HIDRÓGENO ECONÓMICO USABLE ENERGÉTICAMENTE (14.07.2014)

This chapter presents both the spatial and temporal variability of (near-) saturated and unsaturated K obtained from HI-and HYPROP®_{-experiments from May-16 to Apr-17.}

Results are discussed with respect to Hypothesis D and E pertaining to Objective 2) as outlined in Section 1.5.

4.3.1 (Near-) saturated hydraulic conductivity

The tillage treatment had a distinct influence on the overall topsoil K obtained from hood infiltration runs with clear temporal patterns (Table 4). Chopped straw and stubbles from the previous harvest mixed into the top 15 and 30 cm on RT and CT, respectively, had a distinct effect on K at and near saturation. The decomposing organic material likely led to the

formation of macropores (Strudley et al., 2008) which in turn resulted in an increase in K during the first three occasions from May-16 to Aug-16. The increase in mean pore radius due to wetting-drying cycles (Bodner et al., 2013b, 2013a) may have further promoted a re-

formation of soil structure as already discussed in Section 4.1.1. In this section, it was shown that the pore volume fraction of transmission pores (⌀ 50 – 500 μm) and fissures

(⌀ > 500 μm) increased on those plots during that time. Transmission pores were also shown to be correlated to changes in Kslab. This indicated that these pore size classes are important factors for infiltration processes at saturation, especially on CT. The increases in K were

stronger on RT, which may be explained by the relatively higher availability of organic material in the top 15 cm compared to CT. On CT, a similar amount of material was distributed over a larger vertical profile of 30 cm caused by the moldboard plowing as also observed in consistently lower SOC stocks in the top 5 cm compared to RT (Figure 15). Variability in May-16 was high on CT and RT but then decreased with time (Figure 8). This may point to a continuing homogenization of the soil where rather the loose soil matrix influences K than individual large pores that may be subject to a more heterogeneous distribution and therefore exhibit a more variable infiltration behavior.

Stubble breaking post-harvest on RT (Table 1) interrupted the rise in K by cutting through the pore network that had developed up to that point. Due to that, pore continuity, an important metric for K near saturation (Soracco et al., 2019), was reduced. On the other hand, in the conventionally tilled soil with comparingly less ‘building material’ for the soil aggregates, the stubble breaking lead to a further increase in K by loosening the soil matrix as seen in decreased ρb (Figure 14). On NT, K near saturation was comparably lower (Table 4) which is probably associated to the overall increased ρb (Figure 14). Along with a distinct lack of fissures and transmission pores compared to tilled soil (Weninger et al., 2019; Figure 10), this limited infiltration through the soil matrix. In their study on loam and silt loam soils, Soracco et al. (2019) also reported reduced (near-) saturated K on NT compared to CT. While the continuity of big macropores or biopores (⌀ > 1 mm) expressed in a continuity index (Lozano et al., 2013) was higher on untilled soil, porosities of those biopores were lower. In part, this could also be seen in reduced fissures on NT (Figure 10), although the bigger biopores were probably not captured entirely by the evaporation experiments. The higher continuity was attributed to a well-developed soil structure, biological activity and abundant root channels on NT.

Temporal variation on NT was comparably low as also observed by Schwen et al. (2011b) and Keskinen et al. (2019). Unlike the tilled plots, organic material from the previous harvest remained on the soil surface acting as a moderator for soil moisture and temperature reducing the effects of wetting-drying cycles and the destructive impact of the kinetic energy received from heavy rainfall (see also Discussion-Section 4.1.1). Overall, ρb was higher which prevented a preferred infiltration through the soil matrix. Only following seedbed preparation for sugar beets in Apr-17, comparable levels to those of CT and RT were observed.

Nevertheless, ρb remained high indicating little influence of this mechanical manipulation on the spatial arrangement of the soil aggregates on NT. The effect of seedbed preparation was

also limited as it only affected the top 8 cm (Table 1) while the comparably denser layer below (Jacobs et al., 2015) introduced a resistance towards infiltrating water.

Many studies looked into the effects of conventional and conservation agriculture on soil hydraulic properties with rather ambiguous results especially when it comes to K (Blanco- Canqui and Ruis, 2018). Weninger et al. (2019) analyzed the overall differences in (near-) saturated K on this field as well as on two sites in Austria (both Chernozem with silt loam texture). Like Blanco-Canqui and Ruis (2018), they did not find systematic differences

between tillage treatments. However, NT exhibited a tendency for reduced K. As discussed in the introduction, temporal variation might explain part of the disagreement between studies, especially if only ‘snapshot’ measurements were done (Strudley et al., 2008). A good example is

Ks in May-16, where overall values were rather similar with 459, 315 and 511 cm d-1 for CT, RT and NT, respectively. This was also true for associated CVs with 146, 177 and 127 %, respectively. Less than one month later with pesticide spraying as the only management action in between (Table 1), Ks was significantly increased on RT compared to NT. On CT, Ks was now greater than on NT. Schwen et al. (2011a) significantly improved soil water modeling by using time-variable hydraulic parameters despite much lower variations in topsoil (near-) saturated K than those observed here. This shows that one-off estimations of this hydraulic property are not sufficient to determine differences in soil hydraulic properties of

conventional and conservation agriculture, and variability in (near-) saturated K needs to be included in the modeling process.

4.3.2 Unsaturated hydraulic conductivity

Under drier conditions, i.e. more negative h, overall variability of K decreased due to

decreasing influence of larger pores emphasizing the increasing relevance of soil texture over soil structure (Bodner et al., 2013b; Schwen et al., 2011b) Nevertheless, a significant temporal variation of kpF2.0 could be seen on CT which may also be an artifact of the parameterization with the bimodal Mualem model as with increases in Ks and especially K-2cm (see also positive correlation in Figure 13), kpF2.0 decreased and vice versa. However, the same pattern was not present on RT and NT. Given the generally low values of kpF2.0, kpF2.5 and kpF3.0, the temporal variability may not be meaningful for modeling studies and it may be reasonable to assume an invariant K in this h-range.

Laboratory data in the unsaturated moisture range in combination with the field measurements representing the soil structural part is essential for an adequate description of the HCC (Weninger et al., 2018). These results showed that there is a difference between field

and lab data from three to almost five orders of magnitude. For the simulation of unsaturated flow processes this data combination is essential in the parametrization of the respective hydraulic models (Schwärzel and Bohl, 2003). Consideration of only Ks in the parametrization of the HCC as many modeling studies do may lead to overestimations of K at h between -10 and -100 cm (De Pue et al., 2019).

4.3.3 Summary: Objective 2) Hypothesis D and E

Spatial variability across the HCC, as expressed in the GCVs, was shown to be reduced with increasing h. This was mostly true for NT and on some occasions for CT and RT. Therefore, Hypothesis D can partly be confirmed. Temporal variability of the HCC was comparably high on plots under tillage (CT and RT), especially at and near saturation, where GCVs differed greatly between occasions. Structural changes throughout the season seemed to occur long after primary tillage and initial void closure. Abundant fissures and transmission pores under CT and NT were linked to changes in (near-) saturated K. With its denser soil matrix, i.e. higher ρb, K near saturation was temporally more stable on untilled soil and Hypothesis E can be confirmed.

4.4 Factors influencing water transmission and its temporal