Species diversity of Chinese beech forests in relation ... - Springer Link

3 downloads 0 Views 1MB Size Report
Species diversity of Chinese beech forests in relation to warmth and climatic disturbances. KUN-FANG CAO 1 AND ROB PETERS 2. Center for Ecological ...
Ecological Research (1997) 12, 175-189

Species diversity of Chinese beech forests in relation to warmth and climatic disturbances KUN-FANG CAO1 AND ROB PETERS2

Center for Ecological Research, Institute of Botany, Chinese Academy of Sciences, 20 Nanxincun, Xianshan, Haidian District, Beijing 100093, China and 2Department of Forestry, Wageningen Agricultural University, PO Box 342, 6700 A H Wageningen, The Netherlands The trends in the occurrence of climatic disturbances in the Chinese Fagus range are described, and the relationship between woody species diversity and climatic factors in eight old-growth Chinese beech forests is characterized. In the Chinese Fagus range that lies in the humid mountains of southern China, wind storms and heavy rain frequency increase towards the eastern coast. Thunderstorm frequency increases southwards. Snowfall frequency increases northwards. Glaze storm frequency peaks in the center near Lake Dongtian, but much higher in the east than in the west. Hailstorm frequency also peaks in the center. The forests sampled in this study are widely separated. Their canopies consist of either deciduous broad-leaved trees or a mixture of evergreen and deciduous broad-leaved trees. Their species diversity increases towards warmer sites and towards the east. The importance of the evergreen trees in relation to warmth and minimum temperature increases southwards. Our analysis suggests that wind storms and heavy rains enhance the species diversity of Chinese beech forests. Cold disturbances such as glaze and snow diminish the diversity and canopy dominance of evergreen broad-leaved trees but favor deciduous broad-leaved trees, especially beech. The annual precipitation received by the forests in this study varies from 1400-2550 mm. This is not correlated with diversity, however, probably because all of these forests grow in humid conditions with sufficient water being supplied by precipitation throughout the year. Key words: disturbance-diversity theory; energy-diversity theory; evergreen versus deciduous; old-growth

Fagus forests; southern China.

INTRODUCTION Temperature and precipitation are key factors that determine species ranges (Wolfe 1979; Box 1981; Walter 1985; Woodward 1987). Five beech species occur in southern China in a wide climatic range:

Fagus longipetiolata, Fagus lucida, Fagus engleriana, Fagus hayatae and Fagus chienii (Cao et al. 1995). Water deficit prevents beeches from occurring in temperate northern China (Cao et al. 1995). Chinese beeches are restricted to the mountains of southern China where the climate is mainly temperate and humid or very humid (Cao et al. 1995). The beech region is mainly in the evergreen broadleaved forest region of China (region 3 in Hou Received 22 April 1996. Accepted 13 February 1997.

1983, fig. 44). This region has a rich flora partly because glaciers barely made an impact on it (Hsii 1983). Beech species dominate many montane forests of this region (Wu 1980). In the north and at high altitudes, the canopies of these beech forests are dominated by deciduous broad-leaved trees only, and in the south and at low altitudes, by both evergreen and deciduous broad-leaved trees (Wu 1980; Qi 1990). These forests largely vary in species diversity (Cao 1995). Species diversity of forests is influenced by many abiotic factors such as temperature, precipitation, floristic history, disturbances, topography and soil. Although the species diversity of similar biomes varies among different geographical regions and the highest diversities occur away from the equator (e.g. Latham & Ricklefs 1993), species diversity generally increases towards the equator (Whittaker

176

K-E Cao and R. Peters

1975; see also review by Rohde 1992 and references therein). Climatic heat is considered to be the primary cause of species diversity (Currie & Paquin 1987; Currie 1991), an idea that is referred to as the 'energy-diversity theory' below. It is increasingly appreciated that disturbances largely affect the species diversity of biomes such as forests (e.g. Connell 1978; Platt & Strong 1989; Oldeman 1983, 1990; Ashton 1993). Many kinds of climatic disturbances can open up the forest canopies and change forest structure and function locally (Pickett & White 1985). Beneath canopy openings, the abiotic and biotic resources, and occasionally microtopography, differ from those under a closed canopy (Bormann & Likens 1979; Putz 1983; Nakashizuka 1989; Platt & Strong 1989). The opening up of the forest canopy gives tree seedlings a chance to establish themselves. Furthermore, suppressed trees may resume their growth and some trees eventually grow into the forest canopy. Consequently, the climatic events that can open up the forest canopy largely determine the presence and dominance of trees in forests. Connell (1978) suggested that a forest receiving frequent and intense disturbances should be dominated by short-lived, light-demanding trees and have a low diversity; a forest with little disturbance is dominated by longlived, shade-tolerant species and also has a low diversity; and one with intermediate disturbance has the greatest diversity. This concept is referred to as the 'disturbance-diversity theory' below. Climatic disturbances in temperate forests may involve wind storms, heavy rains, glaze storms, and snow storms (Pickett & White 1985; Bryant 1991). Southern China is regularly affected by typhoons coming from the Pacific Ocean, and wind storms occur more frequently and intensely along and near the coast than in the western interior (Feng et al. 1985). Heavy rainfall occasionally causes landslides in mountain forests in southern China (e.g. Liu & Xu 1991). Glaze storms are observed to cause major disturbances in montane forests, including beech forests near Lake Dongtian in central southern China (Wang 1984; Jiang 1991). Glaze affects evergreen broad-leaved trees more than deciduous ones and affects fast-growing trees more than slowgrowing ones (Wang 1984; cf. Downs 1938, in north-eastern USA). Thunderstorms can cause strong gusts and heavy rainfall (Nieuwolt 1977) and, therefore, cause disturbance to forests. Also,

