6
Scientia Horticulturae 164 (2013) 610–615 Contents lists available at ScienceDirect Scientia Horticulturae journal h om epa ge: www.elsevier.com/locate/scihorti Canopy leaf area index for apple tree using hemispherical photography in arid region Chunwei Liu a,b , Shaozhong Kang b,, Fusheng Li c , Sien Li b , Taisheng Du b a Jiangsu Provincial Key Laboratory of Agricultural Meteorology, College of Applied Meteorology, Nanjing University of Information Science and Technology, Nanjing 210044, China b Center for Agricultural Water Research in China, China Agricultural University, Beijing 100083, China c College of Agriculture, Guangxi University, Nanning, Guangxi 530005, China a r t i c l e i n f o Article history: Received 16 July 2013 Received in revised form 28 September 2013 Accepted 4 October 2013 Keywords: Analysis software Apple orchard Canopy structure Gap fraction Hemispherical image Leaf area index a b s t r a c t The hemispherical images of apple orchard were collected using fisheye lens in 2010 and then respec- tively analyzed by the methods of Bonhom, 2000, 2000G, sphere, ellipse, LAI 2000-3 rings, LAI 2000-4 rings and Miller to obtain leaf area index (LAI), plant area index (PAI), gap fraction, leaf inclination angle and clumping index in an arid region of northwest China. Results show that during DOY (day of year) 202–237 (at flourish stage), the estimated LAI by logarithmic ellipse method and PAI by LAI 2000-4 rings method were 1.96 and 1.95 m 2 m 2 , respectively, which were close to the direct LAI (1.85 m 2 m 2 ). How- ever, the direct LAI was higher than the estimated LAI from ellipse method of Hemiview (1.59 m 2 m 2 ) and lower than the estimated PAI from gap fraction analysis method of Photoshop (2.67 m 2 m 2 ). Thus the methods of logarithmic ellipse and the LAI 2000-4 rings were used to estimate the LAI (PAI) from hemi- spherical images. And the estimated LAI by the methods of logarithmic ellipse and the LAI 2000-4 rings were used to analyze seasonal variations of canopy parameters. LAI increased rapidly at leaf expanding stage and then maintained at high level from the vigorous to harvest stage. Canopy gap fraction in the south side was similar to that in the north side. The leaf inclination angle tended horizontally to receive more radiation with the development of apple growth. In addition, the estimated LAI from hemispher- ical images using the methods of logarithmic ellipse and LAI 2000-4 rings was accurate even without calibrating clumping index. Thus the estimated canopy parameters from hemispherical images using the methods of logarithmic ellipse and LAI 2000-4 rings can be used for the modeling of evapotranspiration. © 2013 Published by Elsevier B.V. 1. Introduction Canopy parameters, such as leaf area index (LAI), leaf inclination angle and gap fraction, are used to describe radiation distribu- tion in the canopy and estimate the evapotranspiration in the orchard. LAI is the one-side area of leaves per unit ground area without considering the shape of the leaves (Bréda, 2003; Chen and Black, 1992; Lang, 1991; Smith et al., 2008) and it is crucial to calculate orchard evapotranspiration using Shuttleworth–Wallace model (Shuttleworth and Wallace, 1985). Because canopy parame- ters cannot distinguish the woody area from the direct LAI, Leblanc et al. (2005) suggested that plant area index (PAI) may be useful. References such as LAI, PAI and gap fraction has been monitor- ing to reflect evergreen or deciduous forest canopies that are the main gateways regulating the exchange of carbon and water vapour between terrestrial ecosystems and the atmosphere (Clark and Corresponding author. Tel.: +86 10 62737611; fax: +86 10 62737611. E-mail addresses: [email protected], [email protected] (S. Kang). Murphy, 2011; Seidel et al., 2011) and to calculate the evapotrans- piration in arid orchard where needs precise water management during the growing season (Marsal et al., 2014; Testi et al., 2004). Mean leaf inclination angle is one of important parameters in canopy structure that stands the angle between the normal and vertical directions of the leaf, reflecting mean horizontal condition of the leaves, which determines the radiation intercepted by the forest or orchard canopy (Stuckens et al., 2011; Wang et al., 2007). Direct and indirect methods can be applied to measure LAI in fruit trees. The direct method to measure LAI is more precise than the indirect method, but it is usually destructive to the trees, and labor- or time-consuming. Thus, some indirect methods on the basis of optical principle are used to improve the drawbacks of direct methods, and they can obtain other canopy parameters such as leaf inclination angle and gap fraction as well (Chen et al., 1997; Lang et al., 2010; Jonckheere et al., 2004). Hemispherical photography is one of the indirect methods and it is convenient to operate, collect photography limitlessly and assess photogra- phy simultaneously. Hemispherical photography is collected using fisheye lens and analyzed using the threshold value of the gap 0304-4238/$ see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.scienta.2013.10.009

