Carbon sequestration potential of residues of different types of cover crops in olive groves under mediterranean climate

The maintenance of plant cover between olive grove lanes until the beginning of spring is a soil management alternative that is gradually being adopted by olive growers. As well as protecting the soil from erosion, plant covers have other advantages such as improving the physicochemical properties of the soil, favouring its biodiversity and contributing towards the capturing of atmospheric carbon and its fixation in the soil. A trial was conducted over three growing seasons in an olive plantation situated in southern Spain. It was designed to evaluate the C fixation potential of the residues of the cover species Brachypodium distachyon, Eruca vesicaria, Sinapis alba and of spontaneous weeds; and also to study the decomposition dynamics of plant residues after mowing cover. After 156 and 171 days of decomposition, the species that released the largest amount of C was Brachypodium with values of 2,157 and 1,666 kg ha–1 respectively, while the lowest values of 461 and 509 kg ha–1 were obtained by spontaneous weeds. During the third season (163 days of decomposition) and due to the weather conditions restricting the emergence and growth of cover, spontaneous weeds released the most C with a value of 1,494 kg ha–1 . With respect to the fixation of C, Sinapis records the best results with an increase in soil organic C (SOC) concentration of 7,690 kg ha–1 . Considering the three seasons and a depth of 20 cm, the behaviour sequence of the different species in favouring the fixation of soil organic C was Sinapis > Brachypodium > spontaneous weeds > Eruca. Additional key words: carbon release; cover crops; soil carbon fixation.


Introduction
Tree crops in Spain occupy 4,748,283 ha or 46.5% of the total plantation surface of the area in 15 countries in Europe.The olive tree (Olea europaea L.) is the most common, representing 51% of the area, a figure that is increasing every year due to the lack of profitability of alternative crops.Mediterranean countries account for 98% of the world's olive cultivation area, largely in Spain (2.6 × 10 6 ha), Italy (1.4 × 10 6 ha), Greece (1 × 10 6 ha) and Portugal (0.5 × 10 6 ha).The European Common Agricultural Policy (CAP) budget devoted to olive groves amounts to 2,250 million euros (Beaufoy, 2002).Some 1.5 Mha of the 2.4 Mha olive groves registered in Spain are in Andalusia (MARM, 2010), accounting for over 80% of our production.These groves have traditionally occupied marginal, not very fertile soils broken up by erosion and steep slopes and are hardly suitable for other crops.Only in the last few decades areas with acceptable conditions of soil and climate started to be cultivated.
Currently, olive groves in Andalusia (Spain) suffer from environmental degradation, i.e. erosion, compaction and the risk of diffuse contamination, and also from the loss of soil fertility and the need to replenish the nutrients extracted by the plant or lost in erosion processes.
In order to mitigate this problem, research has been carried out since 1980 to facilitate weed control, improve soil management systems and prevent the mineralization of organic matter (OM) and the loss of soil structure.This has been done using no-till and the establishment of plant covers between the rows of olive trees to protect the soil from erosion (Francia et al., 2006).
The benefits of plant covers recognized in the scientific and technological bibliography are very great: they reduce the pollution of surface waters (Rodríguez-Lizana et al., 2007), improve the water balance in the soil (Bowman & Billbrough, 2004), help to control weeds (Hatcher & Melander, 2003) and recycle the unused nitrogen in the soil (Weiner et al., 2002).liberó la mayor cantidad de C con un valor de 1.494 kg ha -1 .Con respecto a la fijación de C, Sinapis registró los mejores resultados con un incremento de la concentración de C orgánico en suelo de 7.690 kg ha -1 .Considerando las 3 campañas y una profundidad de 20 cm, la secuencia de especies que favorecen la fijación de C orgánico fue Sinapis > > Brachypodium > hierba espontánea > Eruca.
Palabras clave adicionales: carbono liberado; cubierta vegetal; fijación de carbono en suelo.Carbon sequestration potential of residues of cover crops in olive groves

Experiment sites
Experiments were conducted over a period of three agricultural years (2007/08, 2008/09 and 2009/10) at Arenillas olive orchard farm, which was established in 2001 in Fernán Núñez, Córdoba,Spain (37° 40' 1.53" N and 4° 47' W;266 m above mean sea level) on soil with an 11% average slope.The physicochemical characteristics of the soil are shown in Table 1.

