Integrating Microtopography and Vegetation Influences: A Landscape-Level Approach to Estimating CO2 Emissions in Tropical Peatlands

Groundwater level (GWL) and soil temperature are critical parameters for estimating CO₂ emissions from tropical peatlands. However, scaling CO₂ emissions to the landscape level remains challenging because of the heterogeneity of canopy coverage on different land use types that influence soil temperature and the irregularity of peat surface microtopography that regulates water accumulation and drainage. This study aims to address this gap by (1) capturing high-frequency measurements of GWL and soil temperature under varying vegetation canopy covers and microtopographic conditions within the same land use type, and (2) using those measurements along with remote sensing approaches to model CO₂ emissions at landscape level. This work was carried out in a secondary swamp forest (SF) and an oil palm (OP) site in Anjongan Dalam Village, Mempawah, West Kalimantan, Indonesia.  These sites were selected as they represent the predominant land use types in the country’s peatlands. Solinst Levelogger® 5 were installed along with custom-made multi-depth soil temperature sensors at six locations within each site that accounted for varying canopy covers (sparse and dense) in both land use types. To establish a GHG baseline model, we also conducted biweekly monitoring of soil CO₂ flux using LiCOR LI-7810 Trace Gas Analyzer in both land use types.

In the first-quarter of our measurements (September-December 2023), we made three central observations. Firstly, all twelve plots exhibited similar GWL fluctuation patterns in response to wet and inter-storm (no rain) periods, but the magnitude of these fluctuations varied within measurement points in SF and OP (max. discrepancy = 18.6 and 38.5 cm, respectively). LiDAR observations revealed that microtopography controlled these water levels, suggesting that elevation variations influence the magnitude of GWL. Secondly, a comparison of daytime soil temperatures between sparse and dense canopy cover within each land use type revealed differing magnitudes: sparse vegetation areas, likely due to more open canopies, registered significantly higher near surface soil temperatures for SF (mean = 27.3 and 26.8 °C) and OP (mean = 29.6 and 29.4 °C; p-value < 0.01). Thirdly, our findings imply GWL and soil temperature in SF (R² = 0.66 and 0.32) and OP (R² = 0.61 and 0.65, respectively) significantly influence CO₂ emissions. Taken together, these three observations demonstrate how variations in GWL and soil temperature within the same land use may lead to considerable shifts in the resulting CO₂ emission factors. Still, continued monitoring during the dry season is crucial to further elucidate the impacts of microtopography and vegetation cover on CO₂ emissions. The next phase of our research will be focusing on developing our landscape-level GHG model that will incorporate microtopography and canopy covers. Integrating our distributed ground monitoring data with spatiotemporal remote sensing information will be pivotal in introducing new methodology for improving emission factors for other peatlands following our environmental settings.