Canopy leaf area index for apple tree using hemispherical photography in arid region

Embed Size (px)

Citation preview

Page 1: Canopy leaf area index for apple tree using hemispherical photography in arid region

Cp

Ca

Nb

c

a

ARR2A

KAACGHL

1

atowacmteRimb

0h

Scientia Horticulturae 164 (2013) 610–615

Contents lists available at ScienceDirect

Scientia Horticulturae

journa l h om epa ge: www.elsev ier .com/ locate /sc ihor t i

anopy leaf area index for apple tree using hemisphericalhotography in arid region

hunwei Liua,b, Shaozhong Kangb,∗, Fusheng Li c, Sien Lib, Taisheng Dub

Jiangsu Provincial Key Laboratory of Agricultural Meteorology, College of Applied Meteorology, Nanjing University of Information Science and Technology,anjing 210044, ChinaCenter for Agricultural Water Research in China, China Agricultural University, Beijing 100083, ChinaCollege of Agriculture, Guangxi University, Nanning, Guangxi 530005, China

r t i c l e i n f o

rticle history:eceived 16 July 2013eceived in revised form8 September 2013ccepted 4 October 2013

eywords:nalysis softwarepple orchardanopy structureap fractionemispherical image

a b s t r a c t

The hemispherical images of apple orchard were collected using fisheye lens in 2010 and then respec-tively analyzed by the methods of Bonhom, 2000, 2000G, sphere, ellipse, LAI 2000-3 rings, LAI 2000-4rings and Miller to obtain leaf area index (LAI), plant area index (PAI), gap fraction, leaf inclination angleand clumping index in an arid region of northwest China. Results show that during DOY (day of year)202–237 (at flourish stage), the estimated LAI by logarithmic ellipse method and PAI by LAI 2000-4 ringsmethod were 1.96 and 1.95 m2 m−2, respectively, which were close to the direct LAI (1.85 m2 m−2). How-ever, the direct LAI was higher than the estimated LAI from ellipse method of Hemiview (1.59 m2 m−2)and lower than the estimated PAI from gap fraction analysis method of Photoshop (2.67 m2 m−2). Thus themethods of logarithmic ellipse and the LAI 2000-4 rings were used to estimate the LAI (PAI) from hemi-spherical images. And the estimated LAI by the methods of logarithmic ellipse and the LAI 2000-4 ringswere used to analyze seasonal variations of canopy parameters. LAI increased rapidly at leaf expanding

eaf area index stage and then maintained at high level from the vigorous to harvest stage. Canopy gap fraction in thesouth side was similar to that in the north side. The leaf inclination angle tended horizontally to receivemore radiation with the development of apple growth. In addition, the estimated LAI from hemispher-ical images using the methods of logarithmic ellipse and LAI 2000-4 rings was accurate even withoutcalibrating clumping index. Thus the estimated canopy parameters from hemispherical images using themethods of logarithmic ellipse and LAI 2000-4 rings can be used for the modeling of evapotranspiration.

. Introduction

Canopy parameters, such as leaf area index (LAI), leaf inclinationngle and gap fraction, are used to describe radiation distribu-ion in the canopy and estimate the evapotranspiration in therchard. LAI is the one-side area of leaves per unit ground areaithout considering the shape of the leaves (Bréda, 2003; Chen

nd Black, 1992; Lang, 1991; Smith et al., 2008) and it is crucial toalculate orchard evapotranspiration using Shuttleworth–Wallaceodel (Shuttleworth and Wallace, 1985). Because canopy parame-

ers cannot distinguish the woody area from the direct LAI, Leblanct al. (2005) suggested that plant area index (PAI) may be useful.eferences such as LAI, PAI and gap fraction has been monitor-

ng to reflect evergreen or deciduous forest canopies that are theain gateways regulating the exchange of carbon and water vapour

etween terrestrial ecosystems and the atmosphere (Clark and

∗ Corresponding author. Tel.: +86 10 62737611; fax: +86 10 62737611.E-mail addresses: [email protected], [email protected] (S. Kang).

304-4238/$ – see front matter © 2013 Published by Elsevier B.V.ttp://dx.doi.org/10.1016/j.scienta.2013.10.009

© 2013 Published by Elsevier B.V.

