Modeling the biomass allocation of tree resprout in a fire-prone savanna

International audience

English. Young savanna trees can quickly grow back from belowground storage structures after topkill. This capacity is atolerance trait that confers persistence at the plant individual level, enabling them to survive diverse disturbanceregimes.We simulated the growth of a single resprouting stem (without and with fire) of a deciduous tree species thatallocates its photo-assimilates during the vegetation season to reserves, belowground and aboveground parts(leaves and stem). As savannas grow under highly seasonal climates, the model considers that trees are leaflessduring the dry season and following growth is only possible thanks to reserves. Stem architecture constrains theleaf biomass to be proportional to stem length rather than biomass. We compared the success of differentallocation strategies, with and without fire and according to the seasonality. To do so, the height of theresprouting stem after 50 yrs and the time to reach 2 m were modeled for three species of a humid savanna.The viable and faster growth strategies are those for which allocation to belowground parts is <40%. There isvery little sensitivity to allocation to reserves since successful growth is observed for allocation to reserves between 0.5% and 85% of photo-assimilates. In the literature and in our results, there is little impact of fire on thestem height or the time needed to escape the fire trap. Our model suggests that (1) allocation to leaves isdeterminant as leaves are the primary source of assimilates that can then be turned into fire-resistant structures(reserves and roots) and (2) fire only weakly slows down the plant growth compared to dry season