Vegetative and reproductive phenology of the mangrove Kandelia obovata

Vegetative and reproductive phenology of the mangroveKandelia obovatapsbi_367 1..12

MD. KAMRUZZAMAN,* SAHADEV SHARMA* and AKIO HAGIHARA†*Graduate School of Engineering and Science and †Laboratory of Ecology and Systematics, Faculty of Science, University of theRyukyus, Okinawa 903-0213, Japan

Abstract

Mangroves in the subtropical area of Japan are growing to their northern limits, yet littleis known of their phenology. The aim of the present study was to understand bothvegetative and reproductive phenology patterns, such as leaf emergence, leaf fall, budsetting, flowering, fruiting and propagule setting, in the mangrove Kandelia obovata. Thephenology of this species was assessed using litter-fall data for 5 years. Leaf and stipulelitter-falls continued with a clear monthly pattern throughout the years. New leaf pro-duction and leaf fall peaked in summer, immediately after the propagules fell. Leaf andstipule litter-falls were linked to monthly sunshine hour, and monthly mean air tempera-ture and monthly mean air relative humidity, respectively. Kandelia obovata had a dis-tinct flowering period, with the flowering phenophase starting in spring and continuinginto summer. Fruit initiation started at the end of summer and continued into autumn,whereas propagule production occurred during winter and spring. Flowering of K. obo-vata was influenced by monthly sunshine hour and monthly mean air temperature,whereas fruit and propagule litter-falls were not linked to any climatic factors. Thepresent results showed that a small portion (4.4%) of flowers developed into propagules.The average development period from flower buds to mature propagules was approxi-mately 11 months. Kendall’s consistency coefficient suggested that the monthly trends invegetative and reproductive litter-fall components, except for branches, did not changesignificantly among years.

Keywords: development phase, litter-fall, monthly variation, phenogram, time lag.

Received 11 May 2011; revision received 24 September 2011; accepted 18 November 2011

Introduction

Mangrove forests are among the world’s most productiveecosystems and sustain a variety of marine and estuarinecommunities (Lugo & Snedaker 1975; Boto & Bunt 1982).However, mangroves are also one of the world’s mostthreatened tropical and subtropical ecosystems and arebeing degraded in most countries mainly because ofanthropogenic activities and unsustainable exploitation(Khan et al. 2007). Mangroves are the only tall tree forestssituated between the land and sea (Kathiresan & Bingham2001; Alongi 2002) and they are tightly bound to thecoastal environments in which they occur. Once estab-lished, mangroves offer recreational potential, a sustain-

able supply of seafood for aquatic animals and usefulproducts for community subsistence (Alongi 1996; Tyagi& Singh 1998). As a primary producer, mangroves alsoserve as food for herbivores and detritivores. Litter-fall,consisting primarily of mangrove leaves, becomes avail-able following leaf senescence and death.

Leaf litter-fall studies in forest stands have been con-ducted for over a century. More recently, studies havefocused on the ecological role of leaf litter-fall in nutrientcycling and the possible interactions with biotic andabiotic variables (Zhou et al. 2007). Along with providingnutrients for the tree, leaf litter-fall also provides energyand a living environment for soil fauna and microorgan-isms (Guo & Sims 1999). Litter-fall is an indicator of thephenological events of a species (Duke 1990; Clarke 1994).Phenological information is essential for understating eco-system functioning (Janzen 1967) and to better manage

Correspondence: Akio HagiharaEmail: [email protected]

Plant Species Biology (2012) ••, ••–•• doi: 10.1111/j.1442-1984.2012.00367.x

bs_bs_banner

© 2012 The Society for the Study of Species Biology

mangrove ecosystems. Despite the importance of man-grove ecosystems, very few studies have considered thephenology of different mangrove species (Christensen &Wium-Andersen 1977; Wium-Andersen & Christensen1978; Wium-Andersen 1981; Duke et al. 1984; Duke 1990;FAO 1994; Gwada et al. 2000; Tyagi 2003; Coupland et al.2005; Mehlig 2006; Nagarajan et al. 2008). Most phenologi-cal studies have been confined to monitoring floweringand fruiting by direct observation, and very few studieshave combined direct observations with litter-fall mea-surements to assess the phenology of these species (Dukeet al. 1984). Phenological events in mangroves are influ-enced by local or regional environmental conditions, par-ticularly day length, air temperature, rainfall and waterstatus (Saenger & Moverley 1985; Naido 1989; Duke 1990;Fernandes 1999).