Murphy, 2011; Seidel et al., 2011) and to calculate the evapotrans-piration in arid orchard where needs precise water managementduring the growing season (Marsal et al., 2014; Testi et al., 2004).Mean leaf inclination angle is one of important parameters incanopy structure that stands the angle between the normal andvertical directions of the leaf, reflecting mean horizontal conditionof the leaves, which determines the radiation intercepted by theforest or orchard canopy (Stuckens et al., 2011; Wang et al., 2007).

Direct and indirect methods can be applied to measure LAIin fruit trees. The direct method to measure LAI is more precisethan the indirect method, but it is usually destructive to the trees,and labor- or time-consuming. Thus, some indirect methods onthe basis of optical principle are used to improve the drawbacksof direct methods, and they can obtain other canopy parameterssuch as leaf inclination angle and gap fraction as well (Chen et al.,1997; Lang et al., 2010; Jonckheere et al., 2004). Hemispherical

photography is one of the indirect methods and it is convenientto operate, collect photography limitlessly and assess photogra-phy simultaneously. Hemispherical photography is collected usingfisheye lens and analyzed using the threshold value of the gap
Page 2: Canopy leaf area index for apple tree using hemispherical photography in arid region

C. Liu et al. / Scientia Horticultu

Table 1Parameters of experimental trees in the orchard. Truck diameters of 34 trees weremeasured in the orchard, and trunk diameters of 150–200 mm and 200–250 mmoccupy 24% and 71% of total trees, respectively.

Tree no. A1 A2 A3 A4 A5 A6

Trunk diameter (mm) 178 203 214 227 217 235

fwaL12s2

D2adatuodtoalst

2

2

iClsaptedia10

pscfomawris

ln(�)

ractions obtained from digital RGB (red, green and blue) imagesith a sky background (Sandmann et al., 2013). However, many

uthors suggested that clumping index leads to lower estimatedAI by hemispherical photography than the direct LAI (Chen et al.,997; Chen and Cihlar, 1995; Leblanc et al., 2005; Macfarlane et al.,007a). And canopy density determined by exposure time of hemi-pherical image is crucial to precisely estimate LAI (Zhang et al.,005).

A series of software packages, such as Hemiview (Delta-Tevice), Winscanopy (Rich et al., 1993) and Caneye (Weiss et al.,004), have been developed for analyzing hemispherical images,nd each has several calculation methods to obtain the results forifferent sites. However, the applicability of different methods innalyzing hemispherical images needs to be further investigated inhe arid apple orchard. This study collected hemispherical imagessing fisheye lens over the whole season of apple tree in arid regionf northwest China in 2010 and estimated canopy parameters usingifferent software packages. The indirect LAI values from gap frac-ion inversions were validated by the direct LAI values. So thebjectives of this study were to obtain suitable software packages innalyzing canopy hemispherical images in apple orchard and ana-yze seasonal variations of canopy parameters using the selectedoftware packages, so as to provide reliable canopy parameters forhe evapotranspiration model.

. Materials and methods

.1. Experimental outline

Field experiment was conducted in 2010 at Shiyanghe Exper-mental Station for Water-saving in Agriculture and Ecology ofhina Agricultural University (N37◦52′, E102◦51′, altitude 1581 m),

ocated in Wuwei city, Gansu province of northwest China. Theite is a typical continental temperate climate zone with a meannnual sunshine duration of more than 3000 h, mean annual tem-erature of 8 ◦C, annual accumulated temperature (>0 ◦C) of morehan 3550 ◦C, mean annual precipitation of 164.4 mm, mean annualvaporation from a free water surface of 2000 mm and free frostays of 150 days. The groundwater table is below 40–50 m. Exper-

mental soil is irrigated desert soil (Siltigic-Orthic Anthrosols)nd soil texture is sandy loam, with a mean dry bulk density of.46 g cm−3 and mean volumetric water content at field capacity of.30 cm3 cm−3.

Apple trees (Malus domestica Borkh. cv Golden Delicious) werelanted with an east-west row orientation in 1981, with rowpacing of 6 m and plant spacing of 4 m. Irrigation amount wasalculated by mean evapotranspiration over the growing stageor nearly 5 years, and irrigation date was determined by actualrchard management and environmental condition, and irrigationethod was border-irrigation in the plot (4 m × 6 m). Irrigation

mount for each tree was controlled by a water meter. The treesere fertilized with N, P2O5 and K2O of 800, 200 and 150 kg h m−2,

espectively, and received similar pest and weed control and prun-

ng in the orchard. And the trunk diameters of the randomlyelected apple trees are shown in Table 1.