lightning during thunderstorms can directly kill canopy trees (cf. Brtinig & Huang 1989, in a Borneo forest). In Chinese beech forests, tree-falls and branch-breakage may be caused by wind storms (Feng et al. 1985; Zhang et al. 1991), glaze storms (Wang 1984; Jiang 1991), thunderstorms, heavy rains (Sun 1988; Liu & Xu 1991), heavy snowfall (Wang et aL 1965; Anonymous 1986), and hailstorms (Wang 1990; Xu 1991). Chinese beech forests grow in the single evergreen broad-leaved forest region (Hou 1983; Cao et aL 1995) and on similar soils (yellow-brown forest soils; cf. Li & Sun 1990). However, they are subject to various types of climatic disturbances with varying frequency and intensity (see above; Feng et al. 1985). In forests dominated by beech and mixed with evergreen broad-leaved trees, young beeches are rare and evergreen juveniles are abundant (Cao 1995 and references therein). However, Cao (1995) suggests that beech dominance in these mixed forests will be maintained, as disturbances such as glaze and snow storms will occasionally cause major disturbances that particularly affect evergreen broad-leaved trees and provide opportunities for beech to regenerate. In this paper, we describe the general trends in frequency and intensity of major disturbances that create canopy gaps in Chinese beech forests, and characterize the relationship between species diversity and temperature, and climatic disturbances. Wind storms, thunderstorms, heavy rainfall, hailstorms, glaze storms and snow cause climatic disturbances in Chinese beech forests.

METHODS Study sites and plot sampling Eight beech forests widely separated in Chinese beech range were selected for the study (Fig. 1; Table 1). A forest was defined as a beech forest if at least one beech species was dominant. They were all old-growth, natural forests and occurred in the montane belts at altitudes varying from 1100 to 1900 m a.s.1. In the eight forests, the soils were all yellow-brown forest soils (sensu Li & Sun 1990), which belong to dystric and gleyic Cambisols according to the FAO-UNESCO (1988) soil classification, and had a pH range from about 4.5-5.5 (Li & Sun 1990). These eight beech forests experi-

Disturbances and diversity in Chinese beech forests /'95Loo

xool. oo

~

'~*o.st.oo"j

u.c'.oo

~o.oo

--

,

/

~

,~

i

,a *zo~.oo ng

o.oo

~ Shaanxi % ~' e n' a n~Hena ' ~~ " 7 ' ~ } , ~ ~ ~[7~.,~ ' ~ , c[ (~,..z-..Ji~ng'~u

a.,_,__.._.~n Sichu

'

---

~a~L. oe /

l ~'~--,-~

N.,,{.~

:ao.oo j

l:tT';,oo,.-.-,

177

6

/ ~ [ Shanghai

1111

una

Jia

++J

----

"

o,oc

/ /

x,) Vietnam~

,o oo

i

~ ~.,,~_.~

-'/g'~

~- Leg

^



Tmonth -> 2~

If T m o n t h < - 5 ~ then SnowFrac = 1; if T m o n t h > 2~ SnowFrac = 0. In southern China, the average density of fresh snow is about 130 kg m -3 (Hd 1988). Thus, the monthly snowfall (Sf) was calculated as: Sf = SnowFrac

1.8

179

x

Pmnnth X 1000/130

O f course, this method cannot be used to assess snowfall at sites where snowfall occurs at mean monthly temperatures over 2~

Analysis of the relationship between species diversity and climatic factors

1.0

0--

0.8

~ 0.04

I 0.08

,

I 0.12

i

I 0,16

L

I 0,2

Plot area (ha)

Fig. 2 Conversion factors for the number of species (Q) and diversity index H (O) in relation to the area of combined plots of Jiulongshan.

For the analysis, mean annual temperature (Tm~an) and mean annual precipitation were used as the two climatic parameters. We used Tm~an as the thermal index because in the Chinese beech range it is correlated very strongly with other thermal indices such as yearly potential evapotranspiration and

180

K-E Cao and R. Peters

Kira's Warmth Index (Cao et al. 1995). For the sites in Huangshan, Daba and Laoshan, Tmean (Table 1) was extrapolated from the climatic data at the montane weather stations, which are only a few kilometers away from the forest plots sample& For the two plots in Tianpingshan, it was extrapolated from the data of Bamianshan montane weather station, which is about 120 km away in the south-west of Tianpingshan (Fig. 1). For the Miao'ershan site, it was extrapolated from unpublished data of Miao'ershan Nature Reserve, which were recorded in the 1980s at an altitude of 1200 m a.s.1, on the same slope as our sample site. For the remaining three sites, it was extrapolated from the data given in the literature for the same site (Wang et al. 1965; Chen & Tang 1982; Huang 1982). For Tmeanextrapolation, a lapse rate of 0.55~ per 100 m altitude was used (Jiang 1991; Xu 1991). Using the data from the same sources and the references given in Cao (1995) for the local climates, annual precipitation was extrapolated to these sites (Table 1), the precipitation lapse rates varying among local areas. The parameters used for climatic disturbances were the mean annual numbers of days on which wind storms, freezing rains, freezing fogs, snowfall, did rainfall over 50 ram, thunderstorms and hailstorms occurred. These disturbance regimes do not necessarily always cause disturbances such as canopy openings in the forests. But, we assumed that the probability of a disturbance regime causing disturbances in a forest correlates with the frequency of that regime occurring in the forest. Our statistical analysis did not consider the intensity of these disturbance regimes because the data were insufficient. Using the data from the 11 weather stations, the gradients of mean annual frequency of each disturbance regime were interpolated and mapped for the beech zone by means of SURFER software (the Kriging method; normal searching on the nearest four sites with known data; Golden Software 1989). Although there is a fairly wide variation in altitude among the stations, we believed that error in the analysis caused by this altitudinal variation was minimized by the interpolation method. Using these gradient maps, the mean number of disturbance days was interpolated for the eight study sites (Appendix 1). Using the lapse rates given by Jiang (1991), these means were adjusted for the two sites (Tianpingshan and Fanjingshan) whose plots were sampled at two single altitudinal ranges.

