Nitrogen (N) is one of the most important elements in all agricultural ecosystems and significantly influences ecosystem functioning. With the rising costs of synthetic (mineral) fertilizers, cover crops are increasingly used as a source of nutrients, namely N. In Mediterranean environment, cover crops were efficient as N source for maize crop [1]. Integration of cover crops in cash crops can regulate net N mineralization and reduce NO₃^- leaching losses [2]. Understanding patterns of N mineralization and its kinetics is essential to improve N use efficiency in agriculture [3], especially when it comes to meet N demands of crops. Nitrogen mineralization and nitrification produce NH₄⁺ and NO₃^-, which are taken up by plants. These processes also control the supply and magnitude of mineral N from the soil system to the plant [4]. Improving timing, between N demand and supply, is a way to increase and improve N use efficiency in agricultural systems, which is reflected in an increasing of the productivity and environmental quality improvement, because much N is lost through leaching and gas emissions, leading to eutrophication and contributing to climate change [5].

To assess mineralized N, under field conditions, without any profound alteration of the soil and climatic conditions, in situ evaluations are used. In situ field experiments are more labour intensive but provide more realistic estimates of soil N mineralization under environmental conditions and take into account the effects of soil architecture [6, 7]. In this method, mineralized N is estimated by increasing mineral N values in soil samples, incubated under field conditions [8]. Compared with other methods for evaluating mineralized N, in situ incubation has the main advantages of being simple to use, allowing measurements in environmental conditions (for long periods of time) [9], and allowing more reality-like estimates, as soil N dynamics is strongly affected by environmental conditions [10, 11].

Several techniques have been proposed to quantify N mineralization and/or immobilization through in situ incubations [8, 9] such as core sampling, resin-core sampling and buried bags [7]. All these techniques have the disadvantage of underestimating the total mineralized N, as they lack to consider plant presence. Plant presence increases mineralization, due to the release of root exudates, rich in organic N, which promote microbial activity, and contrariwise, reduces immobilized N, due to the competition they establish with microorganisms for N in the soil. However, Monaco et al. [8], consider the buried bags methodology as a good indicator of N mineralization in the soil. In this method, usually called “in situ”, the soil samples were sealed in a sterile polyethylene bag and then reinserted to their respective depth [12]. Thus, the in situ method previously referred to is easy to carry out, sensitive to on-site temperature fluctuations [13], and therefore the results obtained are closer to the real condition.

Net N mineralization (NNM) of soil organic matter is the result of two simultaneous and opposite processes (N mineralization and immobilization), however, in the fields, it also includes processes such as gas losses, leaching, atmospheric deposition and ammonium dynamics (fixation / release) [7]. Clivot et al. [6] highlighted that improved assessment and prediction of NNM rates under field conditions are essential for better management of N in arable cropping systems and improve fertilizer recommendations for farmers because the actual availability of inorganic N depends on the rate of NNM and its transport through the soil.

Previous studies reported that the addition of inorganic N into the soil, derived from cover crop decomposition, by subsequent plant cultivation is important [14] and is recommended as a N source for crop production [1]. Therefore, understanding the impact of different forage legumes on N content and yield of the main crop is important to improve the efficiency of N fertilization in annual crop rotations. Also, the sowing date of cover crops condition, their development and the impact on nutrient accumulation are key factors for N fertilization management [1, 15]. Thus, the delayed sowing of cover crops could affect the carbon:nitrogen (C:N) ratio, which influences the rate of N mineralization [15].

Optimizing N rates for agronomic and environmental considerations continues to receive much attention because N is commonly the most limiting nutrient for agricultural production, representing a large input cost. Losses to the environment usually increase when N is applied in excess [16]. Hence, more studies are needed to obtain knowledge about the N fertilizer value of the residues of cover crops under specific soil and climate conditions.

Synchronization of cover crop decomposition and nutrient release with cash crop uptake can provide benefits to agroecosystems but can be difficult to implement. Thus, in response to the above consideration, we have proposed the following hypotheses: (H1) Net N mineralization would be higher from leguminous than non-leguminous CCs; (H2) leguminous cover crops would be able to supply the N to the cash crop synchronously with its needs. To test these hypotheses, we have conducted a 98-day incubation to compare the N mineralization of leguminous and non-leguminous cover crops.

Our study aimed to evaluate the effects of the winter cover crop (leguminous and non-leguminous), under Mediterranean field in the soil NNM, by measuring in situ NNM incubation in the upper soil (0-150 mm) during maize growing season and by quantification of mineralized N deposited in the soil at the end of incubation.

The quantification of mineralized N in the soil is important for both economic and environmental issues [1, 23]. In this study, both sowing date and residues interfered with mineralized N, although this interference was canceled out with time advancement, after the date of soil residues incorporation into the soil. Nakhone & Tabatabai [22] reported species used as residue conditions for the mineralization of N. Previous studies [1, 23] reported that environmental conditions, such as temperature and humidity, influence the rate of N mineralization. In this study, incubation periods where immobilization ensued coincided with the occurrence of higher temperatures and with the lowest water content in the soil. In addition, Schomberg et al. [24] observed a lower mineralization rate in the presence of higher temperatures. Also, Carsky et al. [25] stated that, with this method, different temperatures can lead to different values of mineralized N.

