Current Research: Joseph Connell

I do research on several different ecological questions, all of which involve long-term study in several different ecosystems

1. What maintains diversity of trees in tropical and subtropical rainforests and of corals on reefs?
2. How do the patterns of recruitment, growth and mortality in these organisms affect their distribution and abundance?
3. What is the pattern of recovery after disturbances in these trees and corals, and in plants in Californian chaparral and in animals and plants in the Californian marine rocky intertidal?
4. How do the patterns of recruitment, growth and mortality affect the distribution of marine intertidal sessile animals in Scotland, and on the Pacific coast of Washington and California?
5. Ecological succession: what mechanisms determine its course after a disturbance?

1. A Long-Term Study of the Community Ecology of Tropical and Subtropical Rainforest Trees, 1963-1999.

• NSF LTREB grant: long-term monitoring of phenology, early seedling germination, and recruitment, growth, and mortality of all sizes of trees. Co-PI: C. A. Gehring, Associates P. T. Green, I. R. Noble, M. D. Lowman.
• NSF Ecology grant: effects of ground-dwelling vertebrates on mycorrhizae of seedlings, as well as on their recruitment, growth and mortality. Co-PI: C. A. Gehring, Associates T. Theimer, I. R. Noble.

General Approach

Since 1963 the overall aim of the present study has been to investigate some of the mechanisms responsible for both the high diversity of species, and the great variety of population structures among those species, in 2 rain forests in Australia. Predictions of some hypotheses proposed to explain this high diversity have been tested non-experimentally by comparing analyses of population dynamics with patterns predicted from the hypotheses. This method was applied by Connell et. al. (1984) in testing predictions of the compensation hypothesis (Connell 1978). Rates of recruitment, growth and mortality were compared to relative abundances of species to see whether frequency-dependence was occurring among species. Because the smaller stages (seedlings and saplings) have greater sample sizes and faster dynamics than the larger trees, we have concentrated our attention on the former. They are also more suitable for field experimental manipulation.

Study region: characteristics of Australian rain forests Australian rain forests occupy a relatively small area of the continent. Prior to settlement by Europeans there was about 2 million ha of forests now classified as rain forests, about 70% of which have since been cleared. These include tropical and subtropical rain forests, as well as temperate rain forests in the south. As in all other regions there is a cline in diversity of trees, with the greatest diversity in the tropics, declining toward temperate regions. In general, the diversity of canopy trees is less than in the richest forests of Amazonia and Malaysia. Australian rain forest plants are chiefly derived from the Gondwanaland flora, and include many relict plants. There is evidence that the extent of rain forest in Australia was greater during the tertiary epoch, and today the rain forests have become fragmented as in some other regions. This fragmentation has been increased by clearing of forests during the last two centuries. The forests show a great variety of structural form and species composition, both of which appear to reflect the predominant environmental conditions.

Long-Term Monitoring Methods

Since 1963 we have monitored the dynamics of trees and seedlings at two rain forest sites in Queensland, Australia, one in tropical rain forest at 17°S (at Davies Creek, one plot of 1.7 ha, 850 m elevation), the other in sub-tropical rain forest at 28°S lat. (at O'Reilly's, two plots of 1.0 ha each, 2 km apart, 850 m elevation). The tropical plot occurs on relatively infertile soils derived from granite bedrock, while the sub-tropical plot occurs on soils of higher fertility over basalt. These plots are unusual, because, unlike many other tropical rain forests, they have never been subject to cutting by humans. Australian aboriginal peoples were not agriculturalists and entered these forests only to hunt and gather, so there was no cutting of trees for shifting cultivation (Flood, J. 1995. Archaeology of the dreamtime. Revised edition, Angus & Robertson, Sydney, Australia, 328 pp.). A second line of evidence is that many of the common large trees on the plots are highly desirable timber species, and would have been among the first to be cut by foresters (R. Keenan, Queensland Department of Forestry, pers. comm.). Both plots are now completely protected in World Heritage areas.

In 1963, following the usual practice of most previous forest ecological studies, all trees > 10 cm DBH were mapped, tagged, measured and identified on both plots. To obtain samples of smaller trees, on permanently-marked belt transects comprising 30% of each plot and extending throughout the mapped area, the same census was done for all "poles", trees between 2.5 and 10 cm DBH. In 1965, all saplings and seedlings <2.5 cm DBH, including all sizes down to tiny, newly-germinated seedlings, were mapped and tagged on permanently-marked narrower strips nested within those for the poles. The saplings and seedlings were mapped on 9.6% and 16.8% of the total plot areas at the tropical and sub-tropical sites, respectively. Thus all woody plants except vines were mapped, the smaller size classes along transects nested within the areas of larger trees. At intervals of 1 to 4 years thereafter, all previously mapped and tagged individuals were censused, and in addition all newly recruited seedlings that had germinated and survived in the interval since the previous census, were mapped, tagged, measured and identified on precisely the same permanently-marked areas used in the original mapping. These censuses were done 16 times between 1965 and 1999, and serve to indicate the temporal and spatial variation in seedling recruitment of all species of trees. Details of the arrangement of the subplots, topography, census dates, and weather patterns are given in Connell et al. (1984) and Connell and Green (2000). At each census, the positions of all fallen trees were mapped, as was the outline of all gaps in the canopy and lower leaf layers caused by tree falls. Gaps were defined as continuous areas larger than 4 sq. m. in which open sky was visible above a height of 2 m.