Bivariate scatterplots suggested a linear association between several climatic variables. Also, a linear association but with obvious outliers was observed between diversity indices and climatic variables such as temperature, and the yearly number of days with thunderstorms, snowfall, freezing rains and freezing fogs. These outliers were due to variation in temperature when plotting the diversity indices against disturbance variables, and due to variation in disturbance frequency when plotting the diversity against temperature. We applied Factor analysis to extract common factors from those interrelated variables (by the PCA method and Oblimin rotation, SPSS Inc. 1990). Two common factors were extracted from the six disturbance variables (Table 2). Factor 1 was mainly influenced by the mean annual numbers of days with freezing rains, freezing fogs and snowfall (positively), and with thunderstorms (negatively). Factor 2 was mainly influenced by the mean number of days with wind storms, and with did rainfall over 50 mm. Bartlett's test of sphericity (P = 0.0019) and the Kaiser-Meyer-Olkin measure (KMO measure= 0.5191) of sampling adequacy indicated that the factor extraction was appropriate (SPSS Inc. 1990). We applied partial correlation (Sokal & Rohlf 1995) to analyze the association between the diversity indices (H, S, E, IDE) and each individual disturbance variable or each common disturbance factor holding Tmean constant, and between each diversity variable and Tmeanholding the two common factors constant. Also, Pearson correlation coefficients between the diversity variables and the temperature were calculated. Table 2 The result of principal component analysis (Oblimin rotation): the loading of the two common factors extracted from the mean annual number of days with wind storm, diel rainfall over 50 mm (heavyrain), thunderstorm, freezing rain, freezingfog and snow, the eigenvalueof the two factors, and the percentage of total variance explained by the two factors.

Wind storm Heavy rain Thunderstorm Snow Freezing rain Freezing fog Eigenvalue Variance percentage

Factor 1

Factor 2

- 0.0679 0.0073 - 0.9316 0.9184 0.6052 0.7473

0.7312 0.9546 0.3382 - 0.0038 0.4564 0.5219

3.1686 52.8

1.6190 27.0

Disturbances and diversity in Chinese beech forests

i81

The nomenclature follows the Beijing Institute of Botany (1972-1976) and Chang and Huang

(1988). r-qo"x

ox

~

'N.

~ O

RESULTS 8

~=.~

~ o

~.~

-=

~ o m M ~ d o ~ o

0 e~

al o~

ca

.~.% o~

g~

@a= Z~ ca

.,=

~ "

# r

.~,,~ ~,-~

~

+-

o .~

.H

gZ ~

's

ILl

~

"~

;:~

0

=P-~

ca

ca ,-~

~oz

8

. ~ .~

e~ 9

d

.go=

~.~,-~ ~ ~ ~ "x~'~

ee~

,= ..~

..~

e~.--~ ~

~

ca

~.u~ ~ 0 "-'~ 9 ~ 9 O-~:Jt.~ ~'~ ~'~ ~ ~

~

~

o

~o ~...~ ~ ~"~ '-' ~ .

~ 0

There was a clear east-west trend in wind storm frequency in the Chinese beech range (Fig. 3a; Table 3). The eastern area near the coast had wind storms much more frequently than the western interior. D i d rains of over 50 mm were more frequent in the eastern and central areas than in the rest of the beech range (Fig. 3b). Also, the maximum amount of did rainfall was much greater in the central and eastern areas than in the western interior (Table 3). In contrast to wind storm, there was not an east-west trend in thunderstorm frequency but rather a north-south trend (Fig. 3c). The thunderstorm occurrences increased southwards. Freezing rains and freezing fogs were most frequent and intense (thicker ice accumulation) in the center of the beech range, which is near Lake Dongtian that covers an area of about 2 8 0 0 k m 2 (Fig. 3d,e; Table 3). The greatest thickness of 1.2 m of ice accumulated on an iron line during a continuous glaze storm was recorded at Nanyue Station in the central part. Freezing rains and freezing fogs in the east were slightly less frequent than in the center but more frequent than in the western interior. The estimated amount of annual snowfall varied from 0 to 70 cm among the 11 weather stations (Table 3). It was larger in the north-eastern and central areas than elsewhere in the beech range. A mean annual snowfall of 106 cm was estimated at the weather station of Huangshan. The snowfall frequency decreased southwards (Fig. 3f). Freezing rain, freezing fog, and snowfall were absent in the southernmost part (e.g. Wenshan, Table3). Hailstorm occurrences peaked in the central part (Fig. 3g). The disturbance gradients shown on the maps were less reliable in the southernmost part due to a lack of weather stations there. But this hardly influences our interpolation to the forest sites in this study because they are located between or fairly close to weather station sites (Fig. 1). The diversity indices (H, S, E and IDE) of the 10 sample units are shown in Table 4. Warmer sites usually had larger species diversity. The southernmost site had the highest species diversity (S and H)

182

K-E Cao and R. Peters

and the smallest importance of deciduous broadleaved trees fIDE). The northernmost site had the lowest species diversity (S and H) and the largest IDE. However, with a similar annual temperature (Tmean), the beech forests in the east (Huangshan and Jiulongshan) had greater diversity than those in the western inland (e.g. Daba, F1 and T1; Tables 1 and 4; Fig. 4). With a similar Tmean, the forest stands in the south (higher altitudes) had a greater abundance of evergreen broad-leaved trees than those in the north (T1 and F1 vs Daba and Jiulongshan vs Huangshan, Table 4). Tmean was correlated positively and strongly with the diversity index (H), number of species (S), and Equitability index (E) (Pearson correlation, Table 5). It was negatively and non-significantly correlated with IDE. Holding the two extracted disturbance factors constant (partial correlation), the temperature was still strongly correlated with H (Table 5). But its correlation with S and E was weakened and became non-significant, and its correlation with IDE totally disappeared. This indicates that the two extracted disturbance factors do influence these diversity parameters. Holding Tme~n constant, Factor 1 (see factor loading in Table 2) was strongly correlated with S negatively and with IDE positively, and Factor 2 was positively and modestly correlated with H (Table 5). Using the unextracted single disturbance factors and holding Wmean constant, the yearly number of wind storm days was strongly and positively correlated with H and E. The yearly number of snow days was strongly and positively correlated with IDE, and the yearly number of thunderstorm days strongly with S positively and with IDE negatively. None of the diversity indices was significantly correlated with annual precipitation. Other partial correlations between the climatic variables and diversity indices at a statistically non-significant level are shown also in Table 5.

