1. Compositional Shifts in Undisturbed Neotropical Forests: Effects of Climate Change? William F. Laurance 1,2 & Richard Condit 1 William F. Laurance 1,2 & Richard Condit 1 1 Smithsonian Tropical Research Institute, Panama 2 Biological Dynamics of Forest Fragments Project, Manaus, Brazil 2 Biological Dynamics of Forest Fragments Project, Manaus, Brazil

3. Atmospheric CO 2 Pre-industrial level was 280 ppm Pre-industrial level was 280 ppm Current level (378 ppm) is highest in last 20 million years Current level (378 ppm) is highest in last 20 million years In experiments, elevated CO 2 increases photosynthesis, growth, and water-use efficiency for most plants (for tropical trees, only seedlings assessed) In experiments, elevated CO 2 increases photosynthesis, growth, and water-use efficiency for most plants (for tropical trees, only seedlings assessed) Data from Mauna Loa, Hawaii

4. Evidence for pervasive changes in Amazonian tree communities W. F. Laurance et al. (2004) Pervasive alteration of tree communities in undisturbed Amazonian forests. Nature 428:171-175

5. 18 1-ha plots in forest interiors 18 1-ha plots in forest interiors Monitored 11-19 years Monitored 11-19 years 115 most abundant genera 115 most abundant genera Study Design:

6. Changes in Central Amazonia 1)Concerted shifts in species composition (14 times more than expected by chance), with fast-growing canopy and emergent trees evidently gaining a competitive advantage over smaller, slower-growing trees 2)Accelerating forest dynamics 3)Accelerating growth for most genera

7. Barro Colorado Island  -1500 ha in area -isolated in 1914

8. 50-hectare plot 

9. BCI 50-ha Plot Established in 1982, recensused in 1985, 1990, 1995, 2000 Established in 1982, recensused in 1985, 1990, 1995, 2000 Subdivided into 50 1-ha plots Subdivided into 50 1-ha plots Included only species present in >10 1-ha plots and >10 cm dbh (142 species) Included only species present in >10 1-ha plots and >10 cm dbh (142 species) Bootstrapping used (at 1% alpha level) to assess changes in density and basal area for each species between 1985 - 2000 Bootstrapping used (at 1% alpha level) to assess changes in density and basal area for each species between 1985 - 2000

10. BCI: Changes in Tree Density and Basal Area Tree Density: -19 increased (13%) -23 decreased (16%) Tree Basal-Area: -22 increased (15%) -20 decreased (14%) Overall: -Nearly 30 times more significant changes than expected by chance

11. Amazonia: Differences in Growth Form

12. BCI: No Differences in Growth Form species

13. Amazonia: Larger Trees are Winning IIII

14. BCI: No Effect of Tree Size

15. Amazonia: Winners Grow Faster

16. BCI: No Difference in Growth Rate species

17. Amazonia: Accelerating Forest Dynamics

18. BCI: Decelerating Forest Dynamics Mortality: -1982-85 is higher than the other intervals -1995-2000 is higher than 1990-95 Recruitment: -1982-85 and 1985-90 are higher than both 1990-95 and 1995-2000 -1990-95 is higher than 1995-2000

19. Amazonia: Accelerating Growth Rates Growth accelerated more for winners than losers –Winners: 57% –Losers: 22% –Other genera: 40%

20. BCI: Decelerating Growth Rates -Growth was higher in 1982-85 than in all other intervals (P<0.0001) -Growth was also higher in 1985-90 than in the final two intervals (P<0.0003)

21. What is Driving the Tree- Community Changes on BCI? 1)BCI is a forest fragment

22. 2)Parts of BCI are recovering from past disturbances –70-150 year-old regrowth near the plot: previous source of successional-plant propagules –Successional species are now declining on the plot

23. 3) BCI has experienced a long-term drying trend and recent severe droughts

24. Population Trends vs. Drought Tolerance

25. Conclusions Tree-community changes in central Amazonia are consistent with elevated productivity, possibly from rising CO 2 levels Tree-community changes in central Amazonia are consistent with elevated productivity, possibly from rising CO 2 levels Changes are BCI seem to be driven by recent droughts and possibly nonequilibrium conditions on the plot Changes are BCI seem to be driven by recent droughts and possibly nonequilibrium conditions on the plot The drivers of change may differ among different forests. Thus, there is a dire need for more long-term studies of tree-community dynamics that explicitly test global-change hypotheses The drivers of change may differ among different forests. Thus, there is a dire need for more long-term studies of tree-community dynamics that explicitly test global-change hypotheses

26. Thank You!