index and all morphophysiological traits were weak and nonsignificant, this index showed a close and significant relationship with morphophysiological traits when it was recomputed as a new index with the NDVI ((940/960)/NDVI) (Figure 3).

The functional relationship between spectral reflectance indices and LAI and CT is presented in Figure 4. In general, the LAI showed better relationships with spectral reflectance indices than CT. Among the four NWIs, NWI-3 and NWI-4 showed a better association with LAI and CT than did NWI-1 and NWI-2. WBI, NDVI, R940/R960:NDVI,

and WBI:NDVI showed also better associations with LAI and CT than did R1000/R1100 and R940/R960 (Figure 4). 3.6. Forward multiple linear regression analysesData presented in Table 6 show the partial and cumulative R2 values as well as the probability of the accepted spectral reflectance indices predicting each morphophysiological trait. The results indicated that WBI, NWI-3, R1000/R1100, WBI/NDVI, R940/R960, (R940/R960)/NDVI, and NWI-4 explained 75.7%, 7.6%, 2.1%, 1.8%, 1.1%, 0.8%, and 0.7% of the total variation in LWC, respectively. NDVI and LAI explained together 55.1% and 74.0% of the total variation in GLN and TDW, respectively. LAI, R940/R960:NDVI, CT, and NWI-2 explained 88.8%, 2.6%, 0.7%, and 0.4% of the total variation in GY (Table 6). For LAI and CT, the

2 –1Green leaf area (cm plant )

Spec

tral r

efle

ctan

ce in

dice

s

B

50 100 150 200 250 300 350 400 450 5000.00.30.60.91.21.51.82.12.42.73.03.33.63.9

R1000/ R1100 R2 = 0.21R940/ R960 R2 = 0.09WBI R2 = 0.42*NDVI R2 = 0.46*R940/ R960:NDVI R2 = 0.51**WBI:NDVI R2 = 0.54**

C

Total dry weight (g plant–1 )3 4 5 6 7 8 9 10 11 12

0.00.30.60.91.21.51.82.12.42.73.03.33.63.9

R1000/ R1100 R2 = 0.20R940/ R960 R2 = 0.08WBI R2 = 0.48*NDVI R2 = 0.70**R940/ R960:NDVI R2 = 0.74**WBI:NDVI R2 = 0.72**

D

Grain yield ( ha–1)2 3 4 5 6 7 8 9

R1000/ R1100 R2 = 0.42*R940/ R960 R2 = 0.11WBI R2 = 0.58**NDVI R2 = 0.71**R940/ R960:NDVI R2 = 0.72**WBI:NDVI R2 = 0.67**

E

Leaf water content (%)60 62 64 66 68 70 72 74 76 78 80 82 84

R1000/ R1100 R2 = 0.57*R940/ R960 R2 = 0.20WBI R2 = 0.76**NDVI R2 = 0.62**R940/ R960:NDVI R2 = 0.62**WBI:NDVI R2 = 0.57**

A

Green leaf number (no plant –1)4 6 8 10 12 14 16 18 20

0.00.30.60.91.21.51.82.12.42.73.03.33.63.9

R1000/ R1100 R2 = 0.17R940/ R960 R2 = 0.06WBI R2 = 0.35* NDVI R2 = 0.49*R940/ R960:NDVI R2 = 0.54**WBI:NDVI R2 = 0.53**

Figure 3. Functional relationship between morphophysiological traits and the different reflectance indices. *, **: Significant at the 0.05 and 0.01 probability level, respectively.

11


NDVI explained 63.4% and 39.0% of the total variation, respectively (Table 6).

4. DiscussionIt can be observed from Figure 1 that with the exception of the first 21 days after sowing the daily average temperature and humidity as well as daily reference evapotranspiration rate were similar in both growing seasons. Between 91 and 111 days from sowing, daily mean air temperature and daily reference evapotranspiration exceeded 25 °C and 10.0 mm day–1, respectively, with low relative humidity. This is enough to limit yield potential during fertilization and grain-filling stages, especially under irrigation water

shortage. Therefore, developing new cultivars with higher yield potential is of high importance to cope with irrigation water shortages under the present conditions of the arid Arabian Desert environment. In order to select wheat genotypes better adapted to these environments in plant breeding, it is essential to evaluate a large set of germplasms using different selection criteria.

