Wood respiration in savannas

Abstract

This dissertation investigates respiration by woody tissues in African savannas. Woody tissues are the major component of the biomass in savanna ecosystems. Since wood is made up of long-chain carbon molecules, and when alive is growing or maintaining its cells and thus respiring, it is an active and important component of the carbon cycle at organism, ecosystem and global scales. The wood respiration flux was measured using a LiCor 8100A respirometer on ten common tree species (Combretum apiculatum, Sclerocarya birrea, Senegalia nigrescens, Spirostachys africana, Burkea africana, Ochna pulchra, Strychnos pungens, Terminalia sericea, Philenoptera violacea, and Vachelia nilotica) at three South African savanna sites: Skukuza, Wits Rural Facility, and Nylsvley Nature Reserve. Simultaneously, several factors that could potentially influence wood respiration rates were measured: (1) stem temperature; (2) stem size (height and diameter); (3) woody tissue nitrogen content, (4) woody tissue densities; (5) season of the year; (6) bark and sapwood thickness; and (7) fraction healthy leaves present during measurements (representing tree phenology). The highest wood respiration rates were observed during the warm and wet growing season (October to March), and the lowest during the cooler dry season (April to September). Sapwood volume was found to be the best basis for expressing the wood respiration rate (rather than bark area, for instance). Species respire at different rates per unit sapwood volume. I found that the wood respiration rate is significantly associated with stem temperature and season of the year, and weakly related to the time of day. Wood respiration has a non linear relationship to bark temperature: during the growing season, it increases with temperature up to maximum around 35 to 45 °C and declines at higher temperatures. During the dormant season, wood respiration rates increased with temperature, levelling off but not declining at high temperatures. Using known allometric relationships and the relationships I established between respiration rates and continuously measured environmental variables, I scaled-up wood respiration from the scale of measurement (a chamber area of 0.01 m2 ) to the sum of wood respiration fluxes from a hectare of savanna at the Skukuza flux site, which has long term iii measurements of ecosystem CO2 exchange and soil respiration. I also scaled up the other main savanna ecosystem CO2 fluxes (soil respiration, leaf respiration by trees and grasses, fire emissions, and respiration by insect and mammal herbivores). The sum of all these fluxes is the Total Ecosystem Respiration (TER), which I compared to the ecosystem respiration (R eco) as measured ‘top-down’ by eddy covariance. I found that TER estimates were on average 42% larger than R eco. It is unclear whether this difference results from an overestimation of one or more fluxes in the TER, or an underestimation of R eco. I found that on average TER > Net Primary Production (NPP) > R eco, suggesting that Skukuza is a small source of CO2, which agrees with long-term Net Ecosystem Exchange (NEE) measured at this site. Soil respiration amounted to about 84% of CO2 efflux from this site, leaf respiration (almost equally divided between grasses and trees) 10%, fire 3%, and wood respiration only 1 – 2%, and respiration by herbivores amounted to just 1%. All these fluxes, except those from fires, peak during the wet and warm growing season.