A statistical approach to determining the uncertainty of peat thickness

This paper presents statistical studies of peat thickness to define its expected maximum variation ( dm( r)) as a function of separation distance Δr. The aim was to provide an estimate of the observational uncertainty in peat depth due to positioning error, and the prediction uncertainty of the computed model. The data were GPS position and ground penetrating radar depth measurements of six mires in different parts of Finland. The calculated observational uncertainty for Finnish mires in general caused, for example, by a 20 m positioning error, is 43 cm in depth with 95 % confidence. The peat depth statistics differed among the six mires, and it is recommended that the mire specific function dm( r) is defined for each individual mire to obtain the best estimate of observational uncertainty. Knowledge of the observational error and function dm( r) should be used in peat depth modelling for defining the uncertainty of depth predictions.