DISCUSSION

ate northern China due to the pronounced continental climate there (Cao et al. 1995). The subtropical/warm temperate green zone, lying in the east next to the Tibeto-Himalayan Highland and extending into Taiwan, southernmost Korea, and southern Japan, distinguishes itself from elsewhere in the world with the same latitudes where there are mainly deserts and arid lands (Kira I991). The existence of this green zone is largely due to the TibetoHimalayan Highland that causes the summer and winter monsoons in the western Pacific to be greatly intensified (Kira 1991). The summer maritime monsoons bring abundant rainfalls and the winter continental monsoons carry coldness to the Chinese subtropical zone. Consequently, here the climate is humid with a hot summer but a relatively cold winter (Wu 1980; Kira 1991). The types of climatic disturbances observed in Chinese beech forests are common in other beech forests in the northern hemisphere (Arakawa 1969; Bryson & Hare 1974; Runkle 1990; Bryson 1991; Peters 1997). However, most Chinese beech forests receive glaze storms more frequently and intensely than other beech forests, some of which are two or four times more frequent than in north-eastern USA (Peters 1997). Occasionally, glaze storms in certain Chinese beech forests are very intense or co-occur with snow storms so that they are very

140 (a)

102

104

106

108

110

~

112

~--~, I 114

116

.

118

120

Longitude

Fagus occurs mainly in humid cool temperate regions in the northern hemisphere, except China (Peters 1997). In China, Fagus is confined to the mountains of the evergreen broad-leaved forest zone in subtropical (sensu Wu 1980 and Hou 1983) or warm temperate China (sensu Kira 1991; Cao et aZ 1995). It is absent from temperate or cool temper-

Fig. 3. Spatial gradients of the averageyearly occurrence (days year-1) of the followingclimaticdisturbaqce factors in montane Chinese beech zones: (a) wind storms (momentary velocity > 17.2 m s-l); (b) diel rainfall over 50 mm; (c) thunderstorms; (d) freezing rains; (e) freezing fogs; (f) snow; (g) hailstorms. They were extrapolated from the data of 11 weather stations located between altitudes of 1165 and 2065 m a.s.l. (Table 3).

Disturbances and diversity in Chinese beech forests

183

(c)

(b) a4-

3

2

~

_1~28-

"

\

24- / 102

1;4

1;6

108

110

112

114

116

118

120

Longitude

Longitude (e)

(d)

3432-

32 .~ ao~ 28-

_~ 282fr~ 24102

24i

104

i

106

108

110

i

112

114

116

/

118

120

102

104

106

i

108

~

110

i

i

i

i

112

114

116

118

120

112

114

146

118

120

Longitude

Longitude

(g)

(f) 34-

32-

32-

i

.~ 30"

0

i 28-

1.1 lo=

Fig. 3.

lo4

16~

16=

11o d= Longitude

d4

d5

A=

1=o

102

106

1;8

110 Longitude

(Continued).

destructive to the forests (cf. Wang 1984; Jiang 1991). Freezing rains are more destructive than freezing fogs due to their longer duration and denser ice;(Jiang 1991). The amount of snowfall in Chinese beech forests is much less than in some Japanese beech forests, where annual snowfall is 2-4 m (Japan Society of Forest Environment 1972).

Frequent heavy rains can also have a strong impact on Chinese beech forests due to the fact that they predominantly occur on slopes. Indeed, our analysis indicated the impact of heavy rain on diversity. Factor 2 was largely influenced by heavy rain (Table 2) and it was correlated with diversity (H) (Table 5). The strong correlations of Tmean with diversity

184

K-E C a o and R. Peters

Table 4 Number of woody species (S), total importance value of all deciduous trees and shrubs fIDE), Shannon-Wiener diversity index (H), and Equitability index (E) in the 10 sample units (sampled woody plants taller than 5 m) Sites

Latitude (o N)

S

32.7 30.1 29.7 29.7 28.3 28.2 27.9 27.9 25.8 24.3

16 20 18 44 56 35 19 24 28 62

Daba Huangshan Tianpingshan (T1) Tianpingshan (T2) Jiulongshan Kuankuoshui Fanjingshan (F1) Fanjingshan (F2) Miao'ershan Laoshan

S' 23 38 36 24 79

IDE

Ha

98 97 87 59 31 45 39 53 27 27

2.46 3.91 1.94 4.68 4.48 3.74 2.88 3.46 4.22 5.07

H'" 3.94 4.46 3.74 3.49 5.26

H i'

E

E'

2.74 3.93 2.84 4.55 4.36 4.39 2.96 3.95 4.43 5.28

0.62 0.90 0.47 0.86 0.77 0.73 0.68 0.75 0.88 0.85

0.87 0.85 0,73 0.76 0.83

aCalculated on the basis of combination of individuals and basal area at 1.3 m height, bCalculated solely on the basis of individuals. Data are the indices directly calculated from the sample units and those (S', H' and E') corrected for 0.20 ha. Table 5

Pearson correlation coefficients between diversity indices and mean annual temperature (T .... a); and partial correlation coefficients between diversity indices and the temperature (T ..... ), holding the two extracted disturbance factors constant, and between diversity indices and disturbance variables, holding T .... constant. H T .... " Tmean Factor 1 Factor 2 Storm Hea W rain Thunderstorm Snow Freezing rain Freezing fog Hailstorm PRCP