The soil N contents observed in this study were higher in the treatments with legume residues. Rodrigues et al. [26], reported that legume residues increase the mineral N content in the soil. These values increased up to 70 days after the incorporation of the residues, and then, progressively decreased. This decrease can be explained by two reasons: (i) the start of irrigation (which started about 1.5 months after maize sowing), and (ii) the greater uptake of N by the main crop, since during the first 6-7 weeks, the plants absorbs between 10 to 20% of the total N present at harvest [27]. The explanation previously stated is validated by the values of the accumulated net mineralization, which is always increasing on most incubation dates. The immobilization observed in this study, between day 70 and 83 after incorporation of cover crops residues, could be related to the lack of water before the start of irrigation and, on the other hand, to excess moisture in the soil in the first days of watering, with the consequent occurrence of leaching or denitrification, due to the decrease in aeration, as well as, coincidence with the period when the highest temperatures were registered. Soil temperature and moisture are the environmental factors that most influence the mineralization rate [9].

The values of NH₄⁺ ranged between 1.26 and 9.96 mg N kg^-1 in this study. Similar values of NH₄⁺ were found by Dias et al. [20], when they evaluated, under field conditions, grasses, legumes and spontaneous vegetation, varying the values of NH₄⁺ from 3 to 9 mg N kg^-1. At the beginning of this study, levels of NH₄⁺ were higher than the levels of NO₃^-, followed by a decrease in residual values, explained by the fact that nitrifying bacteria rapidly transform N from the NH4⁺ form to the NO₃^- form [21]. The daily net mineralization for the different residues was: 0.59 mg N kg^-1 day^-1 for ryegrass, 0.67 mg N kg^-1 day^-1 for balansa clover and 0.74 mg N kg^-1 day^-1 for lupine, for residues from the 1st sowing date, and 0.62 mg N kg^-1 day^-1 for ryegrass, 0.49 mg N kg^-1 day^-1 for balansa clover and 0.68 mg N kg^-1 day^-1 for the lupine of the 2nd sowing date. Values referring to cover crops residues of the 2nd sowing date are the opposite of what was expected, since the residue of the balansa clover presents a mineralization rate lower than the mineralization rate of ryegrass. Li et al. [28], state that legume residues decompose faster than non-legume residues.

Net mineralization accumulated in the plots with the residues of lupine was 82 mg N kg^-1, this value being much higher than that found by Rodrigues et al. [26], also using field incubation and lupine, where the value found was 45 mg N kg^-1. Values higher than those observed in this study, are reported by Nakhone & Tabatabai [22], for other legume species: between 170 and 353 mg N kg^-1.

When analyzing the results of the mineralization model, we found that there were very low mineralization constants (Table 1). The mineralization constant indicates how fast the mineralizable N from residues is released and is not equivalent to the rate of residue decomposition [29, 30]. Such low k values indicate that either the incubation conditions were not satisfactory, or the active N fraction was not easily decomposed by soil microorganisms. This is also in line with the values referring to the half-life (t ½), where it is shown that an appreciable period of time is necessary for the N to mineralize (Table 1). Also, the fact that the highest mineralization rate was observed in the last incubation period, may indicate that the duration of the test was not sufficient to mineralize a large faction of the N. As the test took place until maize harvest, it may imply that there is N still available to be lost by leaching if a new culture is not installed.

The N0 values varied greatly between residues, and even for the same residue, between sowing dates, as was the case for balansa clover (Table 1). The residue from the 1st sowing date had a N0 value of 1385 mg N kg^-1 and the residue from the 2nd sowing date had an N0 value was 131 mg N kg^-1. The value relative to the 1st date residual was an exaggerated value, probably due to an unreliable prediction of the regression. Cordovil et al. [30], state that the value of N0 can vary depending on several factors such as humidity, aeration, temperature, nature and amount of waste incorporated into the soil and other physical, chemical and biological factors. Based on the values found in this study, it is possible to verify that the model used does not adapt to the mineralization of the waste used.

Linear regressions were tested to verify the influence on the mineralization of the N added by the residues and the initial N content in the soil. Although both regressions were significant (p < 0.05 and p < 0.01, respectively), only about 17% of mineralized N is explained by the added N content and about 25% is explained by the initial N content in the soil.

For the residues of cover crops tested and under the edaphoclimatic conditions of this assay, all residues showed N mineralization in most of the incubation periods, with immobilization occurring only in one or two incubation periods, which coincided with the highest recorded temperatures and with the start of irrigation, which allows us to conclude that environmental conditions are of great importance in the mineralization process.

In all treatments, the highest mineralization rate was verified in the last incubation period. The sowing date of the intercropping used as residues influenced the NH₄⁺ and NO₃^- contents in the soil, with these contents being higher in treatments related to early sowing. It was in the treatments with legume residues that a greater increase in N content in the soil was observed. In the treatment with ryegrass, from the second sowing date, there was no net immobilization. It was in the treatment with lupine residue that there was a higher rate of daily mineralization (0.71 mg N kg^-1 day^-1) and, consequently, a higher value of mineralized N (82 mg N kg^-1).

The present study suggests that, under Mediterranean field conditions, the cover crops residues of Italian ryegrass, balansa clover and yellow lupin are important N sources for the maize crop and its use as a source of N is promising, mainly in sustainable agriculture. To better affirm this sustainable solution for the supply of N, it is necessary to carry out more in situ incubation tests, over several years and in other soil-climatic conditions.

Conceptualization, A.P., N.M., H.T. and J.C.; methodology, A. P, and J.C.; software, A.P.; validation, A.P., J.L.S.P., H.T. and J.C.; formal analysis, A.P. and J.L.S.P.; investigation, A.P.; resources, A.P.; data curation, A.P.; writing—original draft preparation, A.P.; writing—review and editing: A.P. and J.L.S.P.; visualization, A.P., J.L.S.P., H.T. and J.C.; supervision, H.T. and J.C.; project administration, A.P.; funding acquisition, A.P. and J.LS.P. All authors have read and agreed to the published version of the manuscript.