Kandelia obovata (S., L.) Yong in regions of China andJapan has been reclassified as a new species that waspreviously recognized as Kandelia candel (L.) Druce (Sheueet al. 2003). According to Sheue et al. (2003), K. obovata isdistributed from the Gulf of Tonkin northeastward toKwangtung, Fukien, Taiwan, the Ryukyus and southernJapan. Although this species grows to its northern limit,and mangrove forests grow under ambient climatic con-ditions, only a few studies have investigated the phenol-ogy of K. obovata. Leaf phenological traits only werereported by Gwada et al. (2000), and no previous studiesexamined the vegetative and reproductive phenology ofK. obovata. Therefore, the aim of the present study was todetermine the vegetative and reproductive phenology of amonospecific stand of K. obovata trees. The present studyalso aimed to identify the effect of climatic variables onK. obovata phenology and to determine the expansionperiod of each reproductive organ.

Materials and methods

Study site

The present study was conducted at the Manko wetland,located on Okinawa Island, Japan (26°11′N, 127°40′E), forover 5 years, starting in April 2006 and ending in March2011. The study area (Fig. 1) was designated as a wetlandof international importance under the Ramsar Conventionand is in a subtropical region; the warmth index of thestudy area (i.e. cumulative temperature of mean monthlytemperatures > 5°C) was 219.96 � 1.31°C month, which iswithin the 180–240°C month of a subtropical regiondefined by Kira (1991). Mangrove soil is muddy and theincoming water from the Kokuba River occasionallycarries a heavy load of sediments. The study site had fourpronounced seasons: winter (December–February),spring (March–May), summer (June–August) and autumn(September–November). Regular tidal inundation occurs

at the study site, which also receives some fresh watersupplied by run-off from an adjacent area (Khan et al.2004). A monospecific stand of K. obovata dominates thestudy site. Other mangrove species, such as Rhizophorastylosa Griff., Bruguiera gymnorrhiza (L.) Lamk. and Exco-ecaria agallocha L., are present in patches. Tree density,mean tree height (H) and mean stem diameter at H/10(D0.1H) were 21 456/ha, 4.27 � 0.87 m (mean � standarderror [SE]) and 5.09 � 0.04 cm (mean � SE) in 2010,respectively.

Climate data

Meteorological data were collected from the OkinawaMeteorological Observatory, Naha, from April 2006 toMarch 2011 (i.e. over the time period in which the field-work was conducted). The temperature fluctuatedapproximately 15°C from the coldest month to the hottestmonth, and normal of the mean annual air temperature

Fig. 1 Location of the study area. The hatched zone indicates themangrove area. The bold line denotes the belt transect.

2 M D . K A M R U Z Z A M A N E T A L .

© 2012 The Society for the Study of Species Biology Plant Species Biology ••, ••–••

during the study period was 23.3 � 0.2°C (Fig. 2a). Themean annual sunshine hour was 1709.7 � 73.8 h/year(mean � SE) (Fig. 2a). Rainfall varied throughout theyear, but during most months it exceeded 100 mm/month (Fig. 2b). The highest rainfall ranged from 20 mm/month in December 2008 to 594 mm/month in August2009 during the study period, and the mean annual rain-fall was 2226.5 � 267.3 mm/year (mean � SE). Most rain-fall occurred during summer with less during winter. Themean monthly relative humidity was 72.6 � 0.6%(mean � SE) (Fig. 2b).

Litter-fall collection

A 125-m long belt transect (5 m wide) was establishedperpendicular to the river current in the dense stand ofK. obovata, and then divided into 25 plots (5 m ¥ 5 meach). The total area of the belt transect was 625 m2, indi-cated by the bold line in Figure 1. The gradient of theforest floor was almost flat. The phenology of this specieswas assessed using litter-fall data. Litter-fall data werecollected using 50 net litter traps with a 1-mm mesh sizeand a mouth area of 0.2 m2. Two litter traps were placed in

Fig. 2 Climatic diagram of the four climatic factors examined (temperature, sunshine, humidity and rainfall). Data were obtained from theOkinawa Meteorological Observatory, Naha, Okinawa, Japan.

P H E N O L O G Y O F K A N D E L I A O B O V A T A 3

Plant Species Biology ••, ••–•• © 2012 The Society for the Study of Species Biology

each plot; the traps were placed > 1 m from the ground toavoid tidal water. The litter traps were emptied monthly,and the collected litter-fall was kept in a cotton bag andcarried to the laboratory where it was separated intoleaves, stipules, branches, bud primordia, flower buds,flowers, fruits and propagules. The separated litter-fallcomponents were dried at 80°C over 48 h and thenweighed using a digital balance (EK-600H; A & DCompany, Tokyo, Japan). The numbers of reproductivecomponents, such as flowers, fruits and propagules, werecounted to estimate the success of fruit and propagulesets. Following Duke (1990), we computed reproductiveunits and estimated the following: (i) fruit set was calcu-lated as the percentage of fruit units (immaturefruits + mature fruits + propagules)/(flowers + immaturefruits + mature fruits + propagules); and (ii) propaguleset was calculated as the percentage of propaguleunits (propagules)/(flowers + immature fruits + maturefruits + propagules).