S

0.82

0.77

0.78

0.47 0.74 0.27 0.32 0.28 0.72 0.35 0.55 0.40 0:51 0.20

- 0.50 0.66 0.72 0.63 0.41 - 0.41 - 0.21 - 0.01 - 0.59 0.19

-

-

E

-

-

0.63 0.61 0.10 0.53 0.71 0.44 0.04 0.24 0.03 0.20 0.45 0.40

IDE - 0.53 0.06 0.73 - 0.28 0.14 - 0.34

100

lO

g O C O

_E0.1

0.01

- 0.79 0.86

- 0.09 0.29 0.i9 - 0.45

Entries bolded P < 0.05, otherwise P > 0.05. See the text or Table 4 for the legends of H, S, E and IDE; PRCP, mean annual precipitation (ram); heaW rain, diel rainfall over 50 ram. See Table 2 for PCA loading of Factors 1 and 2; other disturbance variables include the mean annual number of days with the occurrence of the weather events. indices (Table 5) support the energy-diversity theory (Currie & Paquin 1987; Currie 1991). T h e more energy (heat) available, the m o r e species are able to coexist, hence species richness and species equitability increase. Likewise, Hiura (1995) shows that the cumulative temperature o f the growth season is strongly and primarily correlated with the species diversity o f Japanese beech forests. A m o n g our sites, the warmest southernmost site (Laoshan) had the greatest diversity (H, S; Table 4), which is even comparable to that o f some tropical rainforests. For example, in Los Tuxtlas, Mexico, 0.25 ha plots

~ 0

I 10

~

I 20

Species

i

I 30

J

r

~

40

rank by importance

I

J

50

t 60

~ 70

value

Dominance-diversity curves of the three representative beech forests sampled. The species is ranked from the value of largest to smallest importance. In Laoshan (24~ 106~ 1600 m a.s.l.), the southernmost site, the forest was the most species-rich. With a similar mean annual temperature, the forest in Jiulongshan (28~ 118~ 1560 m a.s.l.) near the coast had a greater diversity than the forest in Daba (32~ 106~ 1450 m a.s.l.) in the north-western interior. The sample plot is 0.12 ha in Laoshan ( - - a t - - ) , and 0,20 ha in both Jiulongshan ( - - O - - ) and Daba ( - - I I - - ) . See Table 1 for the annual temperatures and precipitation in these sites. Fig. 4.

have between 61 and 78 w o o d y species (d.b.h. > 3.3 cm) and a diversity index between 3.7 and 4.8 (Bongers eta/. 1988). In another example in Puerto Rico, a 0 . 7 2 h a plot has 51 w o o d y species (d.b.h. > 4 cm) and a diversity index o f 4.0 (Crow 1980). Southern Chinese beech forests generally have greater species richness than beech forests in N o r t h America and Europe (Cao 1995; Peters 1997). This is partly attributable to the old vegeta-

Disturbances and diversity in Chinese beech forests tion history of southern China, on which the Pleistocene glaciers barely made an impact (Wu 1980; Hsii 1983). Our data also support the disturbance-diversity theory (Connell 1978). The increase of species diversity towards the east in our forests was in accordance with the east-west trend of wind storm frequency (Table 4; Figs 3a and 4). Holding Tmcan constant, the wind storm frequency was correlated strongly and positively with diversity indices H and E (Table 5). Also, Factor 2, which was mainly influenced by wind storm and heavy rain frequency, was positively correlated with H. Indeed, when sites had a comparable Tmean, the eastern sites had greater species diversity than the western sites (Tables 1 and 4; Fig. 4). The low diversity in the beech forest of Daba in the western interior coincided with a low degree of impacts by climatic disturbances (Table 4; Figs 3 and 4; Appendix 1). The hypothesis of Connell (1978) suggests a non-linear relationship between disturbances and diversity. But, our data suggested a linear relationship between disturbances and diversity (Table5). Hiura (1995) has also shown a linear relationship between disturbance and the species diversity in Japanese beech forests. This is likely because Chinese and Japanese beech forests do not cover a range with disturbances varying from very little to very frequently and intensely. Our wind storm data suggest that the eastern sites were more disturbed than the western ones (Fig. 3a), but the disturbances in the east were not so frequent and intense that the species diversity of the forests was reduced. Also, in contrast to the western interior, lightdemanding trees are important canopy components in our eastern sites, such as Liriodendron chinense, Liquidambar acalycina, Magnolia cylindra and Sassafras tzumu in Jiulongshan, and Magnolia cylindra, Prunus serrulata and Pinus taiwanensis, Quercus stewardii in Huangshan (not shown but cf. Chen & Tang 1982 and Zhou 1965). This is again consistent with the disturbance-diversity theory (Connell 1978). We did not include the importance of lightdemanding trees in the analysis because our sample plots were relatively small and focused on the stands dominated by beech, which probably underestimate the importance of light-demanding trees and the diversity indices in sites receiving frequent disturbances. The strong negative correlation between S and

