Contribution of Agro-Environmental Factors to Yield and Plant Diversity of Olive Grove Ecosystems ( Olea europaea L.) in the Mediterranean Landscape

: Olive cultivation ( Olea europaea L.) is one of the most signiﬁcant sources of income for agricultural areas in the Mediterranean basin, and the olive oil industry as well as the environmental protection are an important part of the Greek agricultural sector. Generalized Linear Models were applied in order to investigate the predictive strength of several biodiversity components and agro-environmental factors for yield and herbaceous plant diversity (species richness) in organic and conventional olive groves of Greece. Our study highlights an increase in yields of organic olive groves by increasing manure application and the earthworms’ density. In the conventional olive groves, yields increase by increasing soil organic matter and the application of inorganic fertilizer N. Also, the herbaceous plant species richness increases with increasing the Shannon diversity index of herbaceous plants, the ﬁeld area, the application of organic fertilizer K and the manure in organic olive groves. As for the conventional ones, herbaceous plant species richness increases with the increase of the application of inorganic fertilizer N. Moreover, some plant species could be regarded as indicators of the differently managed olive groves. Conclusively, this study contributes to the integration of biodiversity conservation with ecologically sustainable agriculture and conservation of agroecosystem. Finally, it could be utilized as a decision and management tool to the scientiﬁc and agricultural community reinforcing the knowledge about the agro-environmental impact in olive grove management systems.


Introduction
A long history of human interaction can be found in the Mediterranean basin. Ancient civilizations such as Phoenicia and Egyptm and later, the Greece and Roman Empires, have formed the landscape and the biodiversity pattern of the area [1,2]. The Mediterranean biome is known to support high biodiversity, so the area is considered as a biological "hotspot" [2,3], as it also hosts a high percentage of endemic species. Many parameters of human prosperity, such as health, freedom of choice, security, socialization and necessities of life are directly or indirectly related to biodiversity-directly through ecosystem services, indirectly through supporting these ecosystems. It is also true that there are a lot of people having profited by transforming natural ecosystems into those dominated by human proving that biodiversity can be exploited [4].
It is a well-known fact that the cultivation of the olive tree has played a vital role as far as human nutrition is concerned. The edible fruits and their oil are known for their beneficial properties and play an important part in our diet. Therefore, olive trees (Olea europaea L.) are probably the most financially important cultivated trees of the Mediterranean basin and their produce has been highly rated since antiquity. Since antiquity, olive oil has The present study was carried out and conducted in 10 organic (O1-O10) and 10 conventional (C1-C10) olive groves], during two consecutive years, 2009 and 2010, in western Magnesia Prefecture of central Greece (39 • 03 12.05" N, 22 • 57 11.84" E) ( Figure 1).
found between herbaceous plant species and management systems by using the Indicator Value Analysis (IndVal) in order to identify possible indicator-species for specific olive grove management systems.
The study area is included in the Quercetalia ilicis vegetation zone and Quercion ilicis and Oleo-ceratonium subzones [32]. Among the cultivated olive groves there is a small percentage (10%) of abandoned ones, while the rest of the area (35%) is covered by maquis such as Olea europaea var. sulvestris (Olive), Quercus coccifera (Kermes oak), Arbutus unedo (Strawberry tree), Pistacia lentiscus (Mastic tree) and Rubus fruticosus (Blackberry), and meadows can be found all around. In addition, around the fields there are meadows.

Figure 1. Study area.
The climate is typical Mediterranean with relatively cold and wet winters, and hot, dry summers. The average temperature is 16.8 °C, with July as the warmest month and January-February the coldest. The average annual rainfall reaches 490 mm (National Meteorological Service of Greece). The characteristics and management practices applied to the organic and conventional olive groves are presented in Table 1. Table 1. General characteristics and applied management practices to the organic and conventional olive groves (Amfissa variety) in central Greece.

Organic
Conventional Average field size (ha) 13  The study area is included in the Quercetalia ilicis vegetation zone and Quercion ilicis and Oleo-ceratonium subzones [32]. Among the cultivated olive groves there is a small percentage (10%) of abandoned ones, while the rest of the area (35%) is covered by maquis such as Olea europaea var. sulvestris (Olive), Quercus coccifera (Kermes oak), Arbutus unedo (Strawberry tree), Pistacia lentiscus (Mastic tree) and Rubus fruticosus (Blackberry), and meadows can be found all around. In addition, around the fields there are meadows.
The climate is typical Mediterranean with relatively cold and wet winters, and hot, dry summers. The average temperature is 16.8 • C, with July as the warmest month and January-February the coldest. The average annual rainfall reaches 490 mm (National Meteorological Service of Greece). The characteristics and management practices applied to the organic and conventional olive groves are presented in Table 1.

Sampling
Herbaceous plant species were surveyed in spring season in organic and conventional olive groves. The sampling of herbaceous species was carried out by the Line Point Method [33] (140 vegetation lines in organic and 140 in conventional olive groves totally was measured). Thereby herbaceous plant species richness was estimated. For determination of the plant samples used the "Flora Europaea" [34,35], the "Flora Hellenica" [36] and the Vascular plants of Greece: An annotated checklist [37,38]. Several important biodiversity components such as cover and density of woody plants, isopods and earthworms (see Appendix A, Tables A1-A5), and agro-environmental factors such as fertilizer, soil pH, and humidity that might affect yield and herbaceous plant species richness were evaluated.