Statistical analysis

Kendall’s consistency coefficient, W, was used to deter-mine monthly variation in the vegetative and reproduc-tive components of litter-fall among years. The value of Wis within the range 0 � W � 1. When W = 1 the ranks of acomponent of the monthly litter-fall are the same amongyears and when W = 0 the ranks of the component arecompletely different among years. A stepwise multipleregression analysis was carried out to determine theeffects of the four climatic factors (Fig. 2) on K. obovataphenology using MA-MACRO/MRA (ver. 3.0, PracticalBusiness Education Institute, Tokyo, Japan). The correla-tion coefficient rk was calculated to identify the time lag kin maturation period between reproductive organs usingtheir time-series data.

Results

Vegetative phenology

New leaf production and leaf fall of K. obovata occurredcontinuously throughout the year, but showed a clearmonthly trend. Leaf litter-fall was closely linked withstipule litter-fall. Figure 3a shows leaf litter-fall, whichfollowed a monthly pattern, that is, the highest peakoccurred in July and the lowest in January. Stipule litter-fall, which is an indicator of leaf emergence, was highestduring summer (April–August). Figure 3b shows that thehighest peak of stipule litter-fall was in July and the lowestwas in January. Branch litter-fall, including small twigs,bark and large branches, did not show a clear monthlytrend (Fig. 3c). During typhoons, green leaves weredamaged and fell onto the ground individually or

together with living broken branches. These green leavescontributed excessive leaf litter-fall (Fig. 3a). Figure 3cpresents the excessive branch litter-fall data (e.g. July 2006,July 2007 and October 2010); this excessive litter-fall wasattributable to typhoon events with wind speeds exceed-ing 17.2 m/s. The Japan Meteorological Agency definesa strong typhoon as having wind speeds of 33–43 m/sand a very strong typhoon as having wind speeds of44–53 m/s. During the study period, 12 strong typhoonswere observed and 105 typhoons approached Okinawa.Leaves were the dominant element of litter-fall over theentire study period. The mean total litter-fall amount was991.1 � 35.4 g/m2/year (mean � SE), of which leaveswere the largest contributor, with an estimated amount of530.6 � 12.8 g/m2/year (Table 1). Table 1 presents theamounts of stipule and branch litter-fall, and their contri-butions to the total litter-fall.

As shown in Figure 4a, vegetative organ (includingleaves, stipules and branches) litter-fall showed a clearmonthly trend, with its highest peak in summer andlowest in winter. The W-values suggested that monthlytrends in the litter-falls of vegetative organs, exceptbranches, did not change significantly among years(Table 2). Leaf litter-fall was significantly correlated withmonthly sunshine hour and no other climatic factors,whereas stipule litter-fall was significantly correlated withthe monthly mean air relative humidity and the monthlymean air temperature rather than monthly rainfall andmonthly sunshine hour (Table 3).

Reproductive phenology

The reproductive cycle of K. obovata had a regular monthlyperiodicity (Fig. 5). Initiation of green bud primordia wasobserved in March and continued until April. It was verydifficult to separate the flower bud primordia from thelitter-fall. In most cases, flower bud primordia were con-fused with inflorescence segments because K. obovata inflo-rescences are dichotomously branched two or three times(Sheue et al. 2003). Bud primordia eventually developedinto green buds and then developed into mature buds.Flower bud litter-fall started in March and continued untilAugust, with a peak in April and May (Fig. 5a). Flower budlitter-fall was not observed in the litter traps from Septem-ber to February. Flowering started in May and continueduntil September with the most abundant flowering occur-ring in July and August, when most flowers fell into thetraps, after this time the number of flowers in the traps wasdrastically reduced (Fig. 5b). If a flower was not fertilized,it was aborted and then fell from the trees within a fewdays. Pollinated flowers also fell. Figure 5b shows that thehighest flower litter-fall was in August and the lowest wasin October. Flower litter-fall was not observed in the littertraps from November to April.



As shown in Figure 5c, fruiting started in August andcontinued until December. The highest fruit litter-fall wasobserved in October and November. No fruit litter-fallwas observed in the litter traps from February to July.Hypocotyls developed from fruits and then turned into

mature propagules. Propagules began to drop in Marchand continued dropping until May (Fig. 5d). In March,mostly immature propagules were found in the littertraps, whereas mostly mature propagules were found inthe litter traps during May. Figure 5d shows the clear

200

150

100

50

020

15

10

5

0300

250

200

150

100

50

0

Lea

f li

tter

-fal

l (g/

m2 )

Sti

pule

litt

er-f

all (

g/m

2 )B

ranc

h li

tter

-fal

l (g/

m2 )

(a)

(b)

(c)

A A A A A A AAA S S S S A SO O O ON N N ND D D DM M M M M M M MMJ J J J J J J J J J J JJ JF F F F O N D MJ F2006 2007 2008 2009 2010 2011

Fig. 3 Phenograms of the litter-fall of the vegetative components.



monthly trend in propagule litter-fall with a peak in Aprilto May. Neither mature nor immature propagules werefound in the litter traps from July to November. Theaverage total reproductive organ litter-fall was185.8 � 25.5 g/m2/year (mean � SE), which contributed18.83% to the total litter-fall (Table 1). Table 1 shows thedetailed amounts of flower bud primordium, flower bud,flower, fruit and propagule litter-falls, respectively, andtheir percentages of contribution to the total litter-fall.

