Filling holes in regional carbon budgets: Predicting peat depth in a north temperate lake district

Abstract

[1] Peat deposits contain on the order of 1/6 of the Earth's terrestrial fixed carbon (C), but uncertainty in peat depth precludes precise estimates of peat C storage. To assess peat C in the Northern Highlands Lake District (NHLD), a $7000 km 2 region in northern Wisconsin, United States, with 20% peatland by area, we sampled 21 peatlands. In each peatland, peat depth (including basal organic lake sediment, where present) was measured on a grid and interpolated to calculate mean depth. Our study addressed three questions: (1) How spatially variable is peat depth? (2) To what degree can mean peat depth be predicted from other field measurements (water chemistry, water table depth, vegetation cover, slope) and/or remotely sensed spatial data? (3) How much C is stored in NHLD peatlands? Site mean peat depth ranged from 0.1 to 5.1 m. Most of the peatlands had been formed by the in-filling of small lake basins (terrestrialization), and depths up to 15 m were observed. Mean peat depth for small peat basins could be best predicted from basin edge slope at the peatland/upland interface, either measured in the field or calculated from digital elevation (DEM) data (Adj. R 2 = 0.70). Upscaling using the DEM-based regression gave a regional mean peat depth of 2.1 ± 0.2 m (including $0.1-0.4 m of organic lake sediment) and 144 ± 21 Tg-C in total. As DEM data are widely available, this technique has the potential to improve C storage estimates in regions with peatlands formed primarily by terrestrialization.

Uncertainty in Peat Volume and Soil Carbon Estimated Using Ground‐Penetrating Radar and Probing

“…Because the quasi-random resampling of the probe data resolved essentially the same subsurface features as both the full set of gridded probe data and the GPR data, it stands to reason that in this case the structure of the sampUng strategy did not significantly affect the result. The basin model results are similar to three-dimensional models of other peatlands presented in studies using similar or lower probe sampling densities (e.g., Buffam et al, 2010;Lowry et al, 2009). The enhanced value of GPR data becomes apparent when considering the possibility of extracting information about the physical composition of the peat, as described below.…”

“…Even though the probe method may encounter problems with misidentification of the basal sediment depth (Rosa et al, 2009), the potential to assess the uncertainty by making repeated measurements is valuable. Nonetheless, our results clearly indicate that an appropriate sampling density must be used to ensure that the uncertainty estimate in volumetric calculations is acceptably small or the peatland must be known to have a simplistic basin geometry that can be resolved with fewer probe locations (e.g., Buffam et al, 2010). Avoiding excessive probe measurements by using GPR may be desired if the peatland will be studied over time.…”

“…The size ofthe subset was based on the approximate number of probe measurements that might be more representative of a typical investigation to characterize the morphology of a peat basin (e.g., n = 44 for an 8.7-ha basin, 0.05 per 100 m^ [Buffam et al, 2010]). Although a regular grid would normally be used in the field for probe surveys (e.g., Buffam et al, 2010), the purpose of this simulation exercise was primarily to characterize the uncertainty related to using fewer sample locations. The random subsampling approach used in the Monte Carlo simulation is appropriate because it yields the maximum level of uncertainty due to the potential spatial clustering in the subsampled groups.…”

Time-Dependent Impacts of Cattail Invasion in a Great Lakes Coastal Wetland Complex

“…Even small hydrological changes can further alter habitat (Desgranges et al 2006) and ecosystem services (see Fisher and Acreman 2004). Sites invaded by Typha for more than 55 years are likely to be drier as organic matter builds up, "terrestrializing" invaded sites (see Kirschner et al 2001;Rooth et al 2003, andBuffam et al 2010). Habitat quality and ecosystem services such as denitrification are likely to decline as the water levels they depend on become lower and wetlands become less wet.…”

Estimates of evapotranspiration from contrasting Wisconsin peatlands based on diel water table oscillations

“…In that study, peatland ET was estimated using an eddy covariance model; and results indicated that rates averaged 2.2 to 3.7 mm/d (maximum 5 mm/d) and that the ratio of P to ET varied from 1.02 to 1.34 over the time period May to September, averaging 1.21 across 5 years. Satellite photos from GoogleMaps and contour maps of peat thickness from Buffam et al (2010). Trout Bog (left) and Crystal Bog (right) showing satellite photos (top) and thickness of peat (bottom).…”