Experimental design
The "Picual" variety trees were planted 5 years before at a distance of 4 m × 8 m.The single plot measured 192 m2 and it consisted of two central olive trees with a cover crop strip of 12 m × 4 m to each side.The experimental design was randomised complete blocks sited perpendicular to the slope, with four replications.
The sowing dates depended on weather conditions: 22 nd October 2007, 25 th November, 2008 and 30 th November 2009.Common mustard and rocket seeds were previously collected from spontaneous wild populations and replicated in the Andalusia Research Center, IFAPA Alameda del Obispo (Córdoba, Spain).Cruciferous seeds were sown and buried 0.5 cm deep following the procedures established in previous field studies (Alcántara et al., 2009) at rates of 10 and 3 kg ha -1 for common mustard and rocket, respectively, three years.Brachypodium was only sown the first year at a rate of 100 kg ha -1 following commercial recommendations.The second and third years, Brachypodium was established from a cover crop strip which had been left alive the first year and left to sow itself the following seasons.

Sampling
The cover in the experimental olive grove plot was mown in April, and from that date onwards and up to the autumn sowing of the new covers, plant residues were periodically sampled during the agricultural years 2008, 2009 and 2010.In each species, in each block, areas with a high accumulation of residue were selected, and three residue collection points established, which made a total of 12 samples per type of cover and sampling day.
In the three growing seasons, the soil was sampled at the beginning and end of the decomposition period of the plant cover at depths of 0-5, 5-10 and 10-20 cm from each plot.The same occurred in the case of the plant remains, three sampling points being considered in each of the four control subplots per species.The samples were extracted with a Veihmeyer tube and transported to the laboratory in a plastic bag.Subsequently, the soils were air dried and run through a 2 mm sieve.

Analysis of samples
The biomass of the stubble residue was estimated in a 0.25 m 2 metal frame which served to mark out the sampling area and was placed at all the selected points.The residue collected was sent to the laboratory where it was washed with distilled water to prevent contamination in subsequent analysis and was placed in an oven at 65 °C until it reached a constant weight and it was possible to estimate the amount of dry matter.The cover percentage was measured following the evaluation per sectors method described by Agrela et al. (2003), which is characterized by the use of a 1 m 2 frame divided into 100 reticules.The method consists of a subjective assessment of the different percentages of cover estimated in each reticule on a scale of 0 to 5 according to the greater or lesser amount of cover.Total C in the residue samples was analysed in a LECO elemental analyser.
The soil samples were air-dried, ground and sieved through a 2 mm mesh sieve for subsequent analysis.The determination of soil organic C is based on the Walkley-Black chromic acid wet oxidation method.Oxidisable matter in the soil is oxidised by 1 N K 2 Cr 2 O 7 solution.The reaction is assisted by the heat generated when two volumes of H 2 SO 4 are mixed with one volume of the dichromate.The remaining dichromate is titrated with ferrous sulphate.The titre is inversely related to the amount of C present in the soil sample (Sparks et al., 1996).

Cover-residue biomass relation
In order to determine the relationship between biomass and cover percentage, a grade 2 polynomial was used, of the type: the same as that used by Lyon (1998) in dryland crops, where M t (kg ha -1 ) is the residue biomass at instant t.Likewise, the exponential model proposed by Gregory (1982) was used: where Cover t is the fraction of cover at the instant t (%), Cover max is the fraction of maximum cover (100% in all cases), k is the coefficient of cover calculated by the model (ha kg -1 ) and M t is the residue biomass at the instant t (kg ha -1 ).