Figure 4b indicates that litter-fall of the reproductiveorgans of K. obovata followed a very specific monthly peri-odicity. The calculated values of W suggest that thesemonthly trends in litter-fall of reproductive organs didnot change significantly among years (Table 2). As shownin Table 3, the flower bud litter-fall of K. obovata was sig-nificantly correlated with monthly mean air relativehumidity and monthly mean air temperature, whereasflower litter-fall was significantly correlated with monthlysunshine hour and monthly mean air temperature ratherthan other climatic factors. However, fruit and propagulelitter-falls showed no correlation with any climatic factors.

Similar to other mangrove species, K. obovata had avery small proportion of flowers that developed intopropagules. The results show that only 4.40 � 0.35%(mean � SE) of flowers developed into propagules, and16.70 � 0.91% (mean � SE) of the flowers developed intofruits, including both mature and immature fruits andpropagules.

As shown in Figure 6a, the development phase fromflower buds to flowers took 2 months. The developmentphase from flower buds to fruits took 5 months (Fig. 6b).After fertilization was completed, flowers turned intofruits. The developmental phase from flower buds tomature propagules took 11 months (Fig. 6c). Develop-ment from flowers to fruits took 3 months (Fig. 6d). As aspecies with viviparous germination, K. obovata followedthe initiation of hypocotyl from fruit. Development fromflowers to mature propagules and fruits to maturepropagules took 9 months (Fig. 6e) and 6 months (Fig. 6f),respectively.

Discussion

Vegetative phenology

Leaf litter-fall (Fig. 3a) and stipule litter-fall (Fig. 3b) ofK. obovata followed the same seasonal pattern, that is, leaffall and leaf production were lowest in winter and highestin summer. Another study in Ohura Bay, Okinawa Island,Japan, recorded a similar pattern of leaf fall and leafrecruitment of K. candel (Hardiwinoto et al. 1989). Thepresent study showed that peaks in leaf and stipule litter-falls occurred in the same month, whereas Duke (1990)reported a time lag of 1 month between leaf emergenceT

able

1A

nnua

lam

ount

ofve

geta

tive

and

repr

oduc

tive

litte

r-fa

llof

Kan

delia

obov

ata

over

the

stud

ype

riod

Lit

ter-

fall

com

pone

nts

1st

year

2nd

year

3rd

year

4th

year

5th

year

Mea

n

Lea

f59

4.58

�14

.05

(56.

8)60

0.75

�13

.73

(49.

7)54

5.30

�16

.82

(60.

0)48

2.28

�10

.05

(52.

6)42

9.96

�9.

09(4

9.2)

530.

58�

12.7

5(5

3.3)

Stip

ule

92.7

7�

2.66

(8.8

7)96

.95

�2.

08(8

.02)

82.7

3�

3.18

(9.1

0)52

.39

�2.

06(5

.72)

59.5

0�

2.14

(6.8

0)76

.87

�2.

42(7

.79)

Bra

nch

204.

82�

14.7

9(1

9.6)

357.

08�

19.2

6(2

9.5)

88.8

9�

10.6

5(9

.78)

159.

72�

9.87

(17.

4)17

8.69

�8.

61(2

0.4)

197.

84�

12.6

4(2

0.1)

Sum

ofth

eve

geta

tive

orga

ns89

2.17

�31

.51

(85.

3)10

54.7

9�

35.0

7(8

7.3)

716.

91�

30.6

5(7

8.9)

694.

39�

21.9

7(7

5.8)

668.

15�

19.8

3(7

6.4)

805.

28�

27.8

1(8

0.7)

Bud

prim

ord

ium

15.5

8�

1.07

(1.4

9)26

.17

�1.

57(2

.17)

21.7

8�

0.74

(2.4

0)32

.31

�0.

72(3

.52)

19.0

6�

0.41

(2.1

8)22

.98

�0.

90(2

.33)

Flow

erbu

d14

.58

�0.

61(1

.39)

10.5

7�

0.58

(0.8

7)5.

93�

0.27

(0.6

5)8.

88�

0.54

(0.9

7)12

.47

�0.

47(1

.42)

10.4

8�

0.49

(1.0

6)Fl

ower

8.72

�1.

12(0

.83)

20.1

1�

2.07

(1.6

6)25

.40

�2.

72(2

.79)

44.2

6�

2.09

(4.8

3)40

.57

�1.

88(4

.64)

27.8

1�

1.98

(2.8

2)Fr

uit

0.85

�0.

21(0

.08)

3.56

�0.

