The CARBONO Project:
Long-term landscape-scale monitoring of tropical rain forest productivity and dynamics
David B. Clark and Deborah A. Clark
The CARBONO Project ("carbono" means carbon in Spanish) was launched in 1996 by Deborah A. Clark, David B. Clark, and Steven F. Oberbauer. From its inception, the project has focused on how tropical rainforest productivity and dynamics are affected by interannual climatic variation and long-term climatic and atmospheric changes. The project was explicitly conceived to serve as a cross-disciplinary research magnet, such that collaborators in many disciplines could build on the core long-term field data-series and the unique landscape-scale plot network. The success of this idea can be seen in the diversity of topics covered in the 100+ publications from this project to date.
Two key and novel ideas motivated the project's design. First was a focus on landscape-scale phenomena. Our interest was the performance of old-growth tropical rainforest across the within-landscape gradients of topography and soil nutrients. Such a perspective cannot be addressed with a single forest-inventory plot, the typical research design at tropical-forest field sites. A second innovation was our focus on forest performance at annual and sub-annual time-scales. Climate is everywhere seasonal, including in the wet tropics. The CARBONO project was designed to make it possible to assess the effects of annual and subannual climatic variation on forest productivity and dynamics, over the long term.
Clark, Clark, Oberbauer and collaborators developed the first grants for the CARBONO Project. Initial support came from NSF, DOE, the Andrew W. Mellon Foundation, and Conservation International's TEAM Initiative. Subsequent funding has been provided from other programs within NSF and NASA. Today (2012) the core measurements are being supported by a 5-yr grant from NSF's LTREB program to David B. Clark and Deborah A. Clark, with additional funding from NASA to D.B. Clark. Steven Oberbauer leads a parallel NSF-funded research program at La Selva, monitoring forest-atmosphere exchanges of CO2, water, and energy using eddy flux methods.
Go to top
The CARBONO project is based on field measurements carried out across the project's network of 18 0.5 ha plots. The 0.5 ha plot size was selected based on both the spatial scale of significant within-landscape variation in forest basal area, and the grain-size of the different habitats of interest, particularly steep slopes. Three edaphic categories were identified as capturing the important axes of upland (non-swamp) landscape structure: (1) moderately fertile flat sites on old alluvium (the youngest oxisols at La Selva); (2) relatively infertile flat sites on ridge tops; and (3) relatively infertile steep slopes. Plot locations were selected using the La Selva GIS to produce a stratified random design to sample the three edaphic categories. We used random numbers to select plot coordinates within the GIS coverages of soil types and slope. Potential plot locations that crossed streams, main trails or protected areas were omitted. Plot selection was therefore stratified-random within edaphic categories. An important feature of the plot-network design was that the researchers had no knowledge of forest structure at a selected plot location. As a result, the network provides a representative, objective assay of the entire landscape. Plots were surveyed with a tripod-mounted surveying compass and were based on slope-corrected distances (subsequently verified by GPS-location of the plot corners). In each plot, permanent monuments were installed at 10-m spacing.
The soils of the CARBONO plots were sampled and characterized chemically in 1998 (total C, N, P, cations, Fe, Mn, Al, pH) by E. Veldkamp, J. Mackensen, and D. B. Clark (data published in Appendix L, Espeleta & Clark. 2007. Ecological Monographs),
Bulked samples from six cores per depth (0–10 cm, 10–30 cm, 30–50 cm, 50–100 cm), regularly spaced across each plot, were taken in each Old Alluvial terrace and Residual Soil plateau plot; bulked samples from six cores per depth, regularly spaced across upper, middle and lower slope positions, were taken from each Residual Soil slope plot. We calculated soil element stocks for each depth by multiplying element concentrations by soil bulk density. Bulk density was determined (Veldkamp et al. 2003) for the two soil types as mean values from each depth in pits adjacent to three plots on each soil (Old Alluvial terrace and Residual Soil plateau; for the Residual Soil slope plots we used Residual Soil plateau bulk density values). The plot network spans an approximately two- to threefold range in total soil stocks of phosphorus (P), potassium (K), and calcium (Ca)..
Core annual-subannual measurements
From project inception in September 1997, the following core long-term field measurements have been continuously maintained in all 18 0.5-ha plots of the CARBONO Project plot network (the current funding extends through the 2013 census):
- Annual aboveground biomass (stocks, increment) and annual stem dynamics (recruitment, mortality), at the stand level and by tree-species & tree functional group
From 1997 onward, all live stems >10 cm diameter have been mapped and identified (all trees and palms to species, lianas treated as a group), as well as all stems 1-<10 cm diameter in a standard 10x10 m subplot per plot (since 2005). In the annual censuses (Sept.-Oct.) all stems are evaluated for survival, new stems recruiting to the minimum size are mapped and identified, and for all live stems the diameter is measured (nearest mm) at a permanently-marked point above buttresses or other irregularities.