Spatial-temporal distribution of stubble residue in the soil
This section evaluates the variability of that percentage under field conditions.For this purpose, 52 field strips were selected, from which samples of 0.25 m 2 were selected over time.
In order to analyse the temporal stability of the cover percentage in the different strips, a method similar to that proposed by Vachaud et al. (1985) was used.This was based on the concept of temporal stability, calculating averages and variance over time.
In this case, unlike the method proposed by the cited authors, we calculated the temporal means of each strip, rather than the relative differences, as this was of interest in order to ascertain the average cover.

AC strip
Cover where AC strip i represents the mean temporal cover in the strip i; n = samplings done in each treatment of residues; i=1,2…,52; Cover (%) it : cover percentage obtained in the strip i, instant t.
with σ(AC_strip i ) denoting the standard deviation of the mean, calculated as an estimator of temporal stability.From this point of view, time-stable locations (strips) are defined as those with a low value of σ(AC_strip i ).

Data analysis
The climatology of the area was monitored in the three years studied, with precipitation and maximum and minimum daily temperature data being evaluated.
The percentage of original dry weight and organic carbon (OC) remaining at each sampling were regressed over time using the linear regression model procedure of SPSS 11.
An analysis of variance (ANOVA) was performed for all the parameters measured and comparison of means was carried out by the Tukey-test with p ≤ 0.05.

Dynamics of residue biomass
Fig. 1 depicts the temporal evolution of weather and biomass of the plant residues of the different species Carbon sequestration potential of residues of cover crops in olive groves of cover crop used in the assay for the three agricultural years.
After mowing in the first year, the species with the largest residue mass was Brachypodium with 7,323 kg ha -1 , followed by Sinapis with 3,141 kg ha -1 , then Eruca with 2,960 kg ha -1 , and the least residue mass was found in spontaneous weeds which, at the beginning of the sampling, recorded 2,148 kg ha -1 .Throughout the decomposition period, significant differences were noted between the weight of the Brachypodium remains and that of the other covers (Table 2).These differences decreased as the plant remains grew.
In the second year, after mowing, the data recorded were as follows: 11,038 kg ha -1 for Brachypodium, 7,461 kg ha -1 for Eruca, 6,732 kg ha -1 for spontaneous weeds and 5,850 kg ha -1 for Sinapis (Fig. 1).It can be seen that there was a higher production of residue in the second year in all cases as a consequence both of the seeding of the different species and of the rainfall recorded in the autumn and winter months, which favoured the growth of the cover.In this case, the significant differences in the weight of the residue of the different species are not so clear and vary from one sampling date to another (Table 2).
The exceptional weather conditions during autumn (September-December) 2010, with 550 mm of rainfall in the experimental area, which is equivalent to a normal annual average, allowed plants to emerge and grow normally, thereby restricting the production of residues to a great extent in the third agricultural year.This situation caused a decrease in biomass values after the clearing of all covers, spontaneous weeds registering the best data with 5,949 kg ha -1 and Brachypodium the lowest figures with 4,137 kg ha -1 .No significant differences are appreciated from one species to another (Table 2).

Dynamics of residue cover
In the second year, the trend was similar among the different covers and significant differences were only noticed between Brachypodium and the remaining species on the last two sampling dates.In third sample season, Brachypodium once again recorded the highest cover values throughout the decomposition period, but no significant differences were observed in regard to the rest of covers (Table 2).

