Carbon Fluxes and Pools in a Montane Rainforest in Sulawesi, Indonesia
by Florian Heimsch
Date of Examination:2020-11-20
Date of issue:2021-03-18
Advisor:Prof. Dr. Alexander Knohl
Referee:Prof. Dr. Alexander Knohl
Referee:Prof. Dr. Dirk Hölscher
Files in this item
Name:Dissertation_Florian_Heimsch_eDiss.pdf
Size:5.48Mb
Format:PDF
Abstract
English
Forests play an important part in the global carbon cycle. While carbon balances of many forest types have been studied intensively, such information is very sparse for tropical montane rainforests. This thesis aims to establish a robust estimate of carbon storage, respiration and the overall carbon balance of such a forest in the central mountains of Sulawesi, Indonesia. To this end, three approaches were combined. Forest inventories were used to determine forest stand parameters such as aboveground biomass (AGB), basal area (BA) and changes thereof; soil respiration (Rs) measurements provided an insight into the major component of ecosystem respiration and eddy-covariance (EC) measurements were used to estimate net ecosystem exchange (NEE). For long term changes in AGB, results from an inventory carried out in 2006 were used as baseline. These were combined with results from a large scale inventory carried out in 2017, which resulted in an estimated annual increase of AGB of 2.54 Mg ha-1 y-1. Based on the 2017 data, AGB was estimated to be 321.38 ±9.64 Mg ha-1and BA 37.86±0.38 m² ha-1. To quantify Rs and to identify its main drivers, Rs, soil temperature (Ts) and soil moisture (Ms) were measured 21 times at 22 locations at the study site over the span of one month. Forest density in the form of BA and mean DBH were known for each location through the forest inventory mentioned above. I used a range of models to analyse interactions between Rs, Ts, Ms, mean DBH, BA and further variables. BA was the most important factor controlling Rsrates, exceeding the effects of Ts and Ms. Linear mixed effect models including BA reached the best results of all models, explaining up to 88% of Rs variability. Rs over the measurement period was 6.2 ± 1.8 µmol m-2 s-1, whereby spatial variability (CV=137%) exceeded temporal variability significantly (CV=60%). Estimates of NEE based on EC for tropical rainforests are often problematic due to unfavourable measurement conditions at night. Such conditions exist at this study site as well and led to the dismissal of the majority of measured night-time fluxes. Four gap-filling strategies were tested of which only two led to defensible annual NEE rates. The most realistic outcome is achieved by replacing all night-time measurements with modelled data. For this purpose, the Mixfor-SVAT model, which has been used for the site before, was re-parameterised with my new Rs measurements. This led to a modelled night-time respiration of 9.01±1.47 µmol CO2 m-2 s-1 which together with day-time EC measurement data resulted in an estimated annual NEE of -4.07 to -5.29 Mg C ha-1 y-1. The three approaches together show that the forest at the study site is a persistent carbon sink and enabled the first defensible annual NEE estimate of a tropical montane rainforest worldwide.
Keywords: tropical montane rainforest; forest inventory; aboveground biomass; basal area; soil respiration; chamber measurements; eddy-covariance; gap-filling; net ecosystem exchange; carbon cycle; Southeast Asia; Sulawesi