The genetic variation for traits under selection and the higher heritability of these traits are necessary as guides to the breeding value and trait utility within the selection process. Therefore, the traits that are used as reliable selection criteria should have higher genetic variation and heritability. The results from heritability and

A

Leaf area index

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0– 0.056

– 0.048

– 0.040

– 0.032

– 0.024

– 0.016

– 0.008

0.000

0.008NWI–1 R2 = 0.11

NWI–2 R2 = 0.17

NWI–3 R2 = 0.54**

NWI–4 R2 = 0.46*

C

Canopy temperature (°C)

26 28 30 32 34 36 38

NWI–1 R2 = 0.06

NWI–2 R2 = 0.08

NWI–3 R2 = 0.30*

NWI–4 R2 = 0.27*

B

Spec

tral

ref

lect

ance

ind

ices

0.3

0.6

0.9

1.2

1.5

1.8

2.1

2.4

2.7

3.0

3.3

3.6R1000/R1100 R2 = 0.41*

R940/R960 R2 = 0.11

WBI R2 = 0.62**

NDVI R2 = 0.70**

R940/R960:NDVI R2 = 0.70**

WBI:NDVI R2 = 0.63**

DR1000/R1100 R2 = 0.17

R940/R960 R2 = 0.04

WBI R2 = 0.24*

NDVI R2 = 0.39*

R940/R960:NDVI R2 = 0.42*

WBI:NDVI R2 = 0.41*

Figure 4. Functional relationship between different spectral reflectance indices and the leaf area index and canopy temperature. *, **: Significant at the 0.05 and 0.01 probability level, respectively.

12


genetic correlation analysis indicated that the different morphophysiological traits investigated in this study could be used as reliable selection criteria to evaluate wheat germplasms under water shortage (Tables 3 and 4), which is in accordance with Chen et al. (2012) and Hackl et al. (2013). However, direct measurement of these traits by classical methods is difficult and time-consuming, especially when a large number of genotypes need to be evaluated across different environments. Therefore, the main goal of this study was to accurately estimate these destructive morphophysiological traits in a rapid and nondestructive manner.

Various studies have demonstrated that different morphophysiological traits can be assessed remotely and estimated simultaneously in a rapid and nondestructive manner if these traits show a significant correlation with the spectral reflectance indices that were calculated from reflectance of leaves or the whole canopy at different wavelengths (Araus et al., 2001; Royo et al., 2005; Babar

et al., 2006; Mirik et al., 2007; Erdle et al., 2013a, 2013b; Kipp et al., 2014a, 2014b). In this study, we observed that some spectral reflectance indices showed strong associations with the destructive morphophysiological traits investigated in this study. Furthermore, some spectral reflectance indices had relatively high to moderate heritability (Table 3). High genotypic correlations between all destructive morphophysiological traits and some spectral reflectance indices such as R1000/R1100, WBI, NDVI, WBI/NDVI, and R940/R960/NDVI were also observed. Among the four NWIs used in the current study, NWI-3 and NWI-4 were better genetically correlated with destructive morphophysiological traits compared to NWI-1 and NWI-2 (Table 4). These results indicate that several spectral reflectance indices examined in this study can be used as indirect selection criteria to predict different morphophysiological traits of a large number of germplasms in a rapid, low-cost, and nondestructive manner.