rae 164 (2013) 610–615 611

2.2. Measurements

2.2.1. Leaf area index by direct methodDirect LAI measurements were taken during 20th July to 30th

August when apple canopies were well developed and the LAI wassteady. Since it was time-consuming to measure each leaf area ofmature apple trees, Macfarlane et al. (2007a,b) measured total LAIby dividing and weighing tree’s branches into three groups. Thefollowing equation was used to estimate LAI of three apple trees inour study:

LAI = nAn

S(1)

where LAI is direct leaf area index, n is the number of total effectiveleaves defined as leaves with larger leaf area and thicker leaf stemor two or three small area leaves as an effective leaf; An is meanleaf area of the trees measured randomly for 300 leaves by AM300portable leaf area meter (ADC Ltd. UK); S is canopy area calculatedby canopy diameter of four directions with steel measuring ruler(5 m) and stem base diameter with vernier caliper (50 cm).

2.2.2. Hemispherical photographyHemispherical images were obtained using a Nikon FC-E8 fish-

eye lens attached to a Nikon Coolpix 8400 digital camera supportedby a tripod approximately 0.5 m tall in six apple trees (Table 1). Thecamera was leveled using a bubble level placed in the flash slot, andthe north direction was indicated automatically by the Winscanopy(Winscanopy 2006a, Regent, Quebec, Canada) analysis system. Thecamera was set to auto and manual model in the same position toget images of different exposure time. Since the manual exposurecannot distinguish the leaves from the branches of the apple tree,Zhang et al. (2005) suggested that the manual model of the cameracontributes to the difference between canopy and sky, and the pro-cess is to (1) use the same camera with the same fish-eye lens in avery large opening with no obstructions and the preferred apertureis F5.3 or similar; (2) determine the in-stand exposure by increas-ing the shutter speed by two stops with the aperture unchangedat F5.3 for manual exposure and automatic shutter speed for auto-matic exposure in the experiment of apple orchard. Hemisphericalimages were taken at sunset every 5 days.

2.2.2.1. Number of pixels to calculate canopy references. Hemispher-ical images were analyzed using Adobe Photoshop CS4-11.0.1 asfollows: large gaps between tree crowns of each photograph wereselected using the ‘wand’ tool and the total number of pixels con-tained large gaps (gL) recorded from the Image Histogram; all gapswere then selected using the Select Similar menu item and thetotal number of pixels in gaps (gT) recorded from the Image His-togram. The fractions of foliage cover (ff) and crown cover (fc) werecalculated as (Macfarlane et al., 2007c):

ff = 1 − gT

Q(2)

fc = 1 − gL

Q(3)

where Q is the total pixels of the hemispherical image. Crown poros-ity (�) is calculated from

� = 1 − fffc

(4)

The PAI was estimated using a modified version of theBeer–Lambert law as follows,

PAI = fc �(5)

where � is the extinction coefficient.

Page 3: Canopy leaf area index for apple tree using hemispherical photography in arid region

612 C. Liu et al. / Scientia Horticulturae 164 (2013) 610–615

2sDmi(es(2pb

L

le

P

wtb

P

ww

P

wjrW

˝

wm

Table 2Leaf area index measured by direct method of three experimental trees.

No. A1 A3 A4

Measurement time 20–25 July 21–24 August 24–30 AugustCanopy radius (m) 2.090 2.544 2.931Trunk diameter (m) 0.178 0.214 0.227Canopy area (m2) 14.916 22.074 29.124Total leaves 10,192 15,714 25,702Single leaf area (cm2) 24.44 24.44 24.44Total leaf area (m2) 24.909 38.405 62.816Leaf area index 1.670 1.740 2.157

-30

-20

-10

0

10

20

30

40

50

10987654321

Rel

ati

ve

erro

r(%

)

Fig. 2. Relative error between mean estimated LAI without calibrating clumping

study, the estimated LAI without calibrating clumping index wassimilar to the direct LAI, possibly because the clumping index didnot significantly affect the estimated LAI by natural logarithm (lin)

-10

0

10

20

654321

Rel

ati

ve

erro

r(%

)

Fig. 1. Light incidence of the canopy.

