Light interception and its relation to structural differences in three Australian rainforest canopies
M. D. LOWMAN continuous throughout the year. Within several Zoologj3 Departrnent, University of h'c~' Australian rainforests, there are distinct England, Arrnidale, NSW 2351, Australia seasonal peaks in leaf-fall (Webb et al. 1969), without concurrent replacement of canopy by Abstract leaf emergence (Brasell et al. 1980; Lowman 1982). In particular, the cool temperate rain- Relative diflerences in the seasonality of light forest sheds approximately half of its canopy penetrating the canopies of three Australian duringautumn, but this is not replaced until the i.ainforests were tt~easured directly n,ith a following spring. Both warm temperate and quanturn rneter and indirectbl ~rsing subtropical rainforests exhibit less significant hemispherical photographs. The tnossy (but nonetheless pronounced) leaf-falls in early rnicrophyll~fern.forest had higher incident light spring, with new leaf replacement occurring levels (100-200 pEinstcin r,12 per s) and tnore gradually throughout late spring and summer. sky visible (21.8%) at ground level than the Because these rainforests exhibit degrees of sitnple (5-20 pEinsteiti nli' per s with 11.2°0 seasonality in leaf-fall and leaf replacement, it slkylight) and cornpl~~~(5- 10 pEinstein m-2 per is postulated that the amount of sky visible s with 10.8% skylight) notophyll vine ,forests. from the forest floor up through the canopy will There brvre nosignificant changes in atnounts of exhibit temporal fluctuations relative to the shy visible through the canopy over tirne (10 proportion of leaf loss. In addition, the rnonths) .for anjl canopy, despite the ,fact that incidence of l~ght through the canopy is Icafl,fall ~vasa seasonal event. Profile diagratns predicted to vary according to canopy density ~vercconstructed at each site to illustrate the and structure. To examine these hypotheses, depth and stratification of the canopy, and the two types of light measurements were con- differences in these structural features were ducted in three rainforest formations. First, related to the variation in light attenuation. hemispherical photographs (Evans & Coombe Ecological aspects of'light interception through 1959; Anderson 1964; Horn 1971) were taken rainforest canopies are discussed, including tree monthly during 1 year to indicate the relative branch growth patterns and leafing phenologies, changes in the leafiness of the canopy by and seedling regeneration on the,forcst ,/loor. measuring the proportion of sky visible at ground level, and to indicate the canopy struc- Introduction ture by comparing the amounts of sun flecks with light gaps. Second, direct light readings Light penetration through a forest canopy were made at different levels within the three maintains understorey vegetation and deter- canopies to compare available light for leaves mines the degree of suppression or vigour of its under different canopy conditions. Profile growth. In deciduous forests, the annual loss of diagrams were constructed to illustrate the leaves from the canopy results in a large structural features of each canopy, and were ex- seasonal change in light influx to the forest floor amined in relation to light interception through (Horn 1971), and many herbaceous under- the canopies. Light was defined as any portion storey species leafout early in the spring, before of unobscured sky visible through the tree the new canopy in the overstorey flushes and canopy. This study utilized light to measure intercepts light (Jackson 1966). indirectly the density and deflective capacity of In evergreen rainforests, a complete seasonal three leafy canopies. The measurement of light loss of canopy foliage never occurs: leaf-fall is in its true physical sense was not attempted. Methods All woody plants greater than 1 m in height within a 70 X 8 m transect were included. Tree heights and canopy depth were calculated with Structural descriptions of the canopies the aid of a hypsometer, and drawn to scale. Profile diagrams of forest transects were Canopy depth was defined as the cumulative constructed according to standard methods vertical length intersecting elements of canopy described by Davis and Richards (1933) and taken from the top of the uppermost canopy Richards (1952). One site (containing at least down through the lowest-level foliage (Fig l), one of the photographic locations) was selected and mean canopy depth was calculated from 15 as representative of a mature, undisturbed replicate measurements taken at 5 m intervals canopy in each of the three rainforest along the profile diagrams. formations. The locations of the three rain- forests and their respective classifications according to Webb (1 959, 1968) are: Light measurements (1) Cool temperate or microphyll fern forest (MFF) at 1400 m elevation on a slight eastern Direct light readings Light readings were aspect in the New England National Park taken in different regions of the canopy using a (32" 30's); Lambda quantum meter. Access to the upper (2) Warm temperate or simple notophyll canopy was attained using ropes and technical vine forest (SNVF) at 60 m elevation on the climbing apparatus (Lowman 1982, 1984). In east-sloping bank of a tributary of the Hacking most cases, an emergent tree was climbed in River in the Royal National Park (34" 10's); order to reach across to the upper canopy of a and neighbouring tree. The same sites were (3) Subtropical or complex notophyll vine measured under clear, sunny conditions and forest (CNVF) at 800 m elevation in the repeated in cloudy conditions. At least five Dorrigo National Park (30" 20tS), on the measurements were taken in each region. eastern slopes of the New England plateau. Canopy regions were delineated as follows:
FIG. 1. Profile diagram of a cool temperate rainforest, New England National Park, NSW. Speciesabbreviations: Am Acacra ~i~clatio~~lon;Ar ..llbx~a rusc!lolla; Co Cordj~lineslrlda; Cq C'oprosma quadrihda; Cs Cryplocarjri sp.; Da Drcksonra anlarclrca: Ds I1or~phor.a.sas.sufia.s; Eh Elacocarpus holopc~alus;Nm Nolhofagus moorcl; Oe Orilcs c,~cclsa;TI Trochocarpa laurina; Ts ksfnania slip~lala. (1) Upper canopy: measured, and instead sky visibility was used as Above - over the uppermost canopy leaves; an indicator of canopy leafiness, corrections for exposed to direct light. zenith and time of year were not necessary Within - approximately 10 cm below the (Horn 197 1). The spherical photographs were uppermost canopy leaves; light diffused with projected onto a horizontal screen to a standard only one leaf layer overhead. size, and areas of light marked with black ink on (2) Lower canopy: acetate sheets. The light was separated into 'sun Sun fleck - leaf surface within 2 m ofground flecks', defined as any small spherical image of level that received a transient beam of direct light; and 'gaps', any larger or less symmetrical light. area of light. An area meter (Lambda Shade - leaf surface within 2 m of ground Instruments, Model 300) was used to calculate level that received no direct light. areas of total light. gaps, and sun flecks; and the Gap - area wider than 5 m on the forest floor amounts were expressed as percentages of the with direct light. area of the circular image. This method does not quantify the physical amount of light Measurement of light through the canopies reaching the forest floor (since it varies with Hemispherical photography was used in an cloud cover. and position of sun on a daily and unorthodox manner to measure relative seasonal basis), rather, it provides a relative changes in rainforest canopy structure as index of canopy cover and density as indicated indicated by amounts of unobscured sky visible by light infiltration. at ground level (Horn 1971). Ten initial light readings were made at random within each forest to determine mean light penetration Results levels, and three permanent photographic sites were chosen in each forest that best represented Structural characteristics of the canopies mean light levels. Photographs were taken every month for I year, at appoximately Figures 1-3 illustrate the profile diagrams 1000 h. Because light quantity was not being constructed in representative sites of each rain-
10 20 3 0 L 0 50 60 70 Distance (m)