41(0

.29)

3.43

�0.

36(0

.38)

6.49

�0.

40(0

.71)

4.27

�0.

70(0

.49)

3.72

�0.

41(0

.38)

Prop

agul

e11

4.35

�27

.35

(10.

9)93

.45

�14

.09

(7.7

3)13

5.77

�22

.20

(14.

9)13

0.24

�22

.98

(14.

2)13

0.35

�22

.13

(14.

9)12

0.83

�21

.75

(12.

2)Su

mof

the

repr

oduc

tive

orga

ns15

4.07

�30

.36

(14.

7)15

3.86

�18

.72

(12.

7)19

2.32

�26

.29

(21.

2)22

2.17

�26

.72

(24.

2)20

6.72

�25

.59

(23.

6)18

5.83

�25

.54

(19.

3)

Tota

l10

46.2

4�

42.5

412

08.6

4�

32.5

090

9.24

�26

.47

916.

56�

48.7

087

4.87

�26

.53

991.

11�

35.3

5

Valu

esar

em

ean

�st

anda

rder

ror

(g/

m2 /

year

).N

umer

als

inpa

rent

hese

sre

pres

ent

the

perc

enta

geof

the

tota

lam

ount

.



and leaf fall for Avicennia marina in Australia, Papua NewGuinea and New Zealand. In contrast to our results, leaffall of K. obovata in Hong Kong, showed a bimodal pattern,with the highest peaks in spring (February–June) and in

late summer (August–November) (Lee 1989). The leaf pro-duction period followed the same seasonal pattern as thatfound in other mangrove species. For example, Rhizophoraapiculata BL. in southern Thailand and A. marina (Forssk.)

500

400

300

200

100

0

500

400

300

200

100

0

100

80

60

40

20

0

Tota

l lit

ter-

fall

(g/

m2 )

Rep

rodu

ctiv

e or

gans

litt

er-f

all (

g/m

2 )V

eget

ativ

e or

gans

litt

er-f

all (

g/m

2 )

(a)

(b)

(c)

A A A A A A AAA S S S S A SO O O ON N N ND D D DM M M M M M M MMJ J J J J J J J J J J JJ JF F F F O N D MJ F2006 2007 2008 2009 2010 2011

Fig. 4 Phenograms of the litter-fall of the vegetative and reproductive components and the total litter-fall.



Vierh., Ceriops australis (White) Ballment, Smith & Stod-dart, Rhizophora stylosa Griff., and Sonneratia alba J. Simth.in northern Australia showed their highest peaks duringthe wet summer season, that is, in the monsoonal seasonwith maximum temperature and rainfall (Christensen &Wium-Andersen 1977; Coupland et al. 2005, 2006).

Litter-fall seasonality is a common feature in moststudies, but the cause of its seasonality is unclear (Gill &Tomlinson 1971; Lugo & Snedaker 1975; Wium-Andersen& Christensen 1978). The data presented in the presentstudy demonstrate that leaf and stipule litter-fall of K. obo-vata are significantly correlated with monthly sunshinehour, and monthly mean air temperature and monthlymean air relative humidity, respectively. Our results aresupported by the findings of Nakagoshi and Nehira(1986), who reported that cumulative temperature wasdirectly proportional to the number of leaves and the rateof leaf production in mangrove seedlings. A similarfinding was observed in K. obovata (Gwada et al. 2000). Incontrast to our results, most species in a tropical ever-green mountain rainforest show a good correlationbetween leaf shedding and precipitation (Bendix et al.2006) and also display a bimodal peak in leaf productionthat is correlated with the rainy periods (Medway 1972).Kandelia obovata grows in intertidal areas and shows adistinctly unimodal pattern, in contrast to tropical rainfor-

ests with a bimodal or multimodal pattern. High speciesdiversity and appropriate humidity in tropical rainforestscould cause multi-mode peaks in litter-fall. The diversityand water conditions in tropical rainforests differ substan-tially from mangrove ecosystems.

The present study showed that the leaves of K. obovatawere dominant (53.3%) in the total litter-fall (Table 1). Ourresults coincide with the findings of Wafar et al. (1997),who reported that the percentage contribution of leaflitter-fall to total litter-fall in the mangrove species Rhizo-phora apiculata Blume., Rhizophora mucronata Lam., Son-neratia alba J. E. Smith and Avicennia officinalis L. on thecentral west coast of India were 61.2, 61, 57.7 and 43%,respectively. The contribution of leaf litter-fall to totallitter-fall in K. obovata was lower than that recorded inother mangrove species such as A. marina in Kenya, whereleaf litter-fall constituted 514.6 g/m2 (83%) of the totallitter-fall (Ochieng & Erftemeijer 2002). Another study onA. marina var. australasica in New Zealand found that leaflitter-fall contributed 140–200 g/m2 (73%) to the totallitter-fall (May 1999). The contribution (7.79%) of thestipule litter-fall of the present K. obovata was similar tothat of R. apiculata (7%) and R. mucronata (6.3%) in India(Wafar et al. 1997).