Water status of plants is commonly used in plant breeding as a useful selection criterion for evaluating different germplasms under different environmental stresses (Chaves et al., 2002; Munjal and Dhanda, 2005). The spectral reflectance indices examined in this study demonstrated that strong relationships existed between the LWC and WBI (R2 = 0.76), NWI-3 (R2 = 0.68), NWI-4 (R2 = 0.64), NDVI (R2 = 0.62), R1000/R1100 (R2 = 0.57), and WBI/NDVI (R2 = 0.57) (Figures 2E and Figure 3E). The results of direct and indirect effects also show that the WBI/NDVI index had the highest positive direct effect (3.818) and negative indirect effect (–4.520) on LWC, while the (R940/R960)/NDVI index possessed the highest negative direct effect (–4.097) and positive indirect effect (3.348) on LWC. NWI-3 and NWI-4 showed more negative direct and indirect effects on LWC than NWI-1 and NWI-2 (Table 5). In addition, the WBI alone explained 75.7% of the total variation in LWC, and it explained 87.3% when it was combined with NWI-3, R1000/R1100, and WBI/NDVI (Table 6). These results indicate that the spectral reflectance indices based on visible, near, and far infrared regions were associated with LWC in wheat under moisture stress conditions. These results generally agree with the findings obtained by Gutierrez et al. (2010) in different wheat genotypes; they found strong associations between certain WBIs such as NWI-3 and the relative water content of flag leaves. Peñuelas et al. (1993) and Peñuelas and Inoue (1999) also found that the spectral reflectance indexes WBI and WBI/NDVI showed strong correlation with relative water content in a variety of crops. Winterhalter et al. (2011) also identified better relationships to the water status of different maize genotypes under moisture stress in the VIS/NIR range. Elsayed et al. (2011) reported that it is advantageous to combine information from both the

Table 6. Multiple linear forward regression between morphophysiological traits and different spectral reflectance indices.

Responsevariables

Predicatorvariables

PartialR2

ModelR2

GLN*LAI 0.497 0.497NDVI 0.054 0.551

GLANDVI 0.463 0.463WBI 0.049 0.512

TDWNDVI 0.704 0.704LAI 0.036 0.740

GY

LAI 0.888 0.888R940/R960:NDVI 0.026 0.914CT 0.007 0.921NWI-2 0.004 0.925

LWC

WBI 0.757 0.757NWI-3 0.076 0.833R1000/R1100 0.021 0.854WBI:NDVI 0.018 0.873R940/R960 0.011 0.884R940/R960:NDVI 0.008 0.891NWI-4 0.007 0.898

LAINDVI 0.634 0.634WBI 0.088 0.722R1000/R1100 0.016 0.737

CT NDVI 0.390 0.390

*The codes for the traits are presented in Table 2.

13


visible and infrared regions (as WBI/NDVI and (R940/R960)/NDVI) to measure the water status of wheat crops such as the leaf water potential. The same authors reported that the spectral reflectance index NDVI was successfully used to describe the leaf water potential in wheat. NWI-3, examined in this study, has also been widely associated with diverse water relation parameters in a variety of crops in other studies (Sim and Gamon, 2003; Prasad et al., 2007). The association between plant water status parameters and the NWIs, WBI, and NDVI is based on the hypothesis that the NIR wavelengths (970 nm) penetrate deeper into the canopy and therefore accurately estimate the water content (Babar et al., 2006; Gutierrez et al., 2010).

A common response of water-stressed plants is accelerated leaf senescence, particularly around anthesis and throughout grain filling, which reduces the photosynthetic capacity of plant (Royo et al., 2004; Lopes and Reynolds, 2012; Kipp et al., 2014a). Thus, GLN and GLA were determined in this study as potential complementary selection criteria for evaluating germplasm under water shortage because both traits determine the ability of plants to intercept light and convert it into biomass and can also represent the ability of genotypes to extract a small amount of extra soil moisture from deep soil profiles (Lopes and Reynolds, 2012). Using spectral reflectance measurements, the NDVI was well related to GLN and GLA, followed by WBI/NDVI, (R940/R960)/NDVI, and WBI (Figures 3A and Figure 3B). There were also significant genotypic correlations between them, with genotypic correlation values ranging from 0.64 to 0.76 (Table 4). In addition, the same abovementioned indices also showed a strong association with TDW and GY (Figures 3C and Figure 3D). Furthermore, the NDVI alone explained 70.4% and 46.3% of the total variation in TDW and GLA, respectively (Table 5). All of these results indicate that the spectral reflectance indices, which have been successfully used to detect LWC in this study, and the used ratios and/or indices of VIS and NIR wavelengths like the NDVI and WBI as well as the WBI/NDVI and (R940/R960)/NDVI, showed good potential to detect differences in green biomass, GLA, and grain yield effectively. Because most reflectance indices estimating the water status of plants successfully explain the variations in total biomass and grain yield, canopy water content plays a very useful role in estimating growth and yield of wheat genotypes under water shortage conditions. Previous studies have shown that various spectral indices are suitable for detecting the variation in plant water status as well as growth and yield in various crops (El-Shikha et al., 2007; Prasad et al., 2007; Mistele and Schmidhalter, 2010; Elsayed et al. 2011). For instance, Elsayed et al. (2011) reported that the NDVI was a powerful and frequently used reflectance index to describe the leaf water potential and green biomass

