Simulation of water table level and peat temperatures in boreal peatlands

Hydrology controls the physical, chemical and biological processes in peatlands and hence could be the most important process regulating their function, development and characteristic biogeochemistry. Models describing hydrological processes and soil heat exchange phenomena are an important tool in understanding the peatland biogeochemical cycles of C and N. Presented in this paper is a peat soil climate model that uses weather data, mire site and peat characteristics as input data. The model is related to the heterogeneous peat characteristics (e.g. bulk density, degree of humification, remains of Sphagnum, Carex, Eriophorum and woody tissues), which in turn vary with depth. Evapotranspiration is related to weather parameters and tree stand characteristics. The model simulations were validated against the observed data collected during 1993 and 1994 growing seasons at a fen and bog sites at the Lakkasuo mire complex located in central Finland. The observed and simulated day-to-day variations in water table levels and soil temperatures during both seasons appeared overall to be in phase. As the model lacks the data needed to characterize the peat matrix hysteresis, the model was found lacking in response to wetting and drying cycles. The mean differences between simulated and observed water table levels during 1994 were -0.3 ± 1.3 and -0.5 ± 2.3 cm for the fen and bog sites, respectively. The model was found to be sensitive to mire surface characteristics and evapotranspiration, particularly for the bog site with an unsaturated zone deeper than at the fen. The absolute mean differences between the simulated and measured peat temperatures from 5 to 150 cm were less than 1.0°C with a maximal standard deviation of 1.6°C. The model deviations for the upper layers showed larger variations compared to deeper layers, implying a greater accuracy in defining the lower boundary of the thermal regime within the peat column.