Assessment of the variability and uncertainty of soil organic carbon inventories in heterogeneous arid and alpine environments


\n
\nSurface soils, forming the largest pool of terrestrial organic carbon, may be able to sequester atmospheric carbon and thus mitigate climate change. So far the soil organic carbon (SOC) literature is dominated by studies in humid, agricultural environments and limited attention has been given to arid and mountain ecosystems that are highly sensitive to environmental change. Thus, our knowledge on the feedbacks between spatial patterns of SOC stocks and temporally and spatially changing environmental conditions (such as land use and climate) in these ecosystems remains insufficient. Analyzing these feedbacks is a major challenge due to the large spatial variability that is caused by the high activity of geomorphic processes in arid and mountain ecosystems. 
\n
\nDue to the increasing interest in reliable estimates of SOC stocks in various environments, this thesis intends to improve our understanding of the linkages between environmental variability and the uncertainty of SOC stock assessments in dynamic geomorphic systems. These uncertainty estimates are expected to contribute to the development of an efficient sampling design with guidelines for the compilation of SOC inventories in heterogeneous environments.
\n
\nThis PhD focuses on three case studies, i) Sede Boquer in the arid Northern Negev desert (Israel), ii) the Kananaskis country in the Canadian Rocky Mountains and iii) the area between the Kleine Scheidegg and Grindelwald in the Swiss Alps. Each study site is characterized by a high geomorphic activity. Based on SOC stocks, which were established for each study site, the main objective of this thesis is to determine the uncertainty associated with SOC assessments that are mainly linked i) to the high spatial variability of the soil forming factors and soil properties, ii) to analytical errors during the measurements of the soil properties, and iii) to uncertainties that arise from the spatial interpolation of local point data with different local spatial interpolation techniques. 
\n
\nThe first case study aimed to identify the relationship between surface characteristics, vegetation coverage, SOC concentration and stocks in the arid northern Negev in Israel. To identify controlling factors of SOC stocks on rocky desert slopes, we compared soil properties, vegetation coverage, SOC concentration and stocks between ecohydrological units. The results show a large spatial variability of SOC, soil bulk density and soil thickness which is mainly attributed to the disconnectivity of overland flows and the local deposition of fine sediments. The calculated SOC stocks indicate that rocky desert slopes represent a significant amount of SOC of soil-covered areas of 1,54 kg C m^2, with an average SOC stock over the entire study area of 0.58 kg C m^2. The spatial variability within the study site is dependent on differences in eco-climate, microtopography, surface processes, soil formation and properties, and vegetation. These differences were mapped within the study site in terms of ecohydrological units, which provide an effective tool to detect spatial patterns and thus to reduce uncertainties of SOC stocks in arid environments. Furthermore, the results indicate that microscale water supply and NPP are the limiting conditions for the formation of SOC in arid, rocky deserts and thus suggest a high sensitivity to potential climate changes. Even though SOC stocks are smaller than in more humid environments, it is of major importance for the functioning and thus conservation of arid ecosystem.
\n
\nMountain environments are heterogeneous and dynamic geomorphic environments that are highly sensitive to land use and climate change. Local geomorphic processes, which are driven by strong topographic gradients, cause a large heterogeneity of the parent material that represent a major challenge in the assessment of SOC stocks in mountain environments.
\n
\nThe first mountain case study is located in the Front Range of the Canadian Rocky Mountains, which is characterized by a very low human impact and a natural boreal forest cover. The second mountain case study, located between the Kleine Scheidegg and Grindelwald (Swiss Alps), is characterized by a long history of agricultural land use. Uncertainties in SOC stocks due to analytical errors and spatial variability of SOC stocks are assessed using a nested sampling design in combination with Gaussian error propagation and Taylor series expansion along several transects that are equally spaced in each study site. Additionally, in Grindelwald the ability of different spatial interpolation methods to cope with data of high spatial variability was tested. 
\nSOC stocks for the upper 30 cm of the mineral soil in Kananaskis and Grindelwald ranged from 3.01 to 24.94 kg C m^-2 (with a mean of 6.40 kg C m^-2) and from 2.52 to 23.46 kg C m^-2 (mean = 8.93 kg C m^-2), respectively. Both studies confirm that multiple regression analysis and ANOVA explain only parts of the SOC variability and that the largest uncertainty is introduced through the large variability of the coarse fraction. Therefore, mountain geomorphic processes, which dominantly control the grain size of the parent material, are responsible for the large uncertainty of SOC stocks in mountain environments. It is thus argued that detailed geomorphological maps, which represent the grains size of the parent material, have a high potential to reduce the uncertainty that is associated with the coarse fraction. Additionally, both studies confirm that stratified nested sampling designs, as applied in this study, are helpful to discriminate the sources of uncertainty and to identify the relevant scales of spatial variability.
\n
\nBased on the results of the three case studies, general guidelines were derived for the compilation of SOC stocks in arid and alpine environments. These guidelines have a strong focus on the assessment on the quantity and quality of SOC stocks in geomorphic active ecosystems.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.