simultaneously. In a study conducted with 25 bread wheat genotypes, NDVI explained around 40% of the variability found in biomass (Reynolds et al., 1999).

Measurements of LAI and CT both provide indirect indicators for various morphophysiological traits such as the water status of plants, the capacity of roots to access available soil water, and the main limitations of photosynthesis, growth, and production under water shortage conditions (El-Hendawy et al., 2005; Pask and Reynolds, 2013). Therefore, both parameters could be used to detect destructively ascertained morphophysiological traits in a rapid, low-cost, and nondestructive manner (Jin et al., 2013). The results of this study showed that LAI and CT had high and medium heritability, respectively (Table 3). In addition, LAI also provided higher genetic correlations, as shown in Table 4, and had positive direct effects on morphophysiological traits compared to CT (Table 5). Furthermore, LAI alone explained 88.8% and 49.7% of the total variation in GY and GLN, respectively (Table 6). All of these results indicate that the LAI could be used as reliable selection criterion to evaluate wheat germplasms under water shortage conditions. The effectiveness and reliability of this trait in evaluating the moisture stress tolerance may be related to the fact that LAI provides an indirect indicator for rapid ground cover and leaf expansion rate under moisture stress, which thus increases LAI, causing soil temperature and evaporation water losses to decrease as a result of shading of the soil surface by the crop canopy; up to 40% of the total soil water may be lost by evaporation in wheat under Mediterranean-type drought environments (Loss and Siddique, 1994; Condon et al., 2004).

Our results also show that the LAI can be assessed remotely and estimated simultaneously when morphophysiological traits are detected. The results of this study showed that the spectral reflectance indices that had shown strong correlations with morphophysiological traits, such as the NWI-3, NWI-4, NDVI, WBI, (R940/R960)/NDVI, and WBI/NDVI, were also significantly correlated with LAI. The genetic correlations between LAI and these indices ranged from 0.75 to 0.89 (Table 4). The results of direct effects also showed that the NDVI and (R940/R960)/NDVI had the highest positive direct effects (0.934 and 1.970) on LAI, respectively, while the WBI/NDVI possessed the highest negative direct effect (–1.314) on LAI (Table 5). These results indicate that the LAI could also be accurately predicted using spectral reflectance measurements. These results generally agree with the findings obtained by Scotford and Mille (2004) and Jin et al. (2013), who reported that some spectral indices have been successfully used to detect LAI at the canopy level indirectly, without using any other direct field measurements. The NDVI was found to be significantly

14


correlated with LAI in the studies of Aparicio et al. (2002) and Leon et al. (2003). Most importantly, previous studies have shown that some spectral reflectance indices were only effective to discriminate between differences in canopies when the LAI of the crop ranged between 2 and 3 (Scotford and Mille, 2004; Heege et al., 2008; Mistele and Schmidhalter, 2010). If the LAI exceeds 3, the spectral reflectance measurements are mainly obtained from upper canopy layers, while the reflectance of vegetation canopies is sensitive to soil background when the LAI is less than 2. Studies conducted in bread and durum wheat have demonstrated that simple ratio increases linearly with increases in LAI, while NDVI shows a curvilinear response. Furthermore, when the LAI of wheat canopies exceeds a certain level, several leaf layers of the canopy do not entail great changes in NDVI. This is because the reflectance of solar radiation from the underlying soil surface or lower leaf layers is largely attenuated due to the ground surface being completely obscured by the leaves (Serrano et al., 2000; Aparicio et al., 2002). Aparicio et al. (2002) also concluded that NDVI becomes relatively insensitive to changes in canopy structure of durum wheat when the LAI values are higher than 3. Therefore, in this study, the curves fitted between the LAI and the NDVI, (R940/R960)/NDVI, and WBI/NDVI indicated better

estimates than the linear fits (Figure 4). These findings indicate that it is important to integrate the values of LAI with spectral reflectance measurements when the latter are used as effective indicators to detect morphophysiological traits.

