Apers, S., M. Bechtold, A. J. Baird, A. R. Cobb, G. Dargie, H. Hidayat, T. Hirano, A. Hoyt, Y. Ishii, A. Katimon, R. D. Koster, M. Lampela, S. P. Mahanama, L. Melling, S. Page, R. H. Reichle, H. Takahashi, M. Taufik, J. Vanderborght, and G. J. M. De Lannoy:
"Extending the PEATCLSM Framework to Tropical Peatlands: Model Evaluation for the Major Tropical Peatland Areas"
Presentation at the AGU Fall Meeting, Online, 2020.

Abstract:
Tropical peatlands have specific water storage dynamics that exert a first-order control on their internal processes and functioning, and distinguish them from surrounding mineral landscapes and northern peatlands. The integration of tropical peat-specific hydrology in an existing global-scale land surface model (LSM) will help our understanding of the peatland sensitivity to external disturbances.

Here, we present the first-ever global-scale LSM specifically adapted for tropical peatland hydrology. Earlier, a module for natural northern peatland processes (PEATCLSM_N,Natural) was embedded within the NASA Goddard Earth Observing System (GEOS) Catchment land surface model (CLSM) by Bechtold et al. (2019). We developed a literature-based parameter set for natural (PEATCLSM_T,Natural) and drained (PEATCLSM_T,Drained) tropical peatlands. As an additional feature, the PEATCLSM_T,Natural scheme includes a plant oxygen-stress function to resolve reduced transpiration at high groundwater tables. The basic CLSM structure and the same global input data were used in all PEATCLSM versions to allow future use in operational GEOS products. A suite of simulations with PEATCLSM_T,Natural, PEATCLSM_N,Natural, and the operational CLSM version (CLSM_O, which includes peat as a soil class) was conducted over the Amazon Basin, the Congo Basin and Indonesia. A PEATCLSM_T,Drained simulation was added to the Indonesian suite to include the large fraction of drained tropical peatland areas. A preliminary evaluation with in-situ observations of groundwater table depth (WTD) and evapotranspiration (ET) shows overall improvements for simulations with PEATCLSM_T,Natural compared to those with the PEATCLSM_N,Natural and CLSM_O versions for each region. Despite these improvements, strong regional differences occur. The PEATCLSM_T,Natural average WTD bias for 9 sites in the Congo Basin is around -0.21 m, compared to -0.02 m for 20 natural sites in Indonesia. When evaluating at 42 drained sites in Indonesia, the average WTD bias is -0.27 m for CLSM_O and 0.56 m for PEATCLSM_N,Natural, and improves to 0.03 m for PEATCLSM_T,Drained. It is expected that regional parameter tuning could further improve the results over the Congo Basin, but this is not common for operational global LSMs. An evaluation over the Amazon Basin is still ongoing.

Home

NASA-GSFC / GMAO / Rolf Reichle