In Mediterranean semiarid conditions, the decomposition and nutrients dynamics were impacted by the crop species (barley, pea and vetch) and the plant part of the residues. At the early stages of residue decomposition, fluctuations in temperature and moisture regimes have a greater influence on mass-loss than litter quality (Berg & Mcclaugherty, 2008). In our study, the general slow decomposition along with low nutrient release from the residues observed during the first 6 weeks may be explained by the lack of rainfall leading to low soil moisture during summer months. In a semiarid dryland, Douglas et al., (1980) pointed out soil moisture as a limiting factor for straw residue decomposition during summer months in a semiarid dryland agroecosystem. However, in our study, from week 9 onwards (late September), residue decomposition and nutrient loss dynamics presented different rates depending on the type of residue (roots and shoots) and the crop species of residue (barley, pea and vetch). This change in the dynamics was attributed to the increase in the soil biological activity. Decomposer organisms are more active when the soil conditions are favorable (Berg & Mcclaugherty, 2008); availability of nutrients contained in the residues plus mild temperatures together with higher soil moisture (September-October) led to a greater activity of soil microorganisms and fauna, thus, accelerating decomposition rates. The higher decomposition rates observed were supported by the increase in the soil C labile content (POXC) available, resulting in an increase of the β-glucosidase activity, key enzyme in the decomposition process (Nevins et al., 2018). In semiarid Mediterranean conditions, Ordóñez-Fernández et al., (2007) estimated a 57% and 91% of mass loss for pea and sunflower, respectively, linked to autumn rains and 35% of wheat mass loss during spring; besides, similarly to our study, unaltered residues were reported during summer months. The litter decomposition process and subsequently, the release of nutrients to the soil, were greatly influenced by the temporal variability, with periods of limiting environmental conditions for decomposition, typical of the Mediterranean semiarid area. In our study conditions, late summer, autumn and spring months were periods with an activation of soil microorganisms resulting in an acceleration of decomposition rates, due to an increase of the soil moisture and temperature, contrarily to winter and summer months.
Shoot and root mass decay patterns and the subsequent nutrient release might be also explained by the level of contact between residues and soil particles. Residues incorporated into the soil have a greater surface area available for microbial degradation, thus accelerating decomposition and nutrient cycling rates (Lupwayi et al., 2004). However, residues left on the soil surface are more exposed to fluctuating environmental conditions as well as a reduced contact with soil decomposers (Lupwayi et al., 2004). This could explain the first rapid drop of root biomass and its subsequent nutrients release followed by a steady period of decomposition compared to the more gradual decomposition of shoot residues. As microbes consume the more readily decomposable material, the chemical content of the residue becomes increasingly more recalcitrant (Johnson et al., 2007). This could explain the similar mass and carbon decay rates observed from week 14 to 48 in roots residues. Thereafter the disappearance of the easily decomposable fraction, the steady decomposition and slow C and N release rates from roots could be linked to the recalcitrant cell wall constituents, which probably remained in the residues (Lupwayi et al., 2004, 2006a).
Residues from different crop species impacted the decomposition rates and its resulting nutrient release to the soil. Several authors observed fast decomposition rates (Lupwayi et al., 2004; Ordóñez-Fernández et al., 2007a; Sievers & Cook, 2018) and nutrients release (Jahanzad et al., 2016; Lupwayi et al., 2006a; Ranells & Wagger, 1996) related to high N contents in legumes residues under field conditions. Thus, in our study, the lower C:N ratios might explain the higher rates of decomposition and nutrients release of legumes compared to barley residues from both plant parts. Lignin content is frequently higher in roots. Thus, the decomposition of the root residues did not seem to have been driven by the lignin content, since legumes root mass presented higher lignin content compared to barley roots. Lupwayi et al., (2004) reported 74.9 % and a 65.3% of mass loss in buried and surface clover residues, respectively, however, only 27.4 % of barley shoots mass decay in 52 weeks of decomposition. Beside, the better fittings observed in the decay models for legumes compared to barley would be also exlained by the faster rates observed. In similar conditions to our study, Ordóñez-Fernández et al., (2007) reported a better fitting for pea residue decomposition than wheat residues using a single exponential decay model. Contrarily to legumes, high C:N ratio of barley residues were associated with slow decomposition rates. Nitrogen is necessary for soil microorganisms in order to synthesize their cellular material and high C dilutes the N concentration in crop residues (Lupwayi et al., 2006a), reducing its decomposition rates. Although, in similar conditions to our study, Douglas et al., (1980) indicated that wheat straw decomposition was well described as a linear function over time associated to slow decomposition rates. However, in our study, both models proposed poorly explained the barley residue decomposition, probably due to its slow rate.
In both plant parts, the general steady concentration of C in the residues over the experimental period derivated in a similar C loss pattern to the observed in the mass decay. Regarding the N release, vetch residues released a major amount of N as NO3- and NH4+, especially in the shoot residues, not only for the major N content in its biomass, but also due to their lower recalcitrant fraction (acid and neutral fibers, cellulose and hemicellulose) in comparison to pea residues.Lupwayi et al., (2006a) estimated a 64 % N release in zero tllage pea residues, a similar value to our 49 and 65% of N release in pea and vetch surface residues, respectively. However, our results indicated that barley residues N release to the soil was extremely low. Residues characterized by low N contents or a high C:N ratios are expected to result in microbial N immobilization (Johnson et al., 2007) due to a higher competition for the available N by microorganisms, consequently decreasing the decomposition and its nutrient release (Kumar & Goh, 1999).Sievers & Cook, (2018) concluded that cereal rye residues may immobilize N because of its high C:N ratio and decompose at a slow rate. Also,Lupwayi et al., (2006b) reported an increase of N remaining in wheat residues under zero tillage that was attributed to microbial N immobilization. Therefore, the N microbial immobilization may explain the N concentration increase of the barley residues. The limited N available in the barley residues and the high demand of N for decomposition could explain the slower rates and its subsequent low nutrient release compared to vetch and pea residues observed.
In our study, the cropping system influenced the residues decomposition and thus the nutrients release to the soil. Independently of the cropping system, all barley shoot residues followed the same decomposition trend. Maintaining a soil cover of lasting long residues, such as a barley shoot cover, can be interesting for preserving water and soil resources Mediterranean agroecosystems (López et al., 2005). However, it has been demonstrated that a major part of legume residues decomposed during autumn months in Mediterranean semiarid conditions, leading to a high N-release from these residues to the soil. High N availability in the soil can be a good strategy not only for increasing the soil fertility but also for reducing the external N-input (Arrúe et al., 2019) at the sowing (autumn months). Reducing or even suppressing the fertilizer application may enhance the sustainability of the acroecosystems. The success of legume crops to be a source of N in the soil, mainly depends on its adaptability to water-stress conditions and its water use efficiency (Álvaro-Fuentes et al., 2009). In our study, vetch residues were observed to release more N compared to pea, demonstrating that this crop might be a suitable candidate to be in rotation with winter cereals under Mediterranean semiarid conditions.