The vertical structure of the vegetation has also been measured on several dates since 1980, as follows. The heights of the tops and bottoms of all vegetation layers were measured up to the top of the canopy; these data indicate the thickness of these leaf layers and the size of the gaps between them. Sometimes these spaces are quite large, constituting "subcanopy gaps" beneath an intact upper canopy. We found that small understory saplings beneath an intact canopy grew faster if they were located in a subcanopy gap than if the vegetation was close above them (Connell et al 1997). The horizontal structure, i. e., the spatial variation in the density of the understory vegetation, has been measured between points spaced 4 m apart along the transect lines, on several dates since 1980. Two observers at adjacent points (4 m apart) each estimated theamount of vegetation between them, by estimating how much the intervening leaves obscured a white board held by the other person. The mean of these cover estimates by the two observers between each pair of points yields a relative estimate of the amount of vegetation between them. This was done at each of three heights above the ground, at 15 cm, 1 m, and 2 m.

The phenology of the tree species on the tropical plot is being studied by Dr. Peter Green. Every month since January 1995 he has observed the same individual adults (an average of 7 trees of each of 105 species) for intensity of flowering and fruit crop. He has also mapped all tree seedlings that have germinated each month on 20 transects, 100 x 1m, located in and adjacent to the plot. For the species with abundant germination he has followed their early survival.

To measure the effects of ground-dwelling vertebrates on recruitment, mortality and diversity of seedlings at the tropical plot, in 1995 Dr. Gehring and Dr. Theimer erected exclosure fences around plots of 6 x 7 m, to exclude these animals. Each of the 16 exclosures has an adjacent unfenced control plot, with a fence across the upper side to stop litter from washing into it, as a control for this effect caused by the fences on the exclosure plot. Censuses were made of all seedlings <30 cm height in marked plots within these exclosures and controls, both before the exclosures were closed in 1995, and every 6 months thereafter. Dr. Gehring is studying the effect of these exclosures on the mycorrhizae of the seedlings, as well as the effects of mycorrhizae on growth of several species, using greenhouse experiments on seedlings with and without mycorrhizae.

2. A Long-Term Study of the Community Ecology of Corals and Algae on Heron Is., Great Barrier Reef, Australia, 1962-1999.

NSF Biological Oceanography, PI: J. H. Connell, grants 1962- present. Associates: T. P. Hughes, Dept. Marine Biology, James Cook Univ., and C. C. Wallace, Museum of Tropical Queensland, Townsville, Queensland, Australia.

We have made observations of the recruitment, abundance and diversity of all species of corals and algae in our study sites at Heron Island, Queensland, Australia, over temporal scales ranging from 2 months up to 37 years, and spatial scales from mm to 2 km. At sites in 4 different shallow water habitats, we established permanently-marked sets of quadrats in 1962-63, and have censused these using photographs, at intervals ranging from 2 months to 3 years, through Oct. 1999. Two sites were located on the wave-exposed north side of Heron reef, 2 others on the more sheltered southwest side (Fig. 1 in Connell et al. 1997). In addition, in later years we have sampled larger areas on permanently-marked belt transects and line transects and have continued to census them to 1999. Some of the transects are in the same habitats as the quadrats, others are at greater depths on the reef slopes. With my colleagues T. P. Hughes, C. C. Wallace, J. Tanner, and L. Goldwasser, we have analyzed the population dynamics and interactions between colonies of all species on the permanent plots.

3. Long-Term Studies in Temperate Ecosystems

A. Marine Algae and Sessile Animals in the Rocky Intertidal

At intervals since 1953 I have been studying recruitment, growth, and mortality of barnacles, and their predators and competitors, on permanently-marked sites at Millport, Scotland (Connell 1961, 1974, 1985). At intervals since 1959 I have been studying recruitment, growth, and mortality of barnacles and their predators and competitors, on permanently-marked sites at Friday Harbor, Washington (Connell 1970, 1985). Since 1975, with Drs. W. Sousa, S. Swarbrick, and S. Schroeter, we have been following changes in the abundance of marine benthic animals and marine algae in an intertidal boulder field near Goleta, CA.

B. Land Plants in California

1. Since 1985, with Dr. C. Tyler and Dr. D. Lohse, we have been studying the recovery of chaparral vegetation after fire, on a site where all woody plants had been mapped.
2. Since 1959, in a site in the Mohave desert, I have been following the survival and growth of bushy species. They were first mapped in 1959, and have been remapped at intervals since.

Joseph Connell | Research | Publications

<< Return to Faculty Page