185

Factor 1 (Table 5) suggests a negative effect on species diversity by the disturbances of freezing rains, freezing fogs and snowfall. These disturbance regimes have a physical effect upon trees that is rather different from that of wind storms and heavy rains. In the Chinese beech zone, glaze and snow storms occur in winter and early spring when deciduous trees are leafless, whereas wind storms occur throughout the year, and heavy rains concentrate in the growing season. Therefore, evergreen broadleaved trees carry heavier loads during ice and snow storms and have a higher risk of breakage than slowgrowing deciduous hardwoods (cf. Wang 1984; Hd 1988). Also, compared to slow-growing hardwood trees, fast-growing light-demanding trees have less dense wood and are physically weaker and therefore more susceptible to ice breakage. An example is the beech forest of Miao'ershan. There we observed that exposed evergreen broad-leaved trees and fastgrowing deciduous trees such as Liquidambar acalycina and Sassafras tzumu, both in the canopy and in canopy gaps, often had broken crowns. The broken branches were probably caused by ice and snow storms. They open access to wood-rot fungi and cause branches and trunks to rot (Shigo 1986). In Miao'ershan, about 60-70% of Castanopsis lamontii trees in the overstory of the mixed beech forest had rotten trunks (Cao 1995). These trees with rotten trunks are certainly susceptible to wind breakage. In contrast, E lucida canopy trees usually have extensive unbroken crowns. The relative resistance of the beech to ice or snow breakage probably rests on its leaflessness in winter and its dense wood. The beech forest on Mt Tianpingshan receives relatively frequent and intense ice storms (Appendix 1; also cf. Bamianshan station in Table 3), and E lucida crowns form an extensive monospecific forest canopy in the montane zone (Qi 1990). Other studies also show that the ice and snow storms favor a limited number of trees in the dominance of forest overstorys (Downs 1938; Wang 1984; Hd 1988). The minimum temperature was not included in our analysis due to a lack of data. But it could be important to influence the importance of the evergreen trees in Chinese beech forests. It is considered as a major factor that controls distribution of evergreen broad-leaved trees (Sakai & Larcher 1987). Some northern and southern Chinese beech forests (higher altitude) experience a similar T but evergreen broad-leaved trees are more important in . . . . ,

186

K-E Cao and R. Peters

the southern forests (e.g. Daba vs T1 and Fanjingshan 1, and Huangshan vs Jiulongshan, Tables 1 and 4). This is not only explained by the negative effect of cold disturbances (glaze and snow) on the evergreens but probably by the minimum temperature (Sakai & Larcher 1987). In Japan, most evergreen broad-leaved trees are not able to tolerate cold extremes below-15~ (Sakai 1975). In the Chinese beech range, as elsewhere in the northern hemisphere, the further north the greater the probability of the occurrence of a low temperature harmful to evergreen broad-leaved trees. In the northernmost study site (Daba), a minimum temperature of -22.5~ occurred within a 3-year period of climatic records (Daba Forest Station, unpubl, data). This is the lowest extreme temperature recorded among all weather stations in the Chinese beech range (Cao et al. 1995). The beech forest in Daba was almost completely composed of deciduous broad-leaved trees (Table 4), except for some small evergreen trees and shrubs in the understory. Also, the positive relationship between thunderstorms and the importance of the evergreen trees (Table 5) is probably not directly due to the impact of thunderstorms but rather the minimum temperature. The thunderstorm frequency increased towards the equator (Fig. 3c), which coincides with the diminishing occurrence of a minimum temperature (see above). Although the mean annual precipitation varies from 1400 to 2550 m m among the study forest sites (Table 1), the analysis did not show any relationship between precipitation and diversity parameters (Table5). In neotropical forests, Gentry (1982) found a positive correlation between species richness per 1000 m 2 plot and rainfall for trees with a diameter above 2.5 cm but no correlation for trees above 10 cm diameter. In West African rainforests, van Rompaey (1993) showed a decrease in the species richness of large trees (d.b.h. > 70 cm) with increasing rainfall. The lack of association between species diversity of Chinese beech forests and amount of precipitation is probably attributable to the fact that all these forests occur in humid climates with sufficient water supplied by precipitation throughout the year (Cao et al. 1995). Several single disturbance variables showed modest correlation to the diversity parameters at a statistically non-significant level. However, this statistical non-significance could be largely due to our relatively small sample size and

does not necessarily imply a lack of impact on diversity by these disturbances (cf. Fowler & Cohen 1990). When they were loaded together into a common factor the correlation improved (Table 5). The data presented here support both the energy-diversity and disturbance-diversity theories. In beech forests in our study, species diversity correlated with warmth and climatic disturbances. Our analysis suggests that species diversity a n d species composition are both influenced by disturbances, wind storms and heavy rains enhancing species diversity, whereas cold disturbances diminish the diversity and dominance of evergreen broad-leaved trees in Chinese beech forests.

ACKNOWLED GEMENTS

We are grateful to F. Bongers, J. Den Ouden, R. A. A. Oldeman, P. van der Meer and P. Schmidt for advice and comments on the earlier manuscript, to G. Gort and J. J. Jansen for advice on statistics, to X. P. Wang for advice and the unpublished plot data of Laoshan. Thanks are due to H&ping Chen, Jin-sheng He, Xiong Lin, Chuan-dong Yang, 'Lao' Yang, 'Xaio' Yang and Yuan-guang Wen for assistance in the field. We appreciate the permission given to conduct field works by the Daba Forest Station, Huangshan National Park, Fanjingshan MAB Nature Reserve, Miao'ershan Nature Reserve, and permission by the Meteorological Information Center Beijing to use the climatic data. This study was co-funded by the National Natural Science Foundation of China through a research project (no. 39070188), Wageningen Agricultural University through a PhD fellowship to KFC, and the Chinese Academy of Sciences and the Royal Academy of Sciences of the Netherlands through a co-operative project.