.2.2.2. Beer–Lambert theory to calculate canopy parameters. Hemi-pherical images were also analyzed using Hemiview 2.1 (Delta-Tevices, Cambridge, UK), Winscanopy 2006a (Regent Instru-ents, Ste-Foy, Quebec) and Can-eye 6.1 (https://www4.paca.

nra.fr/can-eye/) and the estimated canopy parameters such as LAIPAI), gap fraction, openness and leaf inclination angle using differ-nt canopy models, such as the methods of Bonhom, 2000, 2000G,phere and ellipse model by natural logarithm (lin) and logarithmiclog), and the methods of CEV5.1, CEV6.1, P57, LAI 2000-3 rings, LAI000-4 rings and Miller. The leaf area density (l(h)) is the leaf areaer elevation at height h. For broad-leaved trees, LAI is estimatedy the following equation:

AI =∫ H

0

l (h) dh (6)

The probability of direct light passing through a gap betweeneaves on the horizontal surface (P0) is exactly given by the binomialxpression (Lang and Xiang, 1986):

0 =(

1 −(

AL

As

))N

(7)

here N is total effective leaves; AL is each leaf area randomly dis-ributed on a horizontal surface As. Since N tends to infinity, P0 cane calculated as:

0 = limn→∞

(1 − LAI

N

)N

= limn→∞

(1 − LAI

N

)(N/LAI)×LAI

= e−LAI (8)

here LAI = N·AL/As, and the gap fraction of different angles (Fig. 1)as expressed by the Poisson law (Clark and Murphy, 2011):

(�) = eG(�)×˝×LAI/ cos(�) (9)

here � is zenith angle, P(�) is canopy gap fraction, G(�) is the pro-ection of foliage relative to the ground, and ̋ is clumping indexeflecting the overlapping of the leaves, and can be calculated by

inscanopy software as follows

= Lin(LAI)Log(LAI)

(10)

here Lin(LAI) and Log(LAI) are LAI expressed by natural logarith-ic and logarithmic methods, respectively.

index and direct LAI. 1–5 and 6–10 were relative error of Bonhom, 2000, 2000G,sphere and ellipse model by natural logarithm (lin) and logarithmic (log) methods,respectively.

3. Results and discussion

3.1. Leaf area index by direct method

Table 2 shows that canopy radius was 2–3 m and canopy areawas 14–29 m2. The number of total green leaves per tree was10000–26000 and average single leaf area was 24.44 cm2. Thedirect LAI was 1.60–2.16 m2 m−2, with mean value of 1.85 m2 m−2.

3.2. Comparison of estimated leaf area index by differentsoftware packages and direct leaf area index

As shown in Fig. 2, during 20th July and 30th August, mean LAIvalue (trees A1, A3 and A4) obtained from hemispherical imagesusing different methods of Winscanopy software without calibrat-ing clumping index was similar to the direct LAI. The relative errorsof LAI in linear method of sphere model and logarithmic methodof 2000, 2000G and ellipse models were less than 10%, with theestimated LAI of 1.96, 1.93, 1.93 and 1.96 m2 m−2, respectively.However, Chianucci and Cutini (2013) suggested that the estimatedLAI from hemispherical images was lower than the direct LAI, andthe relative error is reduced after calibrating clumping index. In this

Fig. 3. Relative error between mean estimated PAI without calibrating clumpingindex and direct LAI. 1–6 were relative errors of CEV5.1, CEV6.1, P57, LAI 2000-3rings, LAI 2000-4 rings and Miller model, respectively.

Page 4: Canopy leaf area index for apple tree using hemispherical photography in arid region

C. Liu et al. / Scientia Horticulturae 164 (2013) 610–615 613

θa = 0.3906θc + 11.037

R² = 0.8266

15

20

25

30

30252015

Lin

θa= 0.6781θc + 6.3406

R² = 0.7124

30252015

Log

θa(º)

θc(º)

ted by automatic (�a) and manual (�c) exposure photography.

am2t

mwtact

mam2mvw

3h

ptmmmtatm

3W

3

tdi1((

L

LAIa = 0.6539 LAIc

R² = 0.9811

0

1

2

3

6420

LA

IaLAIc

3.4.3. Leaf inclination angleSeasonal variation of mean leaf inclination angle was estimated

by logarithmic ellipse method of Winscanpy software (Fig. 8). The

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

LA

I/P

AI

(m2

m-2

)

(2010 )

PAI(Ca ney e)

LAI( WinsCanopy)

PAI(He miview)

PAI(Photoshop)

Fig. 4. Comparison of mean leaf inclination angle estima

nd logarithmic (log) methods in 2000, 2000G, sphere and ellipseodel of Winscanopy software. Thus logarithmic method in 2000,

000G and ellipse models are suitable models in estimating LAI inhe arid apple orchard.

As shown in Fig. 3, plant area index (PAI) values estimated by theethods of CEV5.1 and LAI 2000-4 rings were 1.79 and 1.95 m2 m−2,hich coincided with the direct LAI. Demarez et al. (2008) indicated

hat the estimated PAI can replace tree LAI in certain range of zenithngles. The methods of CEV5.1 and LAI 2000-4 rings can be used toalculate LAI since the apple tree stem has no significant effect onhe LAI estimation.

