Search

Abstract Peat and other organic soils (e.g., organo‐mineral soils) show distinctive volume changes through desiccation and wetting. Important processes behind volume changes are shrinkage and swelling. There is a long history of studies on shrinkage which were conducted under different schemes for soil descriptions, nomenclatures and parameters, measurement approaches, and terminologies. To date, these studies have not been harmonized in order to compare or predict shrinkage from different soil properties, for example, bulk density or substrate composition. This, however, is necessary to prevent biases in the determination of volume‐based soil properties or for the interpretation of elevation measurements in peatlands, in order to predict carbon dioxide emissions or uptake caused by microbial decomposition or peat formation. This study gives a comprehensive overview of shrinkage studies carried out in the last 100 years. Terminology and approaches are systematically classified. In part I, the concepts for shrinkage characteristics, measurement methods, and model approaches are summarized. Part II is a meta‐analysis of shrinkage studies on peat and other organic soils amended by own measurement data obtained by a three‐dimensional structured light scanner. The results show that maximum shrinkage has a wide range from 11% to 93% and is strongly affected by common soil properties (botanical composition, degree of decomposition, soil organic carbon, and bulk density). Showing a stronger correlation, bulk density was a better predictor than soil organic carbon, but maximum shrinkage showed a large spread over all types of peat and other organic soils and ranges of bulk density and soil organic carbon.

Under natural conditions, peatlands store large amounts of soil organic carbon (SOC). However, they are under threat due to drainage which leads to mineralisation of soil organic matter to carbon dioxide (CO2). This situation is especially severe in Germany, where more than 70 % of peat and other organic soils are used for agriculture. This study assessed the properties of these soils within the framework of the first German Agricultural Soil Inventory. In a nationwide 8 × 8 km grid, soils from a total of 3104 sites were sampled to depths of up to one metre or down to the peat base. Of these sites, 146 were on peat and other organic soils; and 31 % of the 146 sites were being affected not only by drainage but also by changes in horizonation (e.g. mineral covers, deep ploughing). The classification of heavily disturbed sites is limited within the German Manual of Soil Mapping, which has led to the development of an adapted classification scheme for peat and other organic soils under agricultural use in Germany. The respective peat classes showed distinct patterns of SOC and total nitrogen (Nt) contents and stocks, bulk density (BD) and C:N ratios. Overall, a SOC stock of 529 ± 201 t ha-1 and a Nt stock of 29.3 ± 13.9 t ha-1 were found within a depth of 0–100 cm. However, in deeper profiles, 48 % of the total SOC was stored below 100 cm depth down to the peat base. High SOC stocks were also found in peat-derived, mineral-covered and deep-ploughed organic soils, which might be classified as mineral soils depending on the classification system used but are still prone to mineralisation and need to be considered in terms of emissions reporting and mitigation. Logarithmic and quadratic pedotransfer functions were developed to estimate BD and SOC density, respectively, from SOC contents. This is necessary for the calculation of SOC stocks when analyses of BD are absent. The quadratic relationship between SOC content and SOC density clearly showed that heavily degraded organic soils store as much SOC in a defined volume as more natural ones, and that any estimates of differences in potential CO2 emissions should not be based on SOC content, but on SOC density instead.

Capacitance and time domain reflectometry (TDR) probes are frequently used for measurements of the volumetric soil water content. The measurement concept is based on the correlation between volumetric water content and dielectric permittivity (ε). While considerable effort has been made to accurately measure ε in the typical range of mineral soils (40), typical for highly porous media like organic soils. We evaluated the capability of two moisture probe types (TRIME-PICO 64 and GS3) to measure ε across the range from 1 to 80. In the case of the TRIME probes, different equations to calculate ε from transit times were tested. Measuring in a set of reference solutions, the TRIME probes had an RMSE of 18.73 for ε values derived using the manufacturer’s calibration. With a new calibration, the RMSE was decreased to 3.55. The GS3 probes had an RMSE of 3.96. For both probes, uncertainties increased with increasing ε. We also tested the performance for different electrical conductivities of the reference solutions. Accuracy of ε values was unaffected by increasing conductivities for the TRIME probes but decreased for the GS3 probes. The GS3 probes, however, were able to determine electric conductivities accurately, while TRIME probes failed for electrical conductivity although indicated differently by the manufacturer.