REFERENCES

ABER J. D. & FEDERER C. A. (1992) A generalized, lumped-parameter model of photosynthesis, evapotranspiration and net primary production in temperate and boreal forest ecosystems. Oecologia 92: 463-474. ANONYMOUS (ed.) (1986) Collected Papers on Investigation of Agricultural Resources and Agricultural Zonation

Disturbances and diversity in Chinese beech forests

of Beiling Prefecture, Sichuan Province. Special Publication of Belling Agricultural Zonation Office, Beiling, Sichuan, China (in Chinese). ARAKAWAH. (ed.) (1969) Climates of Northern and Eastern Asia. Elsevier, Amsterdam. ASHTON P.S. (1993) The community ecology of Asian rain forests, in relation to catastrophic events. Journal of Biosciences 18: 501-514. BEIJING INSTITUTE OF BO;FANY, CHINESE ACADEMY OF SCIENCES (1972--1976)Iconographia Cormo-

phytorum Sinicorum. Tomus I, IL III, IV & V. Science Press, Beijing (in Chinese). BONGERSF., POPMAJ., MEAVEDEE CASTILLOJ. & CARABIASJ. (1988)Structure and floristical composition of the lowland rain forest of Los Tuxtlas, Mexico. Vegetario 74: 55-80. BORMANN F. H. & LIKENS G. E. (1979) Pattern and Process in a Forested Ecosystem. Springer, Berlin. Box E. O. (1981) Macroclimate and Plant Forms: An

Introduction to Predictive Modelling in Phytogeography. Junk, The Hague. BRONIG E. F. & HUANG Y. W. (1989) Patterns of tree species diversity and canopy structure and dynamics in humid tropical evergreen forests on Borneo and in China. In: Tropical Forests, Botanical Dynamics, Speciation and Diversity (eds L. B. Holm-Nielsen, I. C. Nielsen & H. Balslev) pp. 75-88. Academic Press, London. BRYANTE. (1991). Natural Hazards. Cambridge University Press, Cambridge. BRYSON R. A. & HARE F. K (eds) (1974) Climates of North America. Elsevier, Amsterdam. CAO K. F. (1995) Fagus Dominance in Chinese Montane Forests: Natural Regeneration ofFagus lucida and Fagus hayame vat. pashanica. PhD thesis. Wageningen Agricultural University, The Netherlands. CAO K. F., PETERSR. & OLDEMANR. A. A. (1995) Climatic ranges and distribution of Chinese Fagus species. Journal of Vegetation Science 6:317-324. CHANGY. T. & HUANG C. C. (1988) Notes on Fagaceae (2). Acta Phytotaxonomica Sinica 26:111-119 (in Chinese with English abstract). CHEN Q. P. & TANG Z. C. (1982) Mixed evergreen and deciduous broad-leaved forest on Mt. Jiulongshan. In:

Reports of Integrated Scientific Survey of Natural Resources on ME. Jiulongshan, Zhejiang Province (ed. Anonymous) pp. 150-168. Special Publication of Zhejiang Science and Technology Society, Hangzhou, China (in Chinese). CONNELE J. H. (1978) Diversity in tropical rain forests and coral reefs. Science 199: 1302-1310. CRAWLEYM. J. (1986) The structure of plant communities. In: Plant Ecology (ed. M. J. Crawley) pp. 1-50. Blackwell Science, London.

187

CROW T. R. (1980) A rain forest chronicle: a 30 year record of change in structure and composition at E1 Verde, Puerto Rico. Biotropica 12: 42-55. CURRIED. J. (1991) Energy and large-scale patterns of animal- and plant-species richness. The American Naturalist 137: 27-49. CURRIED. J. & PAQUIN V. (1987) Large-scale biogeographical patterns of species richness of trees. Nature 329: 326-327. DOWNS A. A. (1938) Glaze damage in the birchbeech-maple-hemlock type of Pennsylvania and New York. Journal of Forestry 36: 63-70. FAO-UNESCO (1988) FAO/UNESCO Soil Map of the World, Revised Legend. World Resources Report 60, FAO, Rome. FENG P. Z., D C. C. & LI X. C. (1985) An Analysis of

Major Meteorological Hazards in China (1951-1980). Meteorological Press China, Beijing (in Chinese). FOWLER J. & COHEN L. (1990) Practical Statistics for Field Biology. John Wiley and Sons, New York. GENTRY A. H. (1982) Patterns of neotropical plant species diversity. Evolutionary Biology 15: 1-84. HE Q. T. (1988) Meteorology. Forestry Press China, Beijing (in Chinese). HIURA T. (1995) Gap formation and species diversity in Japanese beech forests: a test of the intermediate disturbance hypothesis on a geographic scale. Oecologia 104: 265-271. Hou H. Y. (1983) Vegetation of China with reference to its geographical distribution. Annals of the Missouri Botanical Garden 70: 509-548. HsI~ J. (1983) Late Cretaceous and Cenozoic vegetations in China, emphasizing their connections with North America. Annals of the Missouri Botanical Garden 70: 490-508. HUANG W. L (ed) (1982) Scientific Survey of The Fanjingshan Nature Reserve. Special Publication of Guizhou Provincial Environmental Bureau and Environmental Society, Guiyang, China (in Chinese with English summary).

JAPAN SOCIETYOF FORESTENVIRONMENT(1972) Forest

Environment Map ofJapan (1:2 000 000), and Explanatory Book. (Trans. H. Yagi). Japan Society of Forest Environment, Tokyo. JIANG D. L (ed) (1991) Climate of Middle and Lower Reaches of Yangtze River. Meteorological Press China, Beijing (in Chinese). KtRA T. (1991) Forest ecosystems of East and Southeast Asia in a global perspective. Ecological Research 6: 185-200. LATHAMR. E. & RICKLEFSR. E. (1993) Global patterns of tree species richness in moist forests: energy-diversity theory does not account for variation in species richness. Oikos 67: 325-333.

