Field guide for measuring wood density(2.2 MB) • Download
Field guide for measuring tree height (0.4 MB) • Download
Total number of cores: 1095
Total number of species: 200 (35 spp at Korup, 61 spp at La Planada, 51 sp at Yasuni, 57 spp at Yasuni-Tivacuno)

Figure. Histogram of the wood density measurements at three of the CTFS sites: Korup (Cameroon), La Planada (Colombia), and Yasuni (Ecuador). The Tivacuno site is also in Ecudaor.

 

Project funded by the CTFS (2005-2007)

Introduction
Tropical forest ecosystems respond to environmental changes, and this may feedback on the global carbon cycle. Most forest biomass is held in large trees, and a full recovery from disturbance takes typically the time for a tree to reach the canopy story, on the order of at least one hundred years. A decreased precipitation regime, for instance due to a higher frequency of El Nino events, may induce sudden losses of ecosystem functions, and increased plant mortality. The impact of such intense droughts on tropical forests have been well documented for the 1982-1983 El Nino on Barro Colorado Island, central Panama (Condit et al. 1995). Moreover, a large concentration of atmospheric CO2 alters the photosynthetic activity of tropical forest ecosystems, leading to a more rapid turnover, and a higher productivity. This in turn may cause an overall increase in tropical forest pools (Malhi and Phillips 2004). However, the study of these effects faces special difficulties: forests typically have a slow dynamics, and detecting trends requires long and well-documented time series.
Permanent forest plots represent unique tools for assessing these ecological changes. Studies on demographic changes have already been published (Laurance et al. 2004) but these do not permit a direct assessment of the consequences of forest disturbances on the carbon cycle. Biomass is a crucial variable to quantify the amount of ecosystem disturbance, and the impact on atmospheric carbon dioxide concentration. Recent research projects have studied these questions, and they led to major advances to our understanding of the ecophysiology of tropical trees. Unfortunately the compilation, standardization, and continuous monitoring of permanent forest plots is lagging behind the development of sophisticated technologies. At present, there is limited information on the long-term response of tropical forests to environmental changes. The most ancient permanent tropical forest plots were established in the 1940 (Manokaran and Swaine 1994, Sheil et al. 2000), and they concern only a few hectares of forest. The number of permanent plots has considerably expanded since the 1970s, but the total sampling effort represents no more than 100 ha, or about 50,000 trees > 100 mm dbh.
The CTFS network is an unparalleled dataset: it spans over 20 years of demographic monitoring for saplings and trees both in the Neotropics and in the Paleotropics. The objective of this project is to standardize existing information, and to collect new field data that will be useful for refining the current biomass estimates. The present project offers to provide accurate estimate for the aboveground biomass of every sapling, tree, and liana in the CTFS sites using allometric relationships, and based on new field data on wood specific gravity and total tree height. Beyond the goal of the present project, wood density and tree height data should form the basis of a new comparative dataset for the CTFS sites. These are important parameters to study the range of life history strategies, which should be useful for testing theories of plant diversity and species coexistence.

Scientific context
The present project offers to provide accurate estimate for the aboveground biomass (AGB) of every sapling, tree, and liana in the CTFS sites using allometric relationships. Previous studies based on direct harvest experiments of over 2,200 old-growth and secondary forest trees across the tropics revealed that most of the variability in tree-level AGB is predicted by the stem dbh, total height, and wood specific gravity (J Chave et al. in press). The AGB of each tree was estimated from information on the tree geometry, on wood specific gravity, based on species-level or genus-level information available in the literature, and knowledge about the forest type.
The strategy for measuring AGB stocks and changes in tropical forests was tested on nine of the CTFS sites during the Fourth CTFS Analytical Workshop, held in the Fushan Forest Station, Taiwan in August 2004. For the purpose of this exercise, I provided sets of allometric models that predict AGB based on tree dbh, wood specific gravity, and the forest type (three types; dry forests: < 1500 mm/yr, moist forests: between 1500 and 3000 mm/yr, and wet forests > 3000 mm/yr and no seasonality, or cloud forests). Dbh is an obvious predictor of AGB , and it is readily available in the census datasets. Wood specific gravity can be obtained from the literature. I collected literature data on a large number of wood specific gravity values for the CTFS plots. For 28% of the species, a literature value could be found at the species level (from 6% for La Planada to 54% for Mudumalai). Including species for which genus-level information is available, this percentage increases to 83% (from 64% for the two African sites to 92% for Lambir).
Tree height was not included as a predictive variable, because data are lacking for most of the sites. Instead, the forest type classification captures some of this variability: moist forest trees tend to be more slender than both dry and wet forests. This appears to be a serious however for some of CTFS sites. For instance, by the above definition, Yasuni should be a wet forest site, but this forest has relatively tall trees with a canopy at around 30-40 m, comparable to BCI, which is a moist forest. Using the wet forest allometry for Yasuni, AGB was estimated to be 202 metric tons per hectare, while using the moist forest allometry a more reasonable value of 278 tons/ha was obtained. Likewise, the Fushan forest plot should be classified as a wet forest, but the canopy seldom exceeds 20 m, due to the high frequency of typhoons that knock down emergent trees. Applying the wet forest equation, AGB is estimated to be 270 tons/ha, an improbably high estimate. From these few observations, I conclude that while the biomass estimation allometry that ignores tree height produces reasonable results for most forests, considerable caution should be exerted when estimating the AGB stock of a number or rarer forest types, in particular lowland wet forests, and forests submitted to frequent disturbances. An obvious way of improving these estimations would be to collect more tree height data.