FIG. 2. Profile diagram ofa warm lemperale rainforest. Royal National Park, NSW. Spec~esabbreviations: As 4cmcnu smilh~i; Ca C'rratopclal~imaprlalu~n; Da Diploglo~trsa~tslralrs; Ds Dorl~phorasassafras; Fc Ficus coronata; La Ln,istona auslralis; Lr L~lsrarr~icula~a; PC Polvosma cunningharnii; Pu Pi//osporu~nu~rd~tlatutn; Sa Sloanea australi,: Sc S,~n:gr~r~prc.oolmtnlanum: Sg Svncarpru ,~/o~nul~lcra:TI Trtslanra laurina: Wh I47lckru h~rc~gc~lrana.Vines: Piper novae-hollandiar; I'lss~tss/ercull FIG. 3. Prolile d~agramof a subtropical rainforest, Dorrigo National Park, NSW. Species abbreviations: Aa Alec/r,~on s~rhcino.r~u~;An Asplrniu~~rnidus; Be Bo~isloaruodirfi)rrrrrs; Ba Brach.vchi/on act~rifoliuttr:Cm Cryptocary mt7issnmr; Da Diploglottrs austra1i.s; De Dendrocnidc3excclsa; Ds Doryphora su.vsafias: Em Endiandra inuellr~ri,Fw Frcus watkrn~iana;Gb (;ca~ccoishcnfha~~rii; Ha Ir'eritirra actinophi~ffa;Lm Lirtospadi.r ~nonosrachi~u;Nd Nc~uf~tst~udmlhutu; Oe Ori/t~sc..rcr,lsa: Sf Si TABLE I. Light measurements (~Einsteinm-' per s) within three Australian rainforest canopies (1000 h) Canopy rcglon Subtropical Warm temperate Cool temperate Sunny day Cloudy day Sunn) day Cloudy day Sunn) day Cloudy day (s.e.m.) Upper canopy Above 2000 60-SO(2.1) 2000 60-80 2000 60-80 Within 100-200(9.2) 20-30(1.5) 100-200 20-30 500-700 30-40 Lower canopy Sun flcck 800-9000(10.2) 10-20(3.5) 600-800 20-30 I I50 60 Shade 10-SO(8.3) 5-1 O(0.3) 10-200 5-20 100-200 8-10 Gap - - - - 1900 60 available to a leaf surface decreased dra- monthly fluctuations in skylight appeared to matically from 2000 pEinstein m-' per s above coincide with leaf-fall patterns; however, these the upper canopy to as low as 5 pEinstein mp' changes were much less apparent than per s in the shaded regions of the lower canopy. hypothesized (Fig. 4). As predicted, the The magnitude of the difference in light levels greatest proportion (- 25%) of light visible on between the upper and lower canopy regions the MFF floor occurred after the autumnal leaf- was progressively greater as the canopy fall peak (March-May) and the great amount in complexity increased (i.e. from MFF to SNVF the CNVF (- 12%) occurred just after the to CNVF). spring leaf-fall peak (October-December). In the understorey, sun flecks represented as although in neither forest was the seasonal much as 50% of direct sunlight levels whereas variation in visible skylight statistically shade regions represented less than 1% of the significant. The SNVF did not show any change amount of direct sunlight. Light levels on in the amount of unobscured sky that cor- cloudy days were approximately one-tenth the related with either leaf-fall or leaf emergence level on sunny days. although this may vary patterns. Instead, the slight fluctuations may be depending on different types of cloud cover due to small scale disturbances (e.g. branch mortality). The branch architectural patterns Measurement of light through the canopies appeared to maintain stable light interception The monthly measurements of visible light in the canopy throughout the year, despite unobscured by the tree canopies were sig- seasonal leaf-fall events. nificantly different among sites (Table 2) but Table 2 summarizes the mean canopy depths not over time (F9,60= 0.46, NS). A sig- and the proportions of sky visible through the nificantly greater percentage of unobscured sky three canopies. The CNVF had a significantly was visible beneath the canopy of MFFthan the deeper canopy, and also the least amount of sky SNVF or CNVF canopies (Table 2). The visible beneath it. The SNVF canopy was TABLE 2. Summary of the comparisons of canopy light interception characteristics at three rainforest sites (mean valuc) Subtropical Warm Cool s.e.m. d.f. F Significance temperate temperate (0 Total amount of sky visible 10.77 = * 11.17 < 21.83 0.35 2,87 160.87 < 0.001 through canopy ('Yo of area 01' hemispherical photograph) - * L~ghtflecks (% of total unobscured 37.8 - 33.2 > 22.6 1.02 2.87 99.1 1 < 0.001 sky visiblc below the canopy) Gaps (Yo of total unobscured sky 62.2 < 66.8 < 77.4 1.01 2,87 70.68 < 0.001 visible bclow the canopy) Ratios of (light fleck :gap) 60.8 = * 49.7 > 29.2 1.03 2,87 59.95 < 0.001 Canopy dcpth (m) 20.03 > 10.6 = * 12.03 0.58 2.42 26.97 < 0.001 ~ - *The two means are not significantly different by the Studcnt-Newman-Keul's test. - Light fleck SONDJFMAMJ 1979 1980 Time (months) FIG. 4. ;Mean amount of unobscured sky visible through the canopies of three Australian rainforests as measured from hemispherical photographs; subtropical -; warm temperate o - - - 0;cool temperate m-.