Branch litter-fall showed a less clear monthly patternthan leaf and stipule litter-fall (Fig. 3). In a typhoonmonth, branch litter-fall was excessive compared with thelitter-fall in other months because all dead branchesincluding broken living branches fell to the ground. Thisphenomenon shows the strong effect of typhoons. Thepresent observations support the findings of Hardiwinotoet al. (1989), who reported that typhoons had strongeffects on branch litter-fall in Ohura Bay, Okinawa Island,Japan.

Reproductive phenology

We found that flower bud initiation in K. obovata startedduring spring (Fig. 5a), whereas flowering occurredduring summer (Fig. 5b). During spring and summer,temperature, sunshine hour, rainfall and humidity wereat their maximum (Fig. 2). Our results showed that flower

Table 2 Kendall’s consistency coefficient, W, for the vegetativeand reproductive organs of Kandelia obovata over the study period

Litter-fall component W c2 (P-value)

Leaf 0.7589 41.74 (<0.001)Stipule 0.7192 39.55 (<0.001)Branch 0.3270 17.99 (0.0813)Flower bud primordium 0.6120 33.66 (<0.001)Flower bud 0.4775 26.26 (<0.01)Flower 0.8043 44.24 (<0.001)Fruit 0.6867 37.77 (<0.001)Propagule 0.6504 35.77 (<0.001)Total litter-fall 0.8151 44.83 (<0.001)

Degrees of freedom of the c2 distribution are 11.

Table 3 Results of the adjusted R2 usingthe stepwise method of multiple regres-sion analysis for vegetative and reproduc-tive litter-fall components for Kandeliaobovata in relation to environmental factors

ComponentTemperature

(°C)Sunshine

(h/month)Humidity

(%)Rainfall

(mm/month)

Leaf — 0.426 (<0.001) — —Stipule 0.324 (<0.001) — 0.292 (<0.001) —Flower bud 0.348 (<0.001) — 0.261 (<0.001) —Flower 0.309 (<0.001) 0.297 (<0.001) — —Fruit — — — —Propagule — — — —

Numerals in parentheses show the significance probability.



8

7

6

5

4

3

2

1

0

6

5

4

3

2

1

080

70

60

50

40

30

20

10

0

30

25

20

15

10

5

0

A A A A A A AAA S S S S A SO O O ON N N ND D D DM M M M M M M MMJ J J J J J J J J J J JJ JF F F F ON D MJ F2006 2007 2008 2009 2010 2011

Flo

wer

bud

litt

er-f

all (

g/m

2 )F

low

er li

tter

-fal

l (g/

m2 )

Frui

t lit

ter-

fall

(g/

m2 )

Pro

pagu

les

litt

er-f

all (

g/m

2 )

(a)

(b)

(c)

(d)

Fig. 5 Phenograms of the litter-fall of the reproductive components.



bud and flower litter-falls of K. obovata were significantlycorrelated with monthly mean air relative humidity andmonthly mean air temperature, and monthly sunshinehour and monthly mean air temperature, respectively, in

agreement with Duke’s (1990) finding that climatic factorssuch as day length and temperature have a strong influ-ence on A. marina flowering in Australia, Papua NewGuinea and New Zealand. Similar findings have been

Fig. 6 Cross correlation coefficients between pairs of the reproductive components of Kandelia obovata.



reported by Wium-Andersen and Christensen (1978),who observed flower buds and flowers of C. tagal insouthern Thailand in April to June, consistent with thewet season. The peak period of flower bud (Fig. 5a) andflower production (Fig. 5b) of K. obovata coincided withthe peak period of leaf emergence (stipule litter-fall)(Fig. 3b), suggesting that growth of reproductive organsaffects apex vigor (Gill & Tomlinson 1971).

Flowering and fruiting of R. apiculata, R. mucronata,S. alba and A. officinalis were found to be confined to spe-cific periods during a year, and the contributions of thesespecies to flower and fruit litter-falls to total litter-fallwere 21.8, 23.2, 17.3 and 18.1%, respectively (Wafar et al.1997). Our results showed that the contribution of thereproductive components of K. obovata to the total litter-fall was very similar to other mangrove species reportedby Wafar et al. (1997). According to May (1999), the flowerand fruit litter-fall of A. marina var. australasica contrib-uted 11% to the total litter-fall.

