cctreat/HPM-Arctic: HPM-Arctic

HPM-Arctic version integrates two earlier models: The Holocene Peat Model (HPM), a coupled carbon-hydrologic model for peatlands [Frolking et al., 2010], the Geophysical Institute Permafrost Lab soil thermal model GIPL2 [Marchenko et al., 2008]. Briefly, HPM simulates the development of a peat profile over millennia, from initiation, using an annual litter cohort approach so that results can be compared to dated peat cores. Rates of peat accumulation and decomposition are a function of plant community composition (litter quality), modified by dynamic environmental conditions, including water table level and water content in the unsaturated zone, and temperature profiles. Plant community composition (i.e., relative litter inputs from different plant functional types) is dynamic, responding to mean growing season water table depth and peat depth as a proxy for nutrient status. Annual net primary productivity (NPP) is set equal to annual litterfall, the carbon input for peat accumulation. NPP temperature sensitivity was modeled as a Q10 function, with a Q10 value of 1.8, based on an empirical relationship between mean annual air temperatures and above-ground net primary productivity for mosses, vascular plants, and trees that was developed for a transect of peatland sites in boreal Manitoba, Canada [Camill et al., 2001]. Peat bulk density in HPM is computed for each annual litter/peat cohort, and increases non-linearly from a minimum to a maximum value (50 – 130 kg m-3) as the cohort loses mass through decomposition (Frolking et al. 2010). The water table level is calculated from a simple water balance model (precipitation minus evapotranspiration plus net run-on/run-off, and the net peat water content determines the water table location, where the peat water content in the unsaturated zone is a function of peat bulk density and distance above the water table (Frolking et al. 2010)
HPM-Arctic has been modified from the original version of HPM in several ways. Principally, it has been coupled to the Geophysical Institute Permafrost Lab soil thermal model GIPL2 [Marchenko et al., 2008], modified to include an accumulating peat layer on the soil surface [GIPL-2-peat; Treat et al., 2013; Wisser et al., 2011]. GIPL-2-peat is a soil thermal model that solves vertical soil heat transfer and phase change using a numerical approximation accounting for soil type and soil water content; it is driven by air temperature and includes a dynamic winter snowpack as a heat transfer layer. Soil temperatures are calculated for a 100-m soil and bedrock column that has varying thermal properties with depth and variable layer thicknesses, thinner at the surface (0.05 m minimum) and thicker down the soil column into the bedrock (5 m maximum). As peat accumulates on the soil surface over centuries to millennia (at rates generally < 0.001 m yr-1), the mineral soil and bedrock layers slowly descend deeper into the simulated soil profile. Simulated soil temperatures are used to constrain rates of peat decomposition, and variation in the active layer thickness is used instead of peat depth as an indication of nutrient status, which impacts net primary productivity of vascular plants. Active layer thickness, updated annually, is determined by identifying the soil thermal layer just above the top-most layer where the temperature remains below 0° C for two years continuously, in accordance to the definition of permafrost [S A Harris et al., 1988]. In addition to soil temperature profiles, HPM-Arctic has (i) reduced the model time step from annual to monthly for model drivers (air temperature and precipitation), peat profile water balance, and decomposition (the soil thermal model operates at a daily timestep, using air temperatures interpolated from the monthly values); (ii) reduced the number of plant functional types to three: moss, herbaceous (including sedges and graminoids), and ligneous (woody species including shrubs), where for the vascular plants, the inputs and decomposition of above and belowground litter are tracked separately; and (iii) incorporated a simple 'old-new' carbon tracking algorithm, whereby after a specified year all moss, sedge, and shrub plant litter gets labeled as 'new', so that its accumulation as peat and loss through decomposition can be tracked separately from the older peat derived from plant litter inputs prior to the specified year. This study used 2015 CE as both present-day and the boundary between new and old carbon inputs.

Citation: Treat, Claire C., M.C. Jones, J. Alder, A.B.K. Sannel, P. Camill, S. Frolking (in review/2021). "Predicted Vulnerability of Carbon in Permafrost Peatlands with Future Climate Change and Permafrost Thaw in Western Canada", J. Geophysical Research Biogeosciences.