-m. Circles indicate the light fleck:gap proportions. Bars represent main periods of leaf emergence (solid) and leaf-fall (open). similar to the CNVFcanopy in light infiltration but the fact that no more than 12% of the sky except that its canopy was shallower. The MFF ever remained unobscured at ground level canopy had the greatest portion of sky visible implies that the leaf density is extremely high. through it, and a higher proportion of gaps to Other measurements of leaf area index sun flecks than the other canopies. affirmed this (Lowman 1982). In addition, the spring leaf flush is simultaneous with leaf senescence, so that the canopy is always Discussion leafy. The Australian evergreen rainforests examined The absence of seasonality in sky visible did not exhibit significant seasonal changes in through the MFF canopy is not as easily the amount of unobscured sky visible through explained by the structural features evident in their canopies, despite the fact that up to one- the profile diagrams. Its thinner, less stratified half of the MFF canopy may be lost during leaf- canopy would appear to allow greater visibility fall peaks and the fact that the CNVF includes of unobscured sky through it during the winter completely deciduous species (e.g. Brachy- months when half the canopy has fallen; but rhiton ucc~r~foliusand Toona azcvtralis) as scat- this was not the case. Subsequent examination tered canopy trees. The original hypothesis is of beech branch architecture may instead therefore rejected. An alternative explanation explain the constant proportions of sky visible for the year-round constancy of light observed through the canopy, despite the seasonal may relate to structural characteristics of the changes of leaf material present. As with other canopy, rather than seasonal growth or members of the family Fagaceae, the branches senescence patterns of leaves. The deep, multi- and leaves are oriented horizontally in antarctic layered canopy in the CNVF creates an obvious beech (Halle et al. 1978). When leaf-fall occurs, structural barrier to light influx into the lower light penetration to the lower beech canopy regions despite intermittent leaf-fall. The may increase, but the horizontal layers of beech SNVF canopy has a relatively shallow depth; leaves intercept light so that the amount of Lig111lhrouglr rainforest canopies 169 unobscured sky seen at ground level is canopy gap is created (i.e. a disturbance occurs). unchanged. Since light levels do not change Without seasonal changes in light infiltration seasonally under the beech canopy, only the through the canopy, there is no opportunity for occurrence of gaps by tree falls may allow the ephemeral growth as in northern temperate, continued monopoly of beech in the upper deciduous forests. The MFF canopy has a canopy by triggering the release of seedlings and higher frequency of gaps, the existence of which suckers already established in the under- have two possible explanations. First, due to storey. more rigorous climatic conditions, disturbance The use ofhemispherical photography in this may occur more frequently; and second, again study did not directly measure the light regime. due to the rigorous climate, a tree may grow Such photographs can only be used to directly more slowly, therefore taking longer to attain measure the light regime if sufficient infor- canopy height and to occupy the gaps than in mation is available on the degree of cloudiness the SNVF and CNVF. Only long-term observa- and variation in proportions ofdiffuse to direct tions can substantiate either hypothesis. sunlight and daily total solar radiation, as well as seasonal variation in the solar track across Acknowledgments the sky (Anderson 1964). The purpose of photography in this study, however, was to I wish to thank Dr Peter Myerscough and compare the seasonal changes in the overall Professor Harold Heatwole for their helpful structure of forest canopies as interceptors of criticism of the manuscript; Dr Anthony skylight, and to relate these to the environment Underwood for statistical advice; Bert Lester of the understorey (Horn 197 