This is the first study to report the maturation periodsof reproductive organs of K. obovata at the northernlimit of its geographical distribution. The developmentalphase of flower bud initiation to mature propagules took11 months (Fig. 6). Gill and Tomlinson (1971) reportedthat the development time to mature propagules forRhizophora mangle L. in Florida was 8–13 months. Thedevelopmental phase of bud initiation to maturepropagules was 6–8 months in Bruguiera cylindrica (L.) BL.and 16–21 months in Ceriops tagal (Perr.) C.B. Rob. insouthern Thailand (Wium-Andersen & Christensen 1978).According to Christensen and Wium-Andersen (1977),R. apiculata in southern Thailand took nearly 3 years todevelop from flower bud primordia to maturepropagules. In comparison with already reported data ofspecies belonging to the family Rhizophoraceae, itappears that K. obovata growing at the northern limit of itsgeographical distribution has a relatively short matura-tion period.

The present study found that a small portion of flowersdeveloped into propagules (4.4%). Although K. obovataproduces a large number of flowers, the flowers are smallin size, self-compatible and pollinated by small insects,similar to K. candel (Tomlinson 1994). A similar observa-tion was made by Gill and Tomlinson (1971), whoreported that 0–7.2% of R. mangle flowers were fertilized.Although 16.7% of the fertilized flowers developed intofruits, not all of the fruits grew, which may have beenbecause of the energy required during development aswell as tree size. This rate was higher than the range of2.4–5.3% for R. stylosa, R. mangle and B. gymnorrhizareported by Tyagi (2003) and for R. stylosa and C. australisreported by Coupland et al. (2006). It has been suggestedthat K. obovata has the least specialized pollination amongthe Rhizophoraceae species as several types of small

insects visit its flowers (Tomlinson 1994). Entomophilouspollination could be connected with the relatively highfruit set of K. obovata.

Acknowledgments

We are grateful to Dr K. Analuddin, Dr R. Suwa, Dr A. T.M. Rafiqul Hoque and Dr W. Min as well as Mr K.Mouctar for their invaluable help during the fieldwork.Thanks go to Ms R. Deshar and Ms C. Fengxia for theirhelp and useful comments on an earlier draft of the manu-script. This study was partially supported by a Grant-in-Aid for Scientific Research (no. 23380094) from theMinistry of Education, Culture, Sport, Science and Tech-nology, Japan.

References

Alongi D. M. (1996) The dynamics of benthic nutrient pools andfluxes in tropical mangrove forests. Journal of Marine Research54: 123–148.

Alongi D. M. (2002) Present status and future of the world’smangrove forests. Environmental Conservation 9: 331–349.

Bendix J., Homeier J., Cueva Ortiz E., Emck P., Breckle W. S.,Richter M. & Beck E. (2006) Seasonality of weather and treephenology in a tropical evergreen mountain rain forest. Inter-national Journal of Biometeorology 50: 370–384.

Boto K. G. & Bunt J. S. (1982) Carbon export from mangroves. In:Clough B. F. (ed.). Mangrove Ecosystems in Australia. ANUPres, Canberra, pp. 239–257.

Christensen B. & Wium-Andersen S. (1977) Seasonal growth ofmangrove trees in southern Thailand. I. Phenology of Rhizo-phora apiculata BL. Aquatic Botany 3: 281–286.

Clarke P. J. (1994) Baseline studies of temperate mangrove growthand reproduction; demographic and litterfall measures ofleafing and flowering. Australian Journal of Botany 42: 37–48.

Coupland G. T., Paling E. I. & McGuinness K. A. (2005) Vegetativeand reproductive phonologies of four mangrove species fromnorthern Australia. Australian Journal of Botany 53: 109–117.

Coupland G. T., Paling E. I. & McGuinness K. A. (2006) Floralabortion and pollination in four species of tropical mangrovesfrom northern Australia. Aquatic Botany 84: 151–157.

Duke N. C. (1990) Phenological trends with latitude in the man-grove tree Avicennia marina. Journal of Ecology 78: 113–133.

Duke N. C., Bunt J. S. & Williams W. T. (1984) Observations on thefloral and vegetative phonologies of north-eastern Australianmangroves. Australian Journal of Botany 32: 87–99.

FAO (1994) Mangrove Forest Management Guidelines. FAO ForestryPaper No. 117. Food and Agricultural Organization of theUnited Nations, Rome.

Fernandes M. E. B. (1999) Phenological patterns of Rhizophora L.,Avicennia L. and Laguncularia Gaertn. f. in Amazonian man-grove swamps. Hydrobiologia 413: 53–62.

Gill A. M. & Tomlinson P. B. (1971) Studies on the growth of redmangrove (Rhizophora magle L.). 3. Phenology of the shoot.Biotropica 3: 109–124.

Guo L. B. & Sims R. E. H. (1999) Litter decomposition and nutri-ent release via litter decomposition in New Zealand eucalypt



short rotation forests. Agriculture, Ecosystems and Environ-ments 75: 133–140.

Gwada P., Makoto T. & Uezu Y. (2000) Leaf phenological traits inthe mangrove Kandelia candel (L.) Druce. Aquatic Botany 68:1–14.

Hardiwinoto S., Nakasuga T. & Takabatake S. (1989) Litter pro-duction and decomposition of a mangrove forest at OhuraBay, Okinawa. Research Bulletins of the College ExperimentForest, Hokkaido University 46: 577–594.

