Soil properties, topography and crop productivity across the field

Negative correlations were found between Feox and SOC, SC-C and rSOC in the Bjertorp field. These results suggest that the Feox may not be a key driver for SOC stabilization in the Bjertorp soil, which is in contrast to the commonly reported role of reactive Fe phases in SOC dynamics (Bailey et al., 2019; Rasmussen et al., 2018). Plausible reasons for these findings in the Bjertorp field are that (1) soil pH at Bjertorp may be too high for reactive Fe phases to effectively protect SOC (Wagai and Mayer, 2007; Bailey et al., 2019) and (2) redox cycling occurred in the field, which reduced the reactive Fe phases and decreased their capacity to protect SOC (i.e. adsorbed SOC is desorbed; Hall et al., 2018). The first explanation is supported by a positive correlation between Feox and soil pH, which means that larger Feox contents were associated with smaller adsorption capacity per mass of Feox. The soil pH in the Bjertorp field (mean = 6.0) is also relatively high compared to acidic soils for which the importance of Feox for SOC stabilization is well

established (Kaiser and Guggenberger, 2000). Regarding the second possi-bility, redox cycles in the field were not measured in the present study, but signs of redox cycling (oxidized and reduced soil colors, Figure 2) were noted in the soil survey and profile descriptions. The positive correlation be-tween clay and Feox and the negative correlation bebe-tween Feox and elevation imply that the larger clay content in the low-lying areas could induce par-tially anaerobic conditions within soil aggregates (Keiluweit et al, 2018), which in turn would reduce the reactive Fe phases and prevent Feox from being effective for SOC stabilization (Hall et al., 2018; Inagaki et al., 2020).

This hypothesis was also supported by the results presented in Papers ІІ and ІІІ, which indicated that larger clay contents may lead to waterlogging and anaerobic condition due to larger water holding capacity and smaller near-saturated hydraulic conductivities.

SOC content was negatively but weakly correlated with clay content in the Bjertorp field, which suggested that clay content itself was not a key driver for SOC stabilization. These results are in line with the results from the combined dataset presented in Figure 6 as well as the studies cited in section 3.2.1. Clay (and fine silt) content has been used for modelling SOC dynamics and for estimating the degree of SOC saturation; however, based on the results in this thesis, I suggest that measuring Alox and Feox content as proxies for reactive mineral phases, in addition to clay content, should be useful for these purposes for arable topsoils in humid continental climate.

Finally, the multiple linear regression analysis indicated that carbon input from crop production, which was estimated from mean relative yield ac-counting for 11 years of yield records, explained ca. 20% of the spatial vari-ation in SOC content in the Bjertorp field. Positive relvari-ationships between SOC and yield have been reported in several studies at different scales (e.g.

Kravchenko and Bullock, 2000, Oldfield et al., 2019). It should be noted that this positive correlation can also be caused by the positive effects of SOC on crop growth (Oldfield et al., 2020) and there should actually be interactive feedbacks between SOC content and crop production (Henryson et al., 2018).

The MRY was also negatively correlated with clay content and positively correlated with elevation, which indicated that the lower lying areas may not be as suitable for crop production as the higher lying areas in the Bjertorp field possibly due to waterlogging, which may cause lower C inputs for soils with larger clay contents. This relationship may also explain why SOC was

negatively correlated with clay and Feox contents and positively with eleva-tion, since these variables were highly correlated with one another.

6.3 Pore size distribution quantified by X-ray tomography and soil water retention

6.3.1 SOC and pore size distribution

The relatively strong association between SOC and the abundance of smaller macropores (<720 µm in diameter) found in this study was also noted by Larsbo et al. (2016), Xu et al. (2018), Singh et al., (2020) and Wang et al.