1). This investi- for photographic assistance; Margaret French gation showed that not only did the amount of for field assistance: Sandra Pont for her clerical skylight vary between the three rainforests, but skills; and Professors Joseph Connell and Peter also the proportions of gaps to sun flecks. Light Ashton for many stimulating discussions on was composed mainly of flecks in the CNVF rainforests. This research was funded in part by with its dense, deep canopy, while gaps were a Sydney University Postgraduate Studentship most frequent in the higher elevation MFF with and an Australian Museum Grant. The manu- its thinner, more disturbed canopy. These script was completed during the tenure of an differences have obvious consequences for Australian Research Grants Scheme grant at regeneration: seedlings appear less limited by the University of New England. light in the MFF where temperature or some other environmental condition is probably the limiting factor. Regrowth is rapid in the gaps of References the CNVF where light is probably a major Anderson M. C. (1964) Light relations of terrestrial plan1 factor required to stimulate regeneration. All communities and their measurement. Biol. Rev. 39, three rainforests had relatively low unobscured 425-86. Brasell H. M., Unwin G. L. & Stocker G. C. (1980) The sky visible through the canopy, and it is well quantity. temporal distribution, and mineral- documented that disturbances (i.e. light gaps) element content oflitterfall in two forest types at two are an important mechanism to stimulate sites in tropical Australia. J. Ecol. 68, 123-39. regeneration of suppressed seedlings (Connell Connell J. H. (1978a) Tropical rain forests and coral reefs as et al. 1984). These patterns of rainforest open non-equilibrium systems. In: Popuiution D.vnarnics: S,y:v,npusiurnof the British Ecological regeneration are currently under investigation Society, London. .4pril 1978 (eds R. M. Anderson, (Webb el al. 1972; Connell 1978a,b). L. R. Taylor and B. Turner) pp. 141-63. Blackwell These analyses of canopies and light Scientific Publications, Oxford. interception emphasize the importance of Connell J. H. (1978b) Diversity in tropical rain forests and disturbance and gap formation in the growth coral reefs. Science 199. 1302-10. Connell J. H., Tracey J. G. & Webb L. J. (1984) dynamics of the rainforest community where Compensatory recruitment, growth and mortality as light penetration does not change seasonally. faclors maintaining rain forest tree diversity. Ecol. Presumably, in the low-light environments of Monogr. 54, 141-64. the CNVF and SNVF canopies, the~;eis little Davis T. W. A. & Richards P. W. (1933) The vegetation of Moraballi Creek, New Guinea: an ecological study of opportunity for understorey individuals to be alimited area ofTropical Rain Forest. Part 1. J. Ecol. released from a suppressed state unless a 21, 350-84. Evans G. C. & Coombe D. E. (1959) Hemispherical and Webb L.J. (1959) A physiognomic classification of woodland canopy and the light climate. J. Ecol. 47. Australian rain forests. J. Ecol. 47, 55 1-70. 103-13. Webb L. J. (1968) Environmental relationsh~ps of lhe Halle F.. Oldcman R. A. A. & Tomlinson P. B. (1978) structural types of Australian rain forest vegetation. Troptcal Trees and Forests - An Architectural Ecologv 49, 296-3 1 I. Ana1,vsrs. Springer-Verlag, New York. Webb L. J., Tracey J. G. & Wlll~ams W. T. (1972) Horn H. (1 97 1) The Aduptrb~>Geotnelrv of Trees. Princeton Regeneration and pattern in the subtropical rain University Press. Princeton, New Jersey. forest. J. Ecol. 60, 675-95. Jackson M. T. (1966) Effects of microclimate on spring Webb L. J.. Tracey J. G.. Wlll~amsW. T. & Lance G. N. phenology. Ecolog~'47, 407-1 5. (1969) The pattern of mineral return in leaf litter of Lowman M. D. (1982) Leafgrowth d-vnamics and herbtvory three subtropical Australian forests. .411rt. For. 33. rn Auslralran rarn ,forest canopres. PhD thesis, 99-1 10. University of Sydney, Sydney. Lowman M. D. (1984) An assessment of techniques for measuring herbivory; IS rainforest defoliation more intense than we thought? Biotropica 16, 264-8. Richards P. W. (1952) The Troprcal Rarn Forerr. Cambridge Un~versityPress, Cambridge. (Final manuscript received November 1985)