Janzen D. H. (1967)) Synchronization of sexual reproduction oftrees within the dry season in Central America. Evolution 21:620–673.

Kathiresan K. & Bingham B. L. (2001) Biology of mangroves andmangrove ecosystems. Advances in Marine Biology 40: 81–251.

Khan M. N. I., Suwa R. & Hagihara A. (2007) Carbon and nitro-gen pools in a mangrove stand of Kandelia obovata (S., L.) Yong:vertical distribution in the soil–vegetation system. WetlandsEcology and Management 15: 141–153.

Khan M. N. I., Suwa R., Hagihara A. & Ogawa K. (2004) Intercep-tion of photosynthetic photon flux density in a mangrovestand of Kandelia candel (L.) Druce. Journal of Forest Research 9:205–210.

Kira T. (1991) Forest ecosystems of East and SoutheastAsia in a global perspective. Ecological Research 6: 185–200.

Lee S. Y. (1989) Litter production and turnover of the mangroveKandelia candel (L.) Druce in a Hong Kong tidal shrimp pond.Estuarine, Coast and Shelf Science 29: 75–87.

Lugo A. E. & Snedaker S. C. (1975) Properties of a mangroveforest in southern Florida. In: Walsh G. E., Snedaker S. C. &Teas H. J. (eds). Proceedings of the International Symposium onBiology and Management of Mangroves. University of Florida,Gainesville, pp. 170–212.

May J. D. (1999) Spatial variation in litter production by the man-grove Avicennia marina var. australasica in Rangaunu Harbour,Northland, New Zealand. New Zealand Journal of Marine andFreshwater Research 33: 163–172.

Medway L. (1972) Phenology of a tropical rain forest in Malaya.Biological Journal of the Linnaean Society 4: 117–146.

Mehlig U. (2006) Phenology of the red mangrove, Rhizophoramangle L. in the Caete Estuary, Para, equatorial Brazil. AquaticBotany 84: 158–164.

Nagarajan B., Pandiarajan C., Krishnamoorthy M. & Sophia P.(2008) Reproductive fitness and success in mangroves: impli-cation on conservation. Sengupta M. & Dalwani R. (eds). Pro-ceedings of Taal 2007. The 12th World Lake Conference, Jaipur,India; 28 October to 2 November 2007, pp. 29–33.

Naido G. (1989) Seasonal plant water relations in a South Africanmangrove swamp. Aquatic Botany 33: 233–242.

Nakagoshi K. & Nehira K. (1986) Growth and mortality of man-grove seedlings transplanted to Hiroshima. Hikobia 9: 439–449.

Ochieng C. A. & Erftemeijer P. L. A. (2002) Phenology, litterfalland nutrient resorption in Avicennia marina (Forssk.) Vierh inGazi Bay, Kenya. Trees 16: 167–171.

Saenger P. & Moverley J. (1985) Vegetative phenology of man-groves along the Queensland coastline. Proceedings of the Eco-logical Society of Australia 13: 257–265.

Sheue C. R., Liu H. Y. & Yong J. W. H. (2003) Kandelia obovata(Rhizophoraceae), a new mangrove species from EasternAsia. Taxon 52: 287–294.

Tomlinson P. B. (1994) The Botany of Mangroves. Cambridge Uni-versity Press, New York.

Tyagi A. P. (2003) Location and interseasonal variation in flower-ing, propagule setting and propagule size in mangrovesspecies of the family Rhizophoraceae. Wetlands Ecology andManagement 11: 167–174.

Tyagi A. P. & Singh V. V. (1998) Pollen fertility and intraspecificcompatibility in mangroves of Fiji. Sexual Plant Reproduction11: 60–63.

Wafar S., Untawale A. G. & Wafar M. (1997) Litterfall and energyflux in a mangrove ecosystem. Estuarine, Coastal and ShelfScience 44: 111–124.

Wium-Andersen S. (1981) Seasonal growth of mangrove trees insouthern Thailand. III. Phenology of Rhizophora mucronataLamk. and Scyphiphora hydrophyllaceae Gaertn. Aquatic Botany10: 371–376.

Wium-Andersen S. & Christensen B. (1978) Seasonal growth ofmangrove trees in southern Thailand. II. Phenology of Bru-guiera cylindrica, Ceriops tagal, Lumnitzera littorea and Avicenniamarina. Aquatic Botany 5: 383–390.

Zhou G., Guan L., Wei X., Zhang D., Zhang Q., Yan J., Wen D., LiuJ., Liu S., Huang Z., Kong G., Mo J. & Yu Q. (2007) Litterfallproduction along successional and altitudinal gradients ofsubtropical monsoon evergreen broadleaved forests inGuangdong, China. Plant Ecology 188: 77–89.



Documents

Vegetative and reproductive phenology of the mangrove Kandelia obovata