Cover-residue mass relation
Table 3 and Fig. 3 show the result of applying the two models to the relationship between the percentage of soil cover in terms of biomass.The relationship between the soil cover percentage and its biomass was not significant between the variables considered for Brachypodium.The high percentage of cover maintained by this species during the whole decomposition period, even with low values of biomass, due to its capacity to regrow, may be the reason for it not fitting the models proposed.Spontaneous weeds provide the best fit to both models, whereas the rest of the species show lower coefficients of determination (R 2 ), especially Sinapis.
In accordance with the experiment data obtained, we calculated the amount of residue necessary to obtain a 30% cover, this being the limit used in the definition of conservation agriculture.The results are presented in Fig. 3 and Table 4.The exponential model (Gregory) shows that we need around 1,000 kg ha -1 for all the species, which is larger than quadratic model, although considering a range of cover of 25-35% and 28-32%, measured values were larger than both models (Table 4).The residue mass was piled up covering a portion of soil without being effectively dispersed, and we ob-

Spatial-temporal distribution of the residue on the soil
Fig. 4 shows the average cover of the 52 strips controlled for the different species and seasons sampled.Covers were highly variable in the 2008 season, recording covers of less than 30% and others of 100%.More specifically, cover ranged from minimums of 29, 26, 21 and 12% to maximums of 100, 89, 94 and 68% for Brachypodium, Eruca, Sinapis and spontaneous weeds.During the same year, the time stability of the cover did not follow a specific pattern and variations in the standard deviation at control points are similar for the different covers.
In 2009, the favourable growth of the covers resulted in less spatial variability in the percentage of residue cover at the selected points in each of the species.The lowest variation was observed in Brachypodium (90%-100% of cover), while the highest variation was recorded by Sinapis (41%-100%).

Cover [%]
Brachypodium Eruca Sinapis Spontaneous  both seasons.The highest standard deviations were observed at the points where the percentage of cover was lowest, while greater stability over time was observed at the points where the percentage of cover exceded 85%.

Release of carbon
Table 5 shows the amount of C released and the reduction in the mass of the different species considered in the study after the decomposition period.Brachypodium recorded the greatest biomass losses in the first and second years, which gives an idea of the ease of decomposition of the residues of this cover which, at the beginning of the decomposition cycle, had a C/N ratio close to 20, which was lower than that for the rest of the species (Sinapis: 34; Eruca: 29; spontaneous weeds: 23).
Due to the weather conditions, which affected the growth of the covers and the amount of residues after they were cleared up, spontaneous weeds registered the largest biomass loss in the third year and the highest values of residue mass at the beginning of the decomposition period.As regards the amount of organic C released during decomposition, these values are highly disparate among the different species (Table 5).

Soil carbon fixation
The effect of the decomposition of the residues of the different covers on the concentration of organic C in the soil has been evaluated.Table 6 shows the values of this parameter in the soil for the samplings carried out and the increase in C estimated in the three years during the decomposition period.
The non alteration of the soil, leaving the cover residues on the surface for three consecutive years, has  Biomass [kg ha   Most olive growers use spontaneous weeds as soil cover.As a result, the increase or decrease in the amount of C that has been fixed by the different species has been represented in Fig. 5 in relation to that estimated for spontaneous weeds.The planting of cover has fixed 47% and 5% more C in 20 cm of soil than the measure recorded in soils where the native weed flora was left.

Discussion
The decomposition process of the harvest residues are influenced by edaphic and environmental factors like: temperature, moisture, availability of nutrients, microflora and soil fauna, by factors inherent to the residue such as their C/N ratio, content of lignin and soluble carbohydrates, and by management factors like the amount of stubble and its size (Thorburn et al., 2001).However, the most important ones are the climate variables and how susceptible residues are to being colonized by microorganisms (Soon & Arshad, 2002).Authors like Ernst et al. (2002) have carried out studies on different residues and concluded that the C/N ratio determines its decomposition rate.Ruffo & Bollero (2003) indicated the need for a better knowledge of residue decomposition through research conducted under more realistic field conditions.
In regard to the dynamics of residue cover, the benefits of conservation agriculture systems with regard to soil protection from erosion and an improvement in soil water balance are associated with the presence of plant residues covering the soil at times when there is no crop.However, the percentage of the cover and its persistence depends on the type of residue and on the climate in the area (Gajri et al., 2002).
In our case, considering all decomposition period, none of the species recorded mean cover values of below 30% that is the limit over which the soil would be protected from erosion agents (Conservation Tillage Information Center, 1990).Snelder & Bryan (1995) investigated the relationship between cover density and soil loss under simulated rainstorms.In their experiment, a critical threshold occurred with a 55% cover, below which erosion rates increased rapidly.
Analysing the stability of the percentage of cover over time makes it possible to define the persistence of a behaviour pattern in each strip in regard to the rest of strips over time and identify areas with little or excessive cover.The accumulation of biomass or cover in certain areas of the terrain could make later operations performed in olive groves more difficult and affect the -2,000 0 2,000 4,000