The estimated LAI from hemispherical images using the ellipseethod of Hemiview software was 1.59 m2 m−2 during 20th July

nd 30th August, which was lower than the direct LAI. The esti-ated PAI from gap fraction analysis of Photoshop software was

.67 m2 m−2, which was higher than the direct LAI. So the logarith-ic ellipse method was used to estimate LAI and analyze seasonal

ariation of canopy parameters in the arid apple orchard of north-est China.

.3. Difference between automatic and manual exposureemispherical images

The leaf inclination angle obtained by manual exposurehotography was correlated with that by automatic exposure pho-ography, with r2 of 0.8266 and 0.7124 for linear and logarithm

ethods, respectively (Fig. 4). Fig. 5 shows that LAI obtained by theanual exposure photography was correlated with that by auto-atic exposure photography, with r2 of 0.9811. Possible reason is

hat the light is insufficient and the threshold to distinguish the skynd canopy is high, so that some small sky area is neglected. Andhe software cannot separate the stem area from leaves area in the

anual exposure photography.

.4. Seasonal variations of estimated canopy parameters byinscanopy software in apple orchard

.4.1. LAISeasonal LAI was estimated by the logarithmic ellipse method in

his study. Fig. 6 shows that the LAI of apple tree was low at the budevelopment and flowering stages (0.35 m2 m−2 in DOY 100), then

ncreased rapidly at leaf expanding stage (1.5–2.5 m2 m−2 in DOY50) and maintained at high level from the vigorous to harvest stageDOY 150–280). The exponential equation between LAI and DOY

day of year) using piecewise linear fitting was shown as follows:

AI2010 ={

0.035DOY − 3.46 DOY ≤ 153

1.93 DOY > 153(11)

Fig. 5. Comparison of mean estimated LAI by automatic (LAIa) and manual (LAIc)exposure photography.

LAI maintained 1.93 m2 m−2 after DOY 153 from Eq. (11). Greenet al. (1995) found that total leaf area in a 4 m-tall apple tree is35.5 m2 and the LAI is about 2.0 m2 m−2 by canopy radiation inter-cept method. Wünsche et al. (1996) indicated that LAI reached themaximum at 60 days after the budding.

3.4.2. Gap fractionGap fraction varied among different trees, and mainly dis-

tributed in south and north sides (Fig. 7), because the apple yard waspruned every winter. However, there was large difference of gapfraction between the west and east sides because the row spacingwas different in these two sides.

30027024021018015012090

DOY

Fig. 6. Seasonal variation of mean LAI in apple orchard.

Page 5: Canopy leaf area index for apple tree using hemispherical photography in arid region

614 C. Liu et al. / Scientia Horticulturae 164 (2013) 610–615

0.00.10.20.30.40.50.60.70.80.9

90 12 0 15 0 18 0 210 24 0 27 0 30 0

Gap

frac

�on

201 0(A1) NESW

0.00.10.20.30.40.50.60.70.80.9

90 12 0 15 0 18 0 21 0 24 0 27 0 30 0

Gap

frac

�on

2010(A3)NESW

0.00.10.20.30.40.50.60.70.80.9

90 12 0 15 0 180 21 0 24 0 27 0 30 0

Gap

frac

�on

2010(A4) NESW

Fig. 7. Seasonal variation of gap fra

θ= 0.0008 5 DOY2 0.39DOY + 63.20 -r² = 0.69

0.0

10.0

20.0

30.0

40.0

50.0

30027024021018015012090

Leaf

ang

le

DOY

MeanLeafAngle-Log

Fig. 8. Seasonal variation of mean leaf inclination angle.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 10 20 30 40 50 60 70 80 90

Lea

f d

ensi

ty(m

2m

-3)

Leaf angle (º)

(a) 2010/5/5

2010/5/24

2010/5/2 8

2010/6/8

2010/6/2 5

2010/7/1 6

2010/9/7

Fig. 9. Seasonal variation of relationship between leaf area density and leaf inclina

ction in different directions.

mean leaf inclination angle was larger in the early stage, and main-tained approximately 20◦ to obtain more direct radiation after DOY150. The leaf area density is leaf area per canopy volume reflect-ing the density of the leaves. With the advance of growth stage, theproportion of leaf inclination angle between 10◦ and 20◦ was larger,and leaf area density in the north side of apple tree was less thanthat in the south side (Fig. 9). Thus apple leaves can adjust theirinclination angles horizontally to receive more solar radiation forbiomass accumulation (Jones, 1992).