In conclusion, our results indicate that different spectral reflectance indices examined in this study have shown the potential to estimate different morphophysiological traits of a wide and diverse range of spring wheat germplasms in a rapid, low-cost, and nondestructive manner under field conditions. The spectral reflectance indices that have the potential to detect the water status of plants were also able to detect differences in green biomass, green leaf area, and grain yield effectively under water shortage conditions. The LAI can be used as an indirect selection criterion to predict most morphophysiological traits related to moisture stress tolerance in a rapid, low-cost, and nondestructive manner, and it can also accurately be predicted using spectral reflectance measurements.

AcknowledgmentThe authors would like to extend their sincere appreciation to the Deanship of Scientific Research at King Saud University for funding this research (Group No. RG-1435-032).

References

Abd El-Kareem THA, Saidy AEA (2011). Evaluation of yield and grain quality of some bread wheat genotypes under normal irrigation and drought stress conditions in calcareous soil. J Bio Sci 11: 156–164.

Aparicio N, Villegas D, Araus JL, Casadesus J, Royo C (2002). Relationship between growth traits and spectral vegetation indices in durum wheat. Crop Sci 42: 1547–1555.

Araus JL, Cairns J (2014). Field high-throughput phenotyping—The new crop breeding frontier. Trends Plant Sci 19: 52–61.

Araus JL, Casadesus J, Bort J (2001). Recent tools for the screening of physiological traits determining yield. In: Reynolds MP, Ortiz-Monasterio JI, McNab A, editors. Application of Physiology in Wheat Breeding. Mexico City, Mexico: CIMMYT, pp. 59–77.

Araus JL, Slafer GA, Royo C, Serret MD (2008). Breeding for yield potential and stress adaptation in cereals. Crit Rev Plant Sci 27: 377–412.

Babar MA, Reynolds MP, van Ginkel M, Klatt AR, Raun WR, Stone ML (2006). Spectral reflectance indices as a potential indirect selection criteria for wheat yield under irrigation. Crop Sci 46: 578–588.

Berger B, Parent B, Tester M (2010). High-throughput shoot imaging to study drought responses. J Exp Bot 6: 3519–3528.

Bürling K, Cerovic ZG, Cornic G, Ducruet JM, Noga G, Hunsche M (2013). Fluorescence-based sensing of drought-induced stress in the vegetative phase of four contrasting wheat genotypes. Enviro Exp Bot 89: 51–59.

Chaves MM, Pereira JS, Maroco J, Rodrigues ML, Ricardo CPP, Osorio ML, Carvalho I, Faria T, Pinheiro C (2002). How plants cope with drought in the field? Photosynthesis and growth. Ann Bot 89: 907–916.

Chen X, Mina D, Yasir TA, Hu Y (2012). Evaluation of 14 morphological, yield-related and physiological traits as indicators of drought tolerance in Chinese winter bread wheat revealed by analysis of the membership function value of drought tolerance (MFVD). Field Crops Res 137: 195–201.

Claudio HC, Cheng Y, Fuentes DA, Gamon JA, Luo H, Oechel W, Qiu HL, Rahman AF, Sims DA (2006). Monitoring drought effects in vegetation 20 water content and fluxes in chaparral with the 970 nm water band index. Remote Sens Environ 103: 304–311.

Cobb JN, DeClerck G, Greenberg A, Clark R, McCouch S (2013). Next-generation phenotyping: requirements and strategies for enhancing our understanding of genotype–phenotype relationships and its relevance to crop improvement. Theor Appl Genet 126: 867–887.

