Variation in carbon cycling among four tree species in a tropical rain forest

Ecologists have long sought to explain how species affect ecosystem processes in tropical forests and more urgently now as tropical forests face climate and land-use change that shift species composition. Roughly 50 million hectares of primary forest diverse in species have been turned into tree plantations mostly in monodominant stands. To describe and predict the ecosystem processes of such plantations for regional and global assessments of forest carbon storage, we must understand how subtle differences in tree species traits affect forest carbon cycling. In a tropical rain forest of Costa Rica, twelve tree species were planted in monoculture plots, and four have survived to have almost two-fold difference in carbon stored in forest plant biomass. I examine the variation and its cause in this dissertation.;I began by examining the variation in foliar respiration and wood CO 2 efflux rates among species and canopy layers in Chapter 2. Understanding how species affect forest carbon cycling requires that autotrophic respiration be measured and placed in a complete carbon budget. However, extrapolating measurements of autotrophic respiration from chambers to ecosystem remains a challenge in tropical forests. High plant species diversity and complex canopy structure may cause respiration rates to vary and introduce bias in extrapolation. I examined if foliar respiration and wood CO2 efflux rates vary among species and canopy layers and whether the variation was related to commonly used scalars, mass, N content, photosynthetic capacity, and wood size.;The variation in foliar respiration and wood CO2 efflux rates showed that vertical sampling reduces bias more than temporal sampling. Foliar respiration rate increased three-fold with height and averaged ~0.74 &mgr;mol m-2 s-1 in overstory and ∼0.25 &mgr;mol m-2 s-1 in the understory. Leaf mass per area, leaf nitrogen, and photosynthetic capacity explained some of the variation in foliar respiration rate, but canopy layer or height explained the most. Foliar respiration rate was similar among species. Chamber measurements of foliar respiration can be extrapolated to the canopy with rates and leaf area specific to each canopy layer or height class. If area-based rates are sampled throughout the canopy, foliar respiration may be extrapolated with total leaf mass by regressing the area-based rate against leaf mass per area to derive mean respiration rate per unit mass. Wood CO2 efflux for overstory trees averaged 1.0 - 1.6 &mgr;mol m-2 s-1 for overstory trees and 0.6 - 0.9 &mgr;mol mm-2 s-1 for understory species. The variation in wood CO2 efflux rate was mostly related to wood size expressed as the ratio of wood surface area to mass. Wood CO2 efflux rate was similar among species. Mean wood CO2 efflux can be extrapolated to the stand using surface-area based rate, derived by regressing CO2 efflux per unit mass against the ratio of surface area to mass, and total woody tissue surface area. The temperature response of foliar respiration was similar among species, and wood CO2 efflux was similar between wet and dry seasons. Air temperature lacked strong seasonal trends.;I then constructed a complete and detailed annual carbon budget in Chapter 3, to examine if species difference in biomass carbon storage resulted from differences in gross primary production (GPP) or partitioning of GPP to components with lower turnover and respiration rates. Gross primary production and its partitioning control the behavior of ecosystem biogeochemistry models used for assessments in tropics and elsewhere, but they are difficult to measure at the detail necessary to parameterize the models and are often unknown at a site, especially for tropical forests. As more primary forests are converted to monodominant plantations, determining how species affect carbon budgets is never more important.;I quantified the annual values of GPP, NPP, respiration, and biomass, and estimated fraction of GPP partitioned to components canopy, wood, and roots, fraction of component flux respired, and biomass turnover rate. Respiration, NPP, and biomass were measured for each component, and summed to estimate GPP and partitioning fractions. Biomass was estimated from annual inventories. Net primary production was estimated by summing biomass turnover and change in biomass carbon stored in components between annual inventories. Biomass turnover was measured using litter traps and fine root ingrowth cores. Respiration was scaled from chamber measurements made in Chapter 2.;I found that species differences in GPP, not partitioning, explained species differences in NPP and respiration for canopy, wood, and