3.4.4. Clumping indexClumping index was 0.8 throughout the growing season of apple

tree (Fig. 10), possibly because the tree is only about 5 m high and ispruned every winter. The leaves were uniformly distributed whenclumping index was close to 1 (Seidel et al., 2011), so the effectof the clumping index can be neglected when the leaf area index

Leaf angle (º)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 10 20 30 40 50 60 70 80 90

Lea

f d

ensi

ty(m

2 m

-3)

(b) 2010/5/5

2010/5/24

2010/5/2 8

2010/6/8

2010/6/2 5

2010/7/1 6

2010/9/7

tion angle. (a) and (b) were north and south sides of apple tree, respectively.

Page 6: Canopy leaf area index for apple tree using hemispherical photography in arid region

C. Liu et al. / Scientia Horticultu

0.74

0.76

0.78

0.80

0.82

0.84

0.86

30027024021018015012090

Clu

mp

ing

in

dex

Wins canopy-El lips

os

4

dLa1tsaeoabguammf

A

eHIt(

R

B

C

C

of productivity in apple production systems: the role of light interception by

DOY

Fig. 10. Seasonal variation of clumping index.

f apple orchard was estimated from hemispherical image in thistudy.

. Conclusions

During DOY (day of year) 202–237 (at flourish stage), theirect leaf area index (LAI) was 1.85 m2 m−2, and the estimatedAI and plant area index (PAI) respectively by logarithmic ellipsend LAI 2000-4 rings methods of Caneye software were 1.96 and.95 m2 m−2, respectively, which were close to the direct LAI, sohe methods of logarithmic ellipse and LAI 2000-4 rings wereuitable in estimating the LAI from hemispherical images in thepple orchard of northwest China. LAI increased rapidly at the leafxpanding stage and then maintained a high level during the vigor-us to harvest stage. The gap fraction in the south and north sides ofpple tree was nearly the same. The leaf inclination angle tended toe horizontal to receive more radiation with the advance of applerowth. Moreover, the estimated LAI from hemispherical imagessing the methods of logarithmic ellipse and LAI 2000-4 rings wasccurate even without calibrating clumping index. Thus the esti-ated canopy parameters from hemispherical images using theethods of logarithmic ellipse and LAI 2000-4 rings can be used

or the modeling of evapotranspiration.

cknowledgments

We are grateful for grants from the National Natural Sci-nce Foundation of China (50939005 and 51309132), the Nationaligh-Tech 863 Project of China (2011AA100502), China-EU

nt’l Collaboration Projects (S2010GR0692) and Open founda-ion of Jiangsu Key Laboratory of Agricultural MeteorologyS5312041001).

eferences

réda, N.J.J., 2003. Ground-based measurements of leaf area index: a reviewof methods, instruments and current controversies. J. Exp. Bot. 54 (392),2403–2417.

hen, J.M., Black, T.A., 1992. Defining leaf area index for non-flat leaves. Plant CellEnviron. 15, 421–429.

hen, J.M., Rich, P.M., Gower, S.T., Norman, J.M., Plummer, S., 1997. Leaf area indexof boreal forests: theory, techniques, and measurements. J. Geophys. Res. 102(D24), 29429–29443.

rae 164 (2013) 610–615 615

Chen, J.M., Cihlar, J., 1995. Plant canopy gap-size analysis theory for improving opti-cal measurements of leaf-area index. Appl. Opt. 34 (27), 6211–6222.

Chianucci, F., Cutini, A., 2013. Estimation of canopy properties in deciduous forestswith digital hemispherical and cover photography. Agric. Forest Meteorol. 168,130–139.

Clark, J., Murphy, G., 2011. Estimating forest biomass components with hemispher-ical photography for Douglas-fir stands in northwest Oregon. Can. J. Forest Res.41, 1060–1074.

Demarez, V., Duthoit, S., Baret, F., Weissb, M., Dedieu, G., 2008. Estimation of leaf areaand clumping indexes of crops with hemispherical photographs. Agric. ForestMeteorol. 148, 644–655.

Green, S.R., McNaughton, K.G., Greer, D.H., McLeod, D.J., 1995. Measurement of theincreased PAR and net all-wave radiation absorption by an apple tree caused byapplying a reflective ground covering. Agric. Forest Meteorol. 76, 163–183.

Jonckheere, I., Flecka, S., Nackaerts, K., Muysa, B., Coppina, P., Weissb, M., Baret,F., 2004. Review of methods for in situ leaf area index determination: part I.Theories, sensors and hemispherical photography. Agric. Forest Meteorol. 121(1–2), 19–35.