Spontaneous weeds
Soil Organic Carbon [kg ha -1   ] (depth: 0-20 cm) Carbon sequestration potential of residues of cover crops in olive groves decomposition of the residue, as well as areas with little cover reducing the level of protection against erosion (Ayed & Mohammad, 2010).
The study undertaken reveals that the greatest spatial variability of the cover was observed in the first and third seasons when less residue was produced and irregular distribution left areas with a cover of less than 30%.
Regardless of the sample season considered, Brachypodium records the lowest standard deviations, which indicates the cover provided by this species is more stable over time.
While most crop residue studies related to erosion control or the effect of tillage on residue retention express residue data primarily as a percentage of soil cover, studies dealing with residue decomposition usually calculate residue losses in terms of mass.Due to the time and work involved in obtaining residue mass data and the difficulty attributed to residue cover determination methods, there is an interest in establishing relationships between residue mass and soil cover for prediction purposes.
As we show in Table 3, the cover coefficients k of the different species are lower than those estimated by Steiner et al. (2000) in crops of barley, oat, spring wheat and winter wheat in field experiments, k between 0.0099 and 0.00162 ha kg -1 , and also lower than that indicated by Ordóñez et al. (2007) for peas (k = 0.0011 ha kg -1 ).This could be due to the fact that the amount of biomass necessary to achieve 100% cover in the different species of plant cover is much greater than that necessary for the crops previously cited.López et al. (2005) indicated that in order to achieve a 100% cover with barley residue, between 2,000 and 3,000 kg ha -1 of biomass are necessary, which contrasts with our data in which 7,500, 5,900 and 6,700 kg ha -1 of residues of Eruca, Sinapsis and spontaneous grass weeds, respectively, are required to achieve maximum cover.In fact, the values of the coefficient k are very similar to that reported by Gregory (1982) in corn crops, with k = 0.0004 ha kg -1 , with residue values of over 8,000 kg ha -1 .
In addition to protecting the soil, another important characteristic of the residues is that they supply C.However, this depends on the composition of the residue and on how easily it decomposes.The C release rate of the different residues was estimated, understanding this to be the difference between the content of this element in the stubble when the covers are mown and that estimated in the residue samples collected on different dates.
According to the results shown in Table 5 and considering three years, the C release rate of Brachypodium residue was 1.5, 1.9 and 1.9 times higher than that of Eruca, Sinapis and spontaneous weeds respectively.At the end of its decomposition cycle, Brachypodium stubble had lost 72%, 42% and 40% of its initial C content.
The edaphoclimatic conditions of the area and the characteristics of the residue played an important role in the evolution of plant residues.In all cases, the highest percentage of C released by the decomposition of the residues was recorded in the first year because of lower rainfall recorded in the second year.Some authors like Aulak et al. (1991) and Baggs et al. (2000), mention that moisture is important as a trigger of the the decomposition process and even more so in the case of residues with a low C/N ratio.Authors like Ernst et al. (2002) have carried out studies on different residues and concluded that the C/N ratio determines its decomposition rate The relationship between the amount of biomass and C released by the decomposition of residues approaches 2.5, regardless of the type of cover and sample season considered.This indicates that 1 kg of C is released for every 2.5 kg of biomass that are lost.It has been amply proven that when changing from traditional agriculture (intensive tillage) to conservation agriculture, the content in OM in the soil increases over time, with all the positive results that this brings with it (Bravo et al., 2007).
The C sequestration values observed in Table 6 were higher than those estimated by Castro et al. (2008) in olive grove soil where plant cover was maintained for 28 years, and similar to those estimated by Márquez et al. (2008) in an olive grove with a cover for 4 years.In both cases, the cover was spontaneous weeds.González et al. (2012) in a study with a statistical analysis of cover crops reviewed studies show that those with native species obtained a sequestration mean of 1.78 Mg ha -1 yr -1 , while those with sowed species reached 1.16 Mg ha -1 yr -1 .In our case, Sinapis and Brachypodium sequestered 3,618 and 207 kg ha -1 more than spontaneous weeds.Considering three years and a soil depth of 20 cm, the behaviour sequence of the different species in favouring the fixation of organic C in the soil was Sinapis > Brachypodium > spontaneous weeds > Eruca.
As final conclusions, under the edaphoclimatic conditions in southern Spain, the plant residues remaining from the olive grove covers when the latter are mowed in April have the two-fold task of protecting the soil from intense spring and summer rain and favouring the maintenance of soil fertility with the release of C and nutrients as they become degraded.The summer protection of the soil by the residues of different species has been assured as, in the worst of the cases (Sinapis), 38% of the cover was lost until the next cover is established.Despite that, a spatial variability study indicates the convenience of uniform residue distribution in order to ensure cover in the entire area and avoid points that lack protection and could restrict the benefits of cover when faced with erosion.
Likewise, the mass-to-cover relationship established in this study for different cover residues could be used to estimate soil cover from residue mass throughout the decomposition period by using a single k coefficient for each species.
As regards the protection of the soil, Brachypodium developed the greatest amount of biomass and maintained the highest and most stable levels of cover throughout the period under analysis.
In reference to the effect of plant covers on C sequestering, Sinapis fixed the most C, namely 7.7 Mg ha -1 in three years, which represents 44%, 47% and 54% more than the amount fixed by Brachypodium, spontaneous weeds and Eruca, respectively.
Although spontaneous weeds are the most popular alternative among farmers when it comes covering the soil of their olive groves, the results of this study reveal that other types of plant covers not only improved soil fertility, but also yield more environmental benefits as regards their contribution towards reducing erosion processes and fighting climate change.