188

K-E Cao and R. Peters

LI C. K. & SUN O (eds) (1990) Soils of China. Science Press, Beijing. LIU G. R. & Xu K. X. (1991) A study of landslides in mountainous areas of western Hubei Province, China. In: Proceedings of Landslides in China, Vol. 8 (ed. Anonymous) pp. 78-91. Railway Press China, Beijing (in Chinese). METEOROLOGICALINFORMATIONCENTREBEIJING(1984) Ground Climatic Data of China (1951-1980). Meteorological Press China, Beijing (in Chinese). NAKaSHtZUKAT. (1989) Role of uprooting in composition and dynamics of an old-growth forest in Japan. Ecology 70: 1273-1278. NIEUWOLT S. (1977) Tropical Climatology, An Introduction to The Climates of The Low Latitudes. John Wiley and Sons, Chichester. OLDEMANR. A. A. (1983) Tropical rain forest, architecture, silvigenesis and diversity. In: Tropical Rain Forest: Ecology and Management (eds S. L. Sutton, T. C. Whitmore & A. C. Chadwick) pp. 139-150. Blackwell, Oxford. OLDEMAN R. A. A. (1990) Forests: Elements of Silvology. Springer, Berlin. PETERS R. (1997) Beech Forests.Geobotany 24. Kluwer, Dordrecht. PICKETT S. T. P. & WHITE P. S (eds, 1985) TheEcology of Natural Disturbance and Patch Dynamics. Academic Press, Orlando. P~ELOUE. C. (1975) Ecological Diversity. John Wiley and Sons, New York. PLATT W. J. & STRONG D. R (eds) (1989) Treefall gaps and forest dynamics. Special Feature, Ecology 70: 535-576. PUTZ F. E. (1983) Treefall pits and mounds, buried seeds, and the importance of soil disturbance to pioneer trees on Barro Colorado Island, Panama. Ecology 64: 1069-1074. QI C. J (ed) (1990) Vegetation oft-Iunan Province, China. Hunan Science and Technology Press, Changsa, China (in Chinese). ROHDE K. (1992) Latitudinal gradients in species diversity: the search for the primary cause. Oikos 65: 514-527. RUNKLE J. R. (1990) Gap dynamics in an Ohio AcerFagus forest and speculations on the geography of disturbance. Canadian Journal of Forest Research 20: 632-641, SAKAI A. (1975) Freezing resistance of evergreen and deciduous broad-leaf trees in Japan with special reference to their distributions. Japanese Journal of Ecology 25:101-111 (in Japanese with English summary).

SAKAIA. & LARCHERW. (1987) Frost Survival of Plants, Responses and Adaptation to Freezing Stress. SpringerVerlag, Berlin. SHIGO A. L. (1986) A New Tree Biology: Facts, Photos, and Philosophies on Tress and Their Problems and Proper Care. Shigo and Trees, Durham. SOKALR. & ROHLFF. J. (1995) Biometry. 3rd ed. W. H. Freeman and Co., New York. SPSS INC. (1990) SPSS/PC+ Manual, Advanced Statistics 4. Ofor the IBM PC/XT/A T and PSI2. SPSS Inc., Chicago. SUN G. Z. (ed) (1988) Representatives of Landslides in China. Science Press China, Beijing (in Chinese). VAN ROMPAEYR. S. A. R. (1993) Forest Gradients in West Africa: A Spatial Gradient Analysis. PhD thesis. Wageningen Agricultural University, The Netherlands. WALTERH. (1985) Vegetation of the Earth (and Ecological Systems of the Geo-biosphere). 3rd ed. Springer, Berlin. WANG J. X. (1984) An unusual glaze storm and its damage to forests in the mountainous area of southern Hunan. In: Proceedings of Meteorological Studies in Forestry (eds A. L. Jiang & Z. M. Song) pp. 68-71. Meteorological Press China, Beijing (in Chinese). WANG X. P., WANGJ. J., CHEN W. L. et al. (1965) Vegetation of Kuankuoshui forest area, Suiyang, Guizhou. Acta Phytoecologia et Geobotanica Sinica 3: 264-286. WANG Y. S. (1990) Hailstorm hazard in China. In: Natural Hazard; in China (eds G. Z. Sun, Y. S. Wang & P. Y. Zhang) pp. 127-135. Academic Press China, Beijing (in Chinese). WHITTAKER R. H. (1975) Communities and Ecosystems. Macmillan, New York. WOLFEJ. A. (1979) Temperature Parameters of Humid to Mesic Forests of Eastern Asia and Relations to Forests of Other Regions of The Northern Hemisphere and Australia. Geological Survey Professional Paper no. 1106.

WOODWARDF. I. (1987) Climate and Plant Distribution. Cambridge University Press, London. Wu Z. Y. (ed) (1980) Vegetation of China. Science Press, Beijing (in Chinese). Xu Y. H. (ed) (1991) Climate of Southwest China. Meteorological Press China, Beijing (in Chinese). ZICANGY. C., HE W. S. & LI S. K. (1991) Meteorological Hazards in Agriculture of China. Meteorological Press China, Beijing (in Chinese). ZHOU J. L. (1965) Plant communities of Mr. Hwangshan (Huangshan). In: Observationes ad Florulam Hwangshanicam (ed. P.C. Chen) pp. 207-266. Shanghai Science and Technology Press, Shanghai (in Chinese).

D i s t u r b a n c e s a n d diversity in C h i n e s e b e e c h forests

APPENDIX 1 Estimated mean annual number of days on which the major climatic disturbances occurred in the study sites. Entries are mean annual number of days on which the climatic disturbances occur. A wind storm was a wind with a velocity exceeding 17.2 m s-1 at any moment. Heavy rain was a diel rainfall over 50 mm. The code for each site is same as in Table 1 and Fig. 1. Sites

Code

Daba Huangshan Tianpingshan Tianpingshan Jiulongshan Kuankuoshui Fanjingshan Fanjingshan Miao'ershan Laoshan

D HS T1 T2 J K F1 F2 M L

Wind storm 22 113 33 20 83 24 30 24 35 21

Heavy rain 2.5 5 5.4 5.4 7 3 4.2 4.2 5.5 2.2

Freezing rain Freezing fog Snowfall Thunderstorm 28 42 58 33 31 37 48 39 48 13

43 53 63 37 42 33 43 35 50 12

38 30 38 30 21 22 23 21 14.5 5.5

43 50 55 55 72 58 62 62 58 71

Hailstorm 1.8 1.5 3.3 3.3 0.9 2.3 2.8 2.8 1.4 1.3

189