Jones, H.G., 1992. Plants and Microclimate: A Quantitative Approach to EnvironmentPlant Physiology. Cambridge University Press, Great Britain, pp. 35–37.

Lang, A.R.G., 1991. Application of some of Cauchy’s theorems to estimation of surfaceareas of leaves, needles, and branches of plants, and light transmittance. Agric.Forest Meteorol. 55, 191–212.

Lang, A.R.G., Xiang, Y., 1986. Estimation of leaf area index from transmission of directsunlight in discontinuous canopies. Agric. Forest Meteorol. 37, 229–243.

Lang, M., Kuusk, A., Mõttus, M., 2010. Canopy gap fraction estimation from digi-tal hemispherical images using sky radiance models and a linear conversionmethod. Agric. Forest Meteorol. 150, 20–29.

Leblanc, S.G., Chen, J.M., Fernandes, R., Deeringd, D.W., Conleye, A., 2005. Method-ology comparison for canopy structure parameters extraction from digitalhemispherical photography in boreal forests. Agric. Forest Meteorol. 129,187–207.

Macfarlane, C., Arndt, S.K., Livesley, S.J., Edgara, A.C., Whiteb, D.A., Adamsd, M.A.,Eamuse, D., 2007a. Estimation of leaf area index in eucalypt forest with verticalfoliage,using cover and fullframe fisheye photography. Forest Ecol. Manage. 242,756–763.

Macfarlane, C., Grigg, A., Evangelista, C., 2007b. Estimating forest leaf area usingcover and fullframe fisheye photography: thinking inside the circle. Agric. ForestMeteorol. 146, 1–12.

Macfarlane, C., Hoffman, M., Eamus, D., Kerpd, N., Higginsone, S., McMurtrief, R.,Adamsa, M., 2007c. Estimation of leaf area index in eucalypt forest using digitalphotography. Agric. Forest Meteorol. 143, 176–188.

Marsal, J., Johnson, S., Casadesus, J., Lopez, G., Girona, J., Stöckle, C., 2014. Fraction ofcanopy intercepted radiation relates differently with crop coefficient dependingon the season and the fruit tree species. Agric. Forest Meteorol. 184, 1–11.

Rich, P.M., Clark, D.B., Clark, D.A., Oberbauer, S.F., 1993. Long-term study of solar radi-ation regimes in a tropical wet forest using quantum sensors and hemisphericalphotography. Agric. Forest Meteorol. 65 (1–2), 107–127.

Sandmann, M., Graefe, J., Feller, C., 2013. Optical methods for the non-destructiveestimation of leaf area index in kohlrabi and lettuce. Sci. Hortic. 156, 113–120.

Seidel, D., Fleck, S., Leuschner, C., Hammett, T., 2011. Review of ground-based meth-ods to measure the distribution of biomass in forest canopies. Ann. Forest Sci.68 (2), 225–244.

Shuttleworth, W.J., Wallace, J.S., 1985. Evaporation from sparse crops-an energycombination theory. Q. J. Roy. Meteor. Soc. 111 (469), 839–855.

Smith, M.L., Anderson, J., Fladeland, M., 2008. Field measurements for forest carbonmonitoring. Forest Canopy Struct. Prop., 179–196 (Chapter 14).

Stuckens, J., Dzikiti, S., Verstraeten, W.W., Verreynne, S., Swennen, R., Coppin, P.,2011. Physiological interpretation of a hyperspectral time series in a citrusorchard. Agric. Forest Meteorol. 151, 1002–1015.

Testi, L., Villalobos, F.J., Orgaz, F., 2004. Evapotranspiration of a young irrigated oliveorchard in southern Spain. Agric. Forest Meteorol. 121, 1–18.

Wang, W.M., Li, Z.L., Su, H.B., 2007. Comparison of leaf angle distribution func-tions: effects on extinction coefficient and fraction of sunlit foliage. Agric. ForestMeteorol. 143, 106–122.

Weiss, M., Baret, F., Smith, G.J., Jonckheere, I., Coppin, P., 2004. Review of methodsfor in situ leaf area index (LAI) determination: part II. Estimation of LAI, errorsand sampling. Agric. Forest Meteorol. 121, 37–53.

Wünsche, J.N., Laksoand, A.N., Robinson, T.L., Lenz, F., Denning, S.S., 1996. The bases

different shoot types. J. Amer. Soc. Hortic. Sci. 121 (5), 886–893.Zhang, Y.Q., Chen, J.M., Miller, J.R., 2005. Determining digital hemispherical pho-

tograph exposure for leaf area index estimation. Agric. Forest Meteorol. 133,166–181.