Modeling spatiotemporal dynamics of global wetlands: comprehensive evaluation of a new sub-grid TOPMODEL parameterization and uncertainties
- 1Dynamic Macroecology, Swiss Federal Research Institute WSL, Zürcherstrasse 111, 8903 Birmensdorf, Switzerland
- 2Institute on Ecosystems and Department of Ecology, Montana State University, Bozeman, MT 59717, USA
- 3Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, 730000, Gansu, China
- 4Department of Environmental Systems Science, Swiss Federal Institute of Technology ETH, 8092 Zürich, Switzerland
- 5Institute of Earth Surface Dynamics, University of Lausanne, 1015 Lausanne, Switzerland
Abstract. Simulations of the spatiotemporal dynamics of wetlands are key to understanding the role of wetland biogeochemistry under past and future climate. Hydrologic inundation models, such as the TOPography-based hydrological model (TOPMODEL), are based on a fundamental parameter known as the compound topographic index (CTI) and offer a computationally cost-efficient approach to simulate wetland dynamics at global scales. However, there remains a large discrepancy in the implementations of TOPMODEL in land-surface models (LSMs) and thus their performance against observations. This study describes new improvements to TOPMODEL implementation and estimates of global wetland dynamics using the LPJ-wsl (Lund–Potsdam–Jena Wald Schnee und Landschaft version) Dynamic Global Vegetation Model (DGVM) and quantifies uncertainties by comparing three digital elevation model (DEM) products (HYDRO1k, GMTED, and HydroSHEDS) at different spatial resolution and accuracy on simulated inundation dynamics. In addition, we found that calibrating TOPMODEL with a benchmark wetland data set can help to successfully delineate the seasonal and interannual variation of wetlands, as well as improve the spatial distribution of wetlands to be consistent with inventories. The HydroSHEDS DEM, using a river-basin scheme for aggregating the CTI, shows the best accuracy for capturing the spatiotemporal dynamics of wetlands among the three DEM products. The estimate of global wetland potential/maximum is ∼ 10.3 Mkm2 (106 km2), with a mean annual maximum of ∼ 5.17 Mkm2 for 1980–2010. When integrated with wetland methane emission submodule, the uncertainty of global annual CH4 emissions from topography inputs is estimated to be 29.0 Tg yr−1. This study demonstrates the feasibility of TOPMODEL to capture spatial heterogeneity of inundation at a large scale and highlights the significance of correcting maximum wetland extent to improve modeling of interannual variations in wetland area. It additionally highlights the importance of an adequate investigation of topographic indices for simulating global wetlands and shows the opportunity to converge wetland estimates across LSMs by identifying the uncertainty associated with existing wetland products.