(2021). A plausible mechanism behind this correlation is that the aggregation of clay-sized particles enhanced by SOC as a binding agent (as discussed in section 6.1) may have contributed to the stabilization of these small macropores. It should be noted that this positive correlation can also be caused by fine roots which can create small macropores (Bodner et al., 2014;

Meurer et al., 2020a; Lucas et al., 2021) and, at the same time, be an im-portant source of carbon (Kätterer et al., 2011). However, bioporosities were not correlated with SOC except for the bioporosity in the 480–720 µm class, which constituted a minor fraction of the total X-ray visible porosity (<0.1%

of total visible porosity) and hence may have little influence on soil processes (e.g. biogeochemical cycling and water dynamics).

The porosity in diameter classes larger than 0.2 µm estimated from soil water retention data were positively correlated with SOC content. In partic-ular, a relatively strong positive correlation between SOC and the porosity in the 0.2–5 µm dimeter class was observed. This relationship may be explained by the finding that SOC was the key binding agent for the silt-sized aggre-gation as shown in Figure 8, which may also enhance the formation of pores in the 0.2–5 µm class between or within the silt-sized aggregates. Again, it should be noted that there is a two-way interaction between SOC and soil pore structure (Meurer et al., 2020b), which means that pores in this diameter class can also physically protect SOC from microbial decomposition (Dugaint et al., 2012; Kravchenko and Guber, 2017). Pore diameters of 0.2 to 5 µm are too small for microbes to colonize and SOC may therefore not be accessible to them (Kravchenko et al., 2020).

The results from the Bjertorp field are in line with the analysis of the data reported in Jensen et al. (2020) and Kirchmann et al. (1999), who showed

that the porosities in the 0.2–30 µm and 1–5 µm diameter classes were posi-tively and most strongly correlated with SOC content (Table S2 and S3 in Paper II). Also, a very recent study by Fu et al. (2021) reported a strong pos-itive correlation between SOC and porosity in the 0.2–7.5 µm diameter class compared with other pore size classes. Interestingly, the causes for variations in SOC content vary between these studies and the Bjertorp field. As dis-cussed above, the variation in SOC in this thesis was most likely caused by variations in Alox content and the C input from crop production, whereas the variation in Jensen et al. (2020) was caused by land use changes (bare fallow, grassland and crop land). In Kirchmann et al. (1999), the variation was caused by different organic and inorganic amendments, while the variation in SOC in Fu et al. (2021) was likely caused by different land use and man-agement (pasture and cropland with and without irrigation). On the other hand, Zhou et al. (2020) reported no correlation between SOC and the po-rosity in the 0.2–10 µm diameter class for Vertisols under arable land use as water contents at both field capacity and wilting point were positively corre-lated with SOC content. This discrepancy between the studies may indicate that the relationships between SOC and pore size distribution are dependent on soil type, particularly clay mineralogy, which can determine water ad-sorption capacity in micropores (Libohova et al., 2018; Lehmann et al., 2021).

6.3.2 Texture effects

Porosities in some of the macropore diameter classes were positively corre-lated with clay content. This can be explained by the fact that a larger clay content enhances the formation of soil cracks on drying (Horn et al., 1994;

Paradelo et al., 2016b; Colombi et al., 2021). The positive correlations be-tween clay content and bioporosity may be associated with a larger macro-faunal activity (e.g. earthworm) in loamy soils compared to sandy soils (Capowiez et al., 1998; Baker et al., 1998; Lindahl, et al., 2009).

Soil water retention and pore-size distribution estimated from soil water retention measurements were generally more strongly correlated with clay content than with SOC content. The results of multiple linear regression anal-yses also indicated that more SOC would be required to achieve a given soil porosity for samples with larger clay content compared with soils of smaller clay content. This is in line with previous findings that soils with larger clay

content will require larger SOC content to achieve a similar degree of aggre-gate stability, soil structural quality and clay dispersibility (Feller and Beare, 1997; Soinne et al., 2016; Johannes et al., 2017; Jones et al., 2020; Prout et al., 2021).

Also, I found that the importance of Feox, POM-C and the C:N ratio of SOM and POM (as indicators of SOM quality) for the pore size distribution was less clear. This is because they were correlated with the clay content, and their possible influence on pore sizes cannot be discerned. Nevertheless, in the Bjertorp soils, pore size distribution could be well explained by clay and SOC content rather than other soil properties.