Condon AG, Richards, RA, Rebetzke GJ, Farquhar GD (2004). Breeding for high water-use efficiency. J Exp Bot 55: 2447–2460.

El-Hendawy S, Hu Y, Schmidhalter U (2005). Growth, ion content, gas exchange, and water relations of wheat genotypes differing in salt tolerance. Austr J Agr Res 56: 123–134.

15


Elsayed S, Mistele B, Schmidhalter U (2011). Can changes in leaf water potential be assessed spectrally? Funct Plant Biol 38: 523–533.

El-Shikha DM, Waller P, Hunsaker D, Clarke T, Barnes E (2007). Ground-based remote sensing for assessing water and nitrogen status of broccoli. Agric Water Manage 92: 183–193.

Erdle K, Mistele B, Schmidhalter U (2013a). Spectral high-throughput assessments of phenotypic differences in biomass and nitrogen partitioning during grain filling of wheat under high yielding Western European conditions. Field Crops Res 141: 16–21.

Erdle K, Mistele B, Schmidhalter U (2013b). Spectral assessments of phenotypic differences in spike development during grain filling affected by varying N supply in wheat. J Plant Nutr Soil Sci 6: 952–963.

Feng B, Yu H, Hu Y, Gao X, Gao J, Gao D, Zhang S (2009). The physiological characteristics of the low canopy temperature wheat (Triticum aestivum L.) genotypes under simulated drought condition. Acta Physiol Plant 31: 1229–1235.

Gutierrez M, Reynolds MP, Raun WR, Stone ML, Klatt AR (2010). Spectral water indices for assessing yield in elite bread wheat genotypes in well irrigated, moisture stressed, and high temperature conditions. Crop Sci 50: 197–214.

Hackl H, Hu Y, Schmidhalter U (2014). Evaluating growth platforms and stress scenarios to assess the salt tolerance of wheat plants. Funct Plant Biol 40: 409–424.

Heege HJ, Reusch S, Thiessen E (2008). Prospects and results for optical systems for site-specific on-the-go control of nitrogen-top-dressing in Germany. Precis Agric 9: 115–131.

Hicks SK, Lascano RJ (1995). Estimation of leaf area index for cotton canopies using the LI-COR LAI-2000 plant canopy analyzer. Agron J 87: 458–464.

Hura T, Grzesiak S, Hura K, Thiemt E, Tokarz K, Wedzones M (2007). Physiological and biochemical tools useful in drought-tolerance detection in genotypes of winter triticale: accumulation of ferulic acid correlates with drought tolerance. Ann Bot 100: 767–775.

Jin X, Diao W, Xiao C, Wang F, Chen B (2013). Estimation of wheat agronomic parameters using new spectral indices. PLoS ONE 8: e72736.

Jones HG (1999). The use of thermography for quantitative studies of spatial and temporal variation of stomatal conductance over leaf surfaces. Plant Cell Environ 22: 1043–1055

Jongdee B, Fukai S, Cooper M (2002). Leaf water potential and osmotic adjustment as physiological traits to improve drought tolerance in rice. Field Crop Res 76: 153–163.

Kipp S, Mistele B, Baresel P, Schmidhalter U (2014a). High-throughput phenotyping early plant vigour of winter wheat. Eur J Agron 52: 271–278.

Kipp S, Mistele B, Schmidhalter U (2014b). Identification of stay-green and early-senescence phenotypes in high-yielding winter wheat and their relationship to grain yield and grain protein concentration using high-throughput phenotyping techniques. Funct Plant Biol 41: 227–235.

Leon CT, Shaw DR, Cox MS, Abshire MJ, Ward B, Wardlaw MC (2003). Utility of remote sensing in predicting crop and soil characteristics. Preci Agri 4: 359–384.

Lopes MS, Reynolds MP (2012). Stay-green in spring wheat can be determined by spectral reflectance measurements (normalized difference vegetation index) independently from phenology. J Exp Bot 63: 3789–3798.

Loss SP, Siddique KHM (1994). Morphological and physiological traits associated with wheat yield increases in Mediterranean environments. Adv Agron 52: 229–276.