roots. Species took up 3070 -- 4490 gC m-2 year-1 as GPP, respired 2120 -- 3200 gC m-2 year-1, allotted 948 - 1280 gC m-2 year-1 to NPP, and stored 7590 -- 139000 gC m-2 as biomass. Species partitioned 33% of GPP to canopy, 26 -- 37% to wood, and 35% to roots. Species respired more than half of fluxes partitioned to components, 59 -- 72% in canopy, 63% in wood, and 81% in roots, and the high respiration rates of all components may have constrained the variation in partitioning fractions among species. Species with greater GPP had proportionally greater NPP, and species with greater wood NPP had larger biomass storage size and faster storage rate. Species difference in LAI explained species difference in GPP. Species difference in LAI was explained by both leaf mass and thickness, indicating that a morphological trait played a critical role in determining carbon cycling at the whole stand level. Accurate representation of leaf thickness in models may account for a majority of species differences in C fluxes. The current (2007 -- 2010) data indicate that species had accumulated LAI since a previous study (2003 -- 2005) but declined in NPP, suggesting that increased LAI alleviated the decline in GPP or NPP and that the LAI -- GPP relationship is empirical. Species will likely further decline in NPP or increase in turnover rate, and current partitioning fractions and turnover rates may not hold across time.;The decline in NPP can only be caused by a decline in GPP or an increase in respiration, and I examine the role of leaf carbohydrates in regulating photosynthesis and respiration to balance carbon budget at the whole tree level in Chapter 4. Trees are thought to balance their carbon budget with feedback from carbohydrate storage. Trees cannot increase storage size forever as NPP decline, and large storage size may decrease photosynthesis or increase respiration to remove excess carbohydrates. Some form of this feedback regulation is used in modeling plant growth and ecosystem biogeochemistry, to prevent carbohydrate storage from becoming too large. However, the storage pool serves as a source of carbohydrates during night and seasonal dormancy, and the pool size likely fluctuates before triggering any regulation. The changes in storage pool required to trigger the feedback regulation of photosynthesis or respiration remain unquantified, and thus their generality and importance in ecosystem processes are still unknown.;I tested the carbohydrate regulation of photosynthesis and respiration using girdling. Girdling severs phloem to stop carbohydrate export while leaving xylem intact for photosynthesis to continue, to accumulate carbohydrate in leaves simulating carbon imbalance. On the branches in the upper canopy, I varied girdling intensity by girdling in quarter increments and surrounded a target branch with fully girdled ones, to create a gradient in leaf carbohydrate content. Light saturated photosynthesis rate was tracked in situ, and foliar respiration rate and leaf carbohydrate contents were measured after destructive harvest at the end of the treatment duration.;Girdling intensity had no effect on leaf carbohydrates, respiration, and reduced photosynthesis only under full girdling suggesting that leaf carbohydrate content is tightly regulated and thus decoupled from whole plant carbon balance. Girdling intensity did not vary leaf carbohydrate content and respiration in any species. Photosynthesis declined only under full girdling in three of four species, and one species did not respond at all. Because girdling also stops the export of hormones and reactive oxygen series, girdling may induce physiological changes unrelated to carbohydrate accumulation and may not be an effective method to study carbohydrate feedback in leaves. Leaf carbohydrate content may be regulated through phloem transport and sink activity (growth and storage elsewhere) in addiction to photosynthesis. Leaf carbohydrate content may be far removed or even decoupled from whole plant carbon balance, and may not play a role in regulating changes of GPP and respiration as trees age.;This dissertation shows that models of ecosystem biogeochemistry may represent carbon budgets of different species in monodominant stands reasonably well with LAI and without species-specific partitioning fractions. Despite morphological and physiological differences regulating carbon fluxes, species partitioned remarkably similar fraction of GPP to components and to respiration. Tree species in this environment may be constrained in partitioning of GPP. However, model predictions of carbon budgets based on LAI will not capture ontogenetic changes. Leaf carbohydrates may be decoupled from ontogenetic changes in whole tree carbon balance.