- Fine litterfall
Fine litterfall (leaf, wood < 1 cm diameter, reproductive material) has been collected biweekly in all plots since project inception. The litterfall is collected from nine pairs (1 standing basket + 1 ground trap for large [>50 cm]) leaves) of 0.25 m2 traps, sited in a standardized grid array across each plot. The litterfall is sorted by category and dried at 65oC to constant mass. The leaf-litter is then ground and incorporated into the longterm leaf-litter archive (maintained to enable subsequent chemical and isotope analysis).
- Coarse woody debris (CWD) (> 10-cm diameter: initial stocks, annual inputs) and 1-10 cm diameter wood-fall (monthly inputs)
In annual censuses of all 18 CARBONO plots, all new pieces of fallen CWD > 30 cm in diameter are mapped, measured, and assessed for decomposition condition (Clark et al. 2002). In the annual CWD census, all new pieces of CWD 10-<30 cm dia. are also mapped, measured and assessed for condition in a standard 10x50 m subplot. The source of all CWD is noted (branch-fall or tree-fall, tree ID). The height and diameter of all standing-dead stems (> 10 cm dia.) are also measured. Intermediate-sized (1-<10 cm dia.) woody litterfall is collected monthly from the fine-litterfall ground-traps (see above) and dried at 65oC to constant mass.
- Soil moisture
Beginning in April 1998, volumetric soil moisture has been assessed during all biweekly fine-litterfall collections with a permanent vertical FDR sensor measuring over the 0-30 cm depth layer of soil in the center of each of the 18 CARBONO plots.
- Canopy dynamics
Annually, canopy heights in each plot are mapped based on measurements at the 231 points of a 5 x 5 m grid across the plot. Heights to 15 m are measured with a telescoping measuring pole, with verticality checked with a clinometer; sites higher than 15 m are classed as “high canopy.”
Go to Top
Results to Date
The high and sustained productivity of the project, approximately 100 publications to date, is evident in the list of publications. The project has amply fulfilled our original goal to serve as a research magnet for collaboration researchers. Over 100 researchers have co-authored papers related to data developed from the CARBONO Project.
The biggest scientific findings to date have centered on the tight coupling of forest performance to climate. Hotter nights and drier dry seasons both decrease wood production. The temperature and drought effects are statistically independent, a fact that only became detectable as the length of the data series increased. Because climate will surely warm and may well become drier in tropical rain forest areas, these are troubling results with global implications. Increasing atmospheric concentrations of CO2 have been hypothesized to be stimulating increased productivity in tropical rain forests. Our data-series, by far the longest to date from this biome, show that any stimulatory effects attributable to "CO2 fertilization" are trivial compared to the negative impacts already occurring from higher temperatures and from water-limitation.
Research findings from the first 12 years of the CARBONO Project have spanned five broad thematic areas:
- Forest structure (aboveground biomass and dynamics, CWD, LAI, plot floristics)
- Aboveground NPP(aboveground wood production/tree growth, litterfall)
- Stand-level physiology, leaf and bole respiration, VOCs
- Soil chemistry (soil nutrients & relation to litter invert.s, soil N processes), soil moisture, dry/wet deposition, throughfall
- Belowground C processes: C allocation, soil respiration, DOC, SOC stocks and dynamics, fine-root stocks and variation through time, mycorrhizae, microbes
- Remote sensing of forest structure and process (biomass, tree growth/mortality, gap dynamics)
D.A. Clark and D.B. Clark are forest ecologists. The research to date in other areas has proceeded in large measure by our seeking out scientists with additional expertise and offering to collaborate on interdisciplinary projects of mutual interest.
Go to top
From 1980 - 2010 D.A. Clark and D.B. Clark lived on-site and managed the project through daily interactions with the three long-term Costa Rican field technicians participating in the project. In 2010 the Clarks moved to the United States. Since then the daily meetings with technicians have been conducted with Skype video calls. At the end of each field day the technicians scan the day's field sheets and email them to the Clarks, who review them before the next day's meeting at the start of the work day. In addition the Clarks make 2-3 trips a year to La Selva to continue technician training, update equipment inventories, and address any issues that can't easily be worked on in video calls. The combination of daily contact supplemented by regular physical visits is working smoothly.
With the annual census of 2013 we will complete the 17th consecutive annual census of the CARBONO plots. The CARBONO Project remains unique for the tropical-forest biome. At no other tropical-forest field site has forest performance been monitored over the long term at the landscape scale and at annual to sub-annual timesteps. The plots have served as a potent magnet to attract collaborating researchers, particularly those in remote sensing ecology. The project's location in the Republic of Costa Rica, which continues its traditional welcoming policy to foreign researchers, has been another key advantage of the project.
Go to top