Year: 2023
Type: dissertation
Abstract: Agricultural soils contribute to the increase in greenhouse gas (GHG) concentrations in the atmosphere, among others through the changes soil organic carbon (SOC) stocks. Agricultural mitigation measures that reduce the atmospheric carbon dioxide (CO2) concentration often focus on reversing the soil from a carbon (C) source to a sink, storing C in the soil over longer time periods. However, such measures require verification of the SOC stock changes. Additionally, legislation requires the completion of GHG budgets, which also include estimates of the GHG emissions from soils at regional and national scale. Recently, the emergence of C trading schemes has created additional need for farmers, stakeholders, companies and researchers to know the SOC stock at field scale and how this stock may be enhanced. In 1986, the Danish soil monitoring network, the National Soil Grid (NSG), was established. Four sampling campaigns have been undertaken in the NSG, approximately once per decade (1986, 1997, 2009, 2019) for multiple soil layers (topsoil, 0-25 cm; subsoil, 25-50 cm; deep subsoil, 50-100 cm). The data collected during the sampling campaigns may be used for calculating the SOC stock in Danish mineral soil under agricultural use. Estimates of the SOC stock are affected by the extent and quality of soil data available for the calculation, both in terms of what is known about the soil (and thus what remains unknown) as well as the quality of the data. Bulk density (BD) is widely regarded as a necessary parameter for calculating the SOC stock; however, using a standard BD value rather than a site-specific BD value can result is significant over- or underestimation of the SOC stock. Similarly, disregarding the rock fragment (RF) content when calculating the SOC stock can result in overestimation proportional the RF content. In the context of the NSG, the use of site-specific BD and RF content data affects the mean SOC stock estimate for various groupings of site relatively little, but the effect for individual sites is immense. The inclusion of site-specific BD and RF content data is therefore essential for calculating accurate SOC stocks at local scales, e.g. for carbon trading schemes and field experiments. Using the fixed-depth approach, the median change in SOC stock, assuming the BD is the same in both 2009 and 2019 in the NSG, was -1.0 Mg C ha-1 in the topsoil, 3.0 Mg C ha-1 in the subsoil, for a median change in the combined layer (0-50 cm) of 2.0 Mg C ha-1. While the median changes in SOC stock between 2009 and 2019 are statistically significant, 42% of sites in the NSG gained SOC stock despite the median loss, while 34 % of sites lost SOC stock in the subsoil. As with the soil parameters, the mean or median values may represent the situation in the NSG, but is likely a poor reflection of the situation at a given site in the NSG. Land use is an influential factor on the change in SOC stock, and farm management controls much of the C input on agricultural land. Long-term field experiments are often used to calculate estimates for the effect of various management practices on the SOC stock, but these experiments are subject to their local soil and climatic conditions. The changes in SOC stock in the NSG were analyzed in relation to the known farm management, including crops, cover cropping, ploughing, straw incorporation and organic fertilizer origin, as well as soil parameters as an attempt to verify the results from field experiments at national scale. However, due to the multicollinearity of the data, the statistical analysis is severely limited and the results potentially biased. A large-scale verification of the effects of farm management practices is therefore not feasible with the NSG. An alternative approach to estimating the change in SOC stock due to land use and management is using soil C turnover models. Using data from the German national soil monitoring network as well as data from 23 flower strips, the effect of implementing flower strips in German cropland as a C sequestration method was investigated using RothC. The result of the modelling exercise show that flower strips on average sequester 0.5 Mg C ha-1 year-1, which is comparable to the rate of other initiatives such as use of cover crops. Additionally, the species diversity of the flower strips affects the efficiency of the C sequestration; the C input in flower strips derives solely from biomass production, and the more diverse flower strips generally had a lower total biomass production and are thereby less efficient at transferring C from the atmosphere to the soil. However, the C sequestration potential is only one of multiple potential benefits that implementation of flower strips may serve; additionally, flower strips serve as habitats for insects, both pollinators and pest predators. However, converting productive cropland to flower strips may decrease the total yield. In conclusion, site-specific information is crucial for accurate estimates of SOC stocks at local scales, while standard or average values will suffice for larger-scale averages. However, while average and median values may represent the overall trend for the whole monitoring network, individual sites are likely to differ significantly. The statistical analysis of the effect of farm management on SOC stock at national scale was not feasible, and other approaches to estimating SOC stock dynamics should be considered, such as modelling. Still, modelling approaches rely on the accuracy and representativeness of both experimental data and the known information about the site being modelled, and accurate local information is thus essential regardless of the approach to SOC stock calculation.
Institution: Aarhus University
Language: English
Subfield: Soil Science
Domain: Life Sciences
Open Access status: green
Award: G3813490308
OpenAlex ID: W4412332036