Fig. 2
Fig.2represents the temporal evolution of the percentage of soil cover for the residues of the different species considered.Brachypodium cover registers the

Figure 2 .
Figure 2. Temporal evolution of the cover percentage for the different species and years considered in the study.

Figure 4 .
Figure 4. Average cover by ranges in the sample strips.Vertical lines represent the standard deviation obtained in each strip.

Table 1 .
Characteristics of the olive grove soil on which the experiment was

Table 2 .
Comparisons of residue biomass (RB) dry weight means and of cover means between species for the different years and dates sampled, based on the analyses of variance and the Tukey test.Different letters between covers represent significant differences at a probability level of p ≤ 0.05 Carbon sequestration potential of residues of cover crops in olive groves tained amounts of residue mass larger in Eruca and Sinapis than spontaneous weeds.

Table 4 .
Residue biomass (kg ha -1 ) necessary for reaching 30% of cover according to the different species and models considered in the study, and measured values (kg ha -1 ) for a range of measured cover between 25-35% and 28-32%

Table 5 .
Loss of residue biomass and release of carbon from plant cover in the experiment plots for the 2008 (157 days of decomposition), 2009 (172 days of decomposition) and 2010 (163 days of decomposition) agricultural years Comparison of the quantity of C fixed by different species of plant cover to that fixed by spontaneous weeds in the entire profile of soil (depth: 0-20 cm).

Table 6 .
Content of organic carbon in the soil (SOC) at the beginning of the 2008 sample year and at the end of the decomposition period of the residues in the third sample year, and the carbon fixed for the three agricultural years considered in the study.Different letters between covers represent significant differences at a probability level of p ≤ 0.05