Reclaiming histosols in Montéregie region, Québec, Canada, increases peat decomposition and compaction rate and decreases the effectiveness of subsurface drainage. The objective of this paper was to use HYDRUS-2D to model the behavior of subsurface drainage systems, in order to evaluate the compaction effect on drain depth and spacing, and to determine the compact layer thickness and saturated hydraulic conductivities (Ksat) resulting in an improvement of subsurface drainage]. The drainage model was calibrated [Nash-Sutcliffe efficiency coefficient (NSE) = 0.958, percent bias (PBIAS) = −0.57%] using Ksat, meteorological data, and matric potential (h) data measured on the project site from June 10 to July 19, 2017. The calibrated and validated model was used to analyze the variation of h values (Δh in cm d−1) as a function of drain spacing (2–7 m) and drain depth (1 and 1.2 m) and to identify the response surface of Δh to various compact layer thickness and Ksat combinations. The results showed that Δh was on average 58% greater below the compact layer than above it and that reducing drain spacing or increasing drain depth does not improve the drainage rate. The analysis of the compact layer thickness and Ksat effect on Δh showed that for a Δh of 40 cm d−1, Ksat actual values in the two uppermost layers should be multiplied by 50 for compact layer thickness varying from 12 to 35 cm. Water percolation in the soil is reduced by the compact layer. Soil management methods for improving Ksat should therefore be better than deepening the drains or and reducing the spacing.

The objective of this study was to design, using field data, a filter to correct the soil penetration resistance (PR) values by considering all the effects (penetration friction component [PFC], varying penetration rate, and shocks) that bias PR interpretation. The data used were collected in cultivated peatlands of Montérégie (Canada) known for their advanced state of compaction. Penetration resistance (standard measurements) and cone resistance (CR) measurements (successive measurements per 10cm layer) were carried out on four peatlands (five stations per peatland). The filter components, a multi-resolution analysis (AMR), a polynomial model, and a quantification and removal of PFC were identified and were used to design the filter and tested by cross-validation. Our study showed that PFC exists (p value < 5%) and does not depend on the peatland on which the measurements are made. An average filter has therefore been designed for the entire site and its effectiveness has been demonstrated. An application to the data collected in 10 fields on the site made it possible to identify the main soil density profiles existing in Montérégie and to quantify the average PFC characterizing these soils. We noticed that, on average, PFC represents between 18.6 and 41% of the PR value. The combined results of the statistical and physical analyses of the filter made it possible to recommend degrees 2 (D2), 3 (D3), and 5 (D5) of the polynomial model to adjust density profiles of cultivated Histosols when there are no, one, or two compact layers.