Mistele B, Schmidhalter U (2010). Tractor-based quadrilateral spectral reflectance measurements to detect biomass and total aerial nitrogen in winter wheat. Agron J 102: 499–506.

Munjal R, Dhanda SS (2005). Physiological evaluation of wheat (Triticum aestivum L.) genotypes for drought resistance. Indian J Genet Pl Br 65: 307–308.

Mutava RN, Prasad PVV, Tuinstra MR, Kofoid KD, Yu J (2011) Characterization of sorghum genotypes for traits related to drought tolerance. Field Crops Res 123: 10–18.

Pask AJD, Reynolds MP (2013). Breeding for yield potential has increased deep soil water extraction capacity. Crop Sci 53: 2090–2104

Peñuelas J, Filella I, Biel C, Serrano L, Save R (1993). The reflectance at the 950–970 mm region as an indicator of plant water status. Int J Remote Sens 14: 1887–1905

Peñuelas J, Inoue Y (1999). Reflectance indices indicative of changes in water and pigment contents of peanut and wheat leaves. Photosynthetica 36: 355–360

Peñuelas J, Piñol J, Ogaya R, Filella I (1997). Estimation of plant water concentration by the reflectance water index WI (R900/R970). Int J Remote Sens 18: 2869–2875.

Prasad B, Carver BF, Stone ML, Babar MA, Raun WR, Klatt AR (2007). Genetic analysis of indirect selection for winter wheat grain yield using spectral reflectance indices. Crop Sci 47: 1416–1425.

Rashid A, Stark JC, Tanveer A, Mustafa T (1999). Use of canopy temperature measurements as a screening tool for drought tolerance in spring wheat. J Agron Crop Sci 182: 231–237.

Reynolds MP, Rajaram S, Sayre KD (1999). Physiological and genetic changes of irrigated wheat in the post-green revolution period and approaches for meeting projected global demand. Crop Sci 39: 1611–1621.

Royo C, Álvaro F, Martos V, Ramdani A, Isidro J, Villegas D, García del Moral LF (2007). Genetic changes in durum wheat yield components and associated traits in Italian and Spanish varieties during the 20th century. Euphytica 155: 259–270.

Royo C, Aparicio N, Blanco R, Villegas D (2004). Leaf and green area development of durum wheat genotypes grown under Mediterranean conditions. Eur J Agron 20: 419–430.

Royo C, García del Moral LF, Slafer G, Nachit MM, Araus JL (2005). Selection tools for improving yield-associated physiological traits. In: Royo C, Nachit MN, Di Fonzo N, Araus JL, Pfeiffer WH, Slafer GA, editors. Durum Wheat Breeding: Current Approaches and Future Strategies. Binghamton, NY, USA: Food Products Press, pp. 563–598.

16


Scotford IM, Miller PCH (2004). Estimating tiller density and leaf area index of winter wheat using spectral reflectance and ultrasonic sensing techniques. Biosys Engin 89: 395–408.

Sims DA, Gamon JA (2003). Estimation of vegetation water content and photosynthetic tissue area from spectral reflectance: a comparison of indices based on liquid water and chlorophyll absorption features. Remote Sens Environ 84: 526–537.

Teal RK, Tubana B, Girma K, Freeman KW, Arnall DB, Walsh O, Raun WR (2006). In-season prediction of corn grain yield potential using normalized difference vegetation index. Agron J 98: 1488–1494.

Winterhalter L, Mistele B, Jampatong S, Schmidhalter U (2011). High throughput phenotyping of canopy water mass and canopy temperature in well-watered and drought stressed tropical maize hybrids in the vegetative stage. Eur J Agron 35: 22–32.

Zadoks JC, Chang TT, Konzak CF (1974). A decimal code for the growth stages of cereals. Weeds Res 14: 412–415.

Zia S, Romano G, Spreer W, Sanchez C, Cairns J, Araus JL, Müller J (2013). Infrared thermal imaging as a rapid tool for identifying moisture stress tolerant maize genotypes of different phenology. J Agron Crop Sci 199: 75–84.

spectral reflectance indices as a rapid and nondestructive...

Documents