Project
Wood specific gravity is one measure of wood density, which measures the volumetric mass of a wood sample. By convention, we use the following definition in the project: wood specific gravity (WSG) is defined as the oven-dry weight (in grams) divided by the green volume (in cm3) of the sample. This measure varies from about 0.1 for balsa (Ochroma pyramidale, Malvaceae) to over 1.2 for lignum-vitae (Guaiacum sanctum, Zygophyllaceae). This measure differs from the classic definition of wood density used by most forestry institutes, and based on the measurement at 12% or at 15% moisture content.
Data on wood specific gravity and on total tree height, not readily available across all of the CTFS sites, and the present project proposes to fill this gap. Such data have already been collected in Barro Colorado Island, central Panama (n > 7,000; data from S O'Brian, S Bohlmann, D King, SJ Wright). A new campaign of tree height and WSG measurements is being carried out, coordinated by H Muller-Landau (2004). We are not planning to acquire more information at this site. Likewise, tree height data were collected recently at Pasoh and at Lambir as part of a CTFS Research Grant Project awarded to Dr. D King, and we hope to establish collaborative effort for these sites, rather than increase the sampling effort more data at Lambir and Pasoh.
We will focus the present project on four of the lesser documented CTFS sites: Yasuni (Ecuador), La Planada (Colombia), Korup (Cameroon), and Sinharaja (Sri Lanka). At each of the site, search of the existing local literature and field data collection will be combined to provide improved wood specific gravity and total tree height datasets. Wood specific gravity measurements will be taken for the species representing the most basal area. Five trees > 100 mm dbh will be cored for each species, if possible for different diameter classes. There is no evidence that coring a tree decreases its survival potential. However, to avoid any risk of surmortality within the permanent plot, trees will be cored outside, and at least 100 m from, the plot. Each cored tree will be tagged, located, identified to the species, and a dbh measurement will be taken. Rare genera and species for which literature data are missing will also be sampled, if individuals can be found outside the plot. Overall, we hope to take samples from 100 to 200 trees at each site. This would take ca. 1 month of labor work.
In addition, the local forestry literature will be searched for additional values of species-specific wood density in the associated institutions operating a CTFS permanent plot. It is very likely that many more data will be found in grey publications. Spreadsheets will be prepared ahead of time, and it will be asked to the collaborators to translate them (if necessary), and to fill them out as precisely as possible (taxonomic information, synonymy, moisture content of the wood). Wood density values will subsequently be converted into an estimate of WSG using a simple model validated for a large range of species (J Chave et al, in preparation). This would be a great personal project for an undergraduate student of the associated universities. Students supervised by the collaborators and who would like to get involved in the project will be most welcome.

Collaborators
Latin America: Yasuni FDP, Ecuador. Collaborator: Dr. Renato Valencia, Pontifica Universidad Catolica del Ecuador, Quito, Ecuador. La Planada FDP, Colombia. Collaborator: Dr. Cristian Samper, Smithsonian Institution, NMNH Washington DC, USA; and Martha Isabel Vallejos, Instituto Alexander von Humboldt, Bogota DC, Colombia.
Africa: Korup FDP, Cameroon. Collaborators: Dr Duncan Thomas, STRI USA, and Dr. George Chuyong, University of Buea, Buea Cameroon.
Asia: Sinharaja FDP, Sri Lanka. Collaborators: Drs. Nimal and Savitri Gunatilleke, University of Peradeniya, Peradeniya, Sri Lanka.






Above: wood density measurement protocol at the Korup Forest Dynamics Plot (photos: Duncan Thomas)