Abstract. Methane emissions from natural wetlands constitute the largest methane source at present and depend highly on the climate. In order to investigate the response of methane emissions from natural wetlands to climate variations, a one-dimensional process-based climate-sensitive model to derive methane emissions from natural wetlands is developed. In the model the processes leading to methane emission are simulated within a one-dimensional soil column and the three different transport mechanisms, diffusion, plant-mediated transport, and ebullition, are modeled explicitly. The model forcing consists of daily values of soil temperature, water table, and net primary productivity, and at permafrost sites the thaw depth is included. The methane model is tested using observational data obtained at five wetland sites located in North America, Europe, and Central America, representing a large variety of environmental conditions. It can be shown that in most cases seasonal variations in methane emissions can be explained by the combined effect of changes in soil temperature and the position of the water table. Our results also show that a process-based approach is needed because there is no simple relationship between these controlling factors and methane emissions that applies to a variety of wetland sites. The sensitivity of the model to the choice of key model parameters is tested and further sensitivity tests are performed to demonstrate how methane emissions from wetlands respond to longer-term climate variations.

Abstract. Methane is an important greenhouse gas which contributes about 22% to the present greenhouse effect. Natural wetlands currently constitute the biggest methane source and were the major source in preindustrial times. Wetland emissions depend highly on the climate, i.e., on soil temperature and water table. To investigate the response of methane emissions from natural wetlands to climate variations, a process-based model that derives methane emissions from natural wetlands as a function of soil temperature, water table, and net primary productivity is used. For its application on the global scale, global data sets for all model parameters are generated. In addition, a simple hydrologic model is developed in order to simulate the position of the water table in wetlands. The hydrologic model is tested against data from different wetland sites, and the sensitivity of the hydrologic model to changes in precipitation is examined. The global methanehydrology model constitutes a tool to study temporal and spatial variations in methane emissions from natural wetlands. The model is applied using high-frequency atmospheric forcing fields from ECMWF reanalyses of the period from 1982 to 1993. We calculate global annual methane emissions from wetlands to be 260 Tg yr -•. Twenty-five percent of these methane emissions originate from wetlands north of 30øN. Only 60% of the produced methane is emitted, while the rest is reoxidized. A comparison of zonal integrals of simulated global wetland emissions and results obtained by an inverse modeling approach shows good agreement. In a test with data from two wetlands the seasonality of simulated and observed methane emissions agrees well. [1996]. It calculates presentday global methane emissions from wetlands based on the amount of decomposed organic carbon, water table, and temperature. However, their model has never been tested against time series of methane emission data and has not been applied to temporal variations. In section 2 the methane model and its application on the global scale are described. Section 3 covers global data sets used. The global wetland distribution is prescribed from the data set of Matthews and Fung [1987]. Global data sets of all model parameters, which are soil depth, rooting depth, relative pore space, and efficiency of plant-mediated 34,189

Abstract. Measurements of atmospheric methane from a globally distributed network of air sampling sites indicate that the globally averaged CH4 growth rate increased from an average of 3.9 ppb yr -• during 1995-1997 to 12.7 _+ 0.6 ppb in 1998. The global growth rate then decreased to 2.6 _+ 0.6 ppb during 1999, indicating that the large increase in 1998 was not a return to the larger growth rates observed during the late-1970s and early-1980s. The increased growth rate during 1998 corresponds to an increase in the imbalance between CH4 sources and sinks equal to -24 Tg CH4, the largest perturbation observed in 16 years of measurements. We suggest that wetland and boreal biomass burning sources may have contributed to the anomaly. An adaptation of a global processbased model, which included soil-temperature and precipitation anomalies, was used to calculate emission anomalies of 11.6 Tg CH4 from wetlands north of 30øN and 13 Tg CH4 for tropical wetlands during 1998 compared to average emissions calculated for 1982~ 1993. For 1999, calculated wetland emission anomalies were negative for high northern latitudes and the tropics, contributing to the low growth rate observed in 1999. Also 1998 was a severe fire year in boreal regions where -1.3 x 105 km 2 of forest and peat land burned releasing an estimated 5.7 Tg CH4.

A process-basemd odelh as beend evelopedin order to calculate methane emissions from natural wetlands as a function of the hydrologic and thermal conditions in the soil. The considered processes in the model are methane production, methane consumption and transport of methane by diffusion, ebullition and through plants. The model has been tested against data from a three-year field study from a Michigan peatland. The interannual and seasonal variations of the modelled methane emissions and methane concentration profiles are in good agreement with the observations. During the growing season the main emission pathway proceeds through plants. Ebullition occurs whenever the water table is above the soil surface, while diffusion is only significant in the first 15 days after a drop of the water table below the peat surface