Statistical Analyses
The study was carried out in a completely randomized design with 10 replicates. The Kolmogorov-Smirnov and Shapiro-Wilk's tests for the confirmation of the normal distribution of data were used. The validity of the homogeneous variance assumption was investigated by Bartlett's test. When normalization was necessary, the [log(x + 1)] transformation and inverse hyperbolic sine transformation when X has 0 values were used [39]. All the above tests were performed before each analysis.
Generalized Linear Models (GLMs) are one of the most widely used statistical methods. We used GLM because is most commonly used to model these data, so we focused on models for this type of data. GLM were used to analyze the agro-environmental factors, explaining most variance of yield and herbaceous plant species richness in olive grove management systems [40][41][42].
Because count data such as yield and plant species richness can never be less than zero, the assumption of ordinary least-squares regression is likely to be broken [43]. We assumed yield and plant species richness to be Poisson-distributed random variables and utilized a logarithmic link function in a Generalized Linear Model [44].
Aiming at decreasing the large number of initial variables in the model (n = 27), we first tested the collinearity among the several variables. The highly correlated ones were defined by using a cut level of R 2 = 0.9 (corresponding to the variance factor 10). The best performing single variable model based on Akaike's Information Criterion (AIC) was used as a start with the number of the rest increasing, until the change in explained deviance D 2 was less than 1% [45]. Each of the best n-variable models concluded by comparing all possible n-variable combinations. In order to define the final models, backward elimination based on AIC was used, testing separately linear and quadratic terms, while insignificant parameters were excluded [46]. The changes resulting in the explained deviance D 2 were indicators of the parameter's-variable's performance. The models' robustness was repeatedly evaluated with 10-fold cross-validations. For robust results, the mean of 100 internal cross-validations was used [47,48].
The model coefficients had been taken from the model of robust standard errors and the formula for the model is the below: Y = the dependent variable (also called the predicted variable). β 1 ,β 2 . . . βp = a weight (also called a coefficient). Determines how much weight one variable contributes to the model. If everything in the equation holds constant, b 0 gives the predicted change in Y for a unit change in X. X 1,2 . . . p = a variable. β 0 = the intercept-always a constant.
All statistical analyses were performed using R Statistical Software (version 3.5.0) [49]. Characteristic plant species indicators of organic and conventional olive groves were assessed by the Indicator Value index (IndVal). The IndVal method of a species has long been the most popular measure, commonly used in ecology, to express species importance in a community. IndVals were calculated for each herbaceous plant species on the basis of its abundance and frequency of occurrence in samples assigned to each olive grove management system. The Indicator Value index takes values between 0 and 100. Only significant IndVals (identified with Monte Carlo statistics) (Indval > 50%) were considered [50]. The characteristic plant species takes the highest value (100) when all individuals of a plant species are found in a single management type (high specificity) and when the species occurs at all forms of this type (high fidelity) [51]. This analysis was accomplished with the application of the IndVal software (for a detailed description of the mathematical background for IndVal, see Dufrene and Legendre 1997) [52].

Factors Influencing Yield of Olive Grove Management Systems
The Generalized Linear Models built for yield of organic and conventional olive groves ( Table 2) showed a very good predictive ability (as shown by the respective adjusted R 2 ) ( Table 3). Especially, the models presented in Table 3 include two variables, for both organic and conventional olive groves, which account for 88% and 85% of the variance (based on adj. R 2 ) of olive grove yields, respectively. More specifically, the yields of organic olive groves increase as the application of manure and the density of earthworms increase, while the yields of conventional olive groves increase with the increase of soil organic matter and inorganic fertilizer applications (N).

Factors Influencing Herbaceous Plant Species Richness of Olive Grove Management Systems
The herbaceous plant species of the study olive groves comprises 107 species belonging to 93 genera related to 35 families. The herbaceous plant species recorded in organic olive groves was 101 while in conventional olive groves was 74 species (see Table A2).
The model for herbaceous plant species richness in organic olive groves comprises four variables [herbaceous plant Shannon diversity index (the Shannon (Sd) plant diversity index was examined by Species Diversity and Richness IV software (Seaby and Henderson 2006), in each olive grove management system [(for a detailed description of the mathematical background for Sd index, see Seaby and Henderson, (2006)] [53], field area, organic potassium fertilization, and manure application] (Tables 4 and 5). The model explains a fairly good percentage of the whole variability of the dependent variable and shows that herbaceous plant species richness increases significantly with increasing Shannon diversity index of herbaceous plants, field area, application of organic, potassium fertilization, and manure application to organic olive groves (Table 5).    The model for the herbaceous plant species richness in conventional olive groves consists of one variable (inorganic, nitrogen rich fertilizer) that accounts for nearly half of the total variability of the dependent variable (Tables 4 and 5). Therefore, herbaceous plant species richness was significantly and positively correlated with inorganic fertilizer application (Table 5).

The Contribution of Agricultural and Environmental Factors in the Yields of Organic and Conventional Olive Groves
The yields of organic olive groves were found to be positively affected by the application of manure and the earthworms' density. This is probably due to the fact that the application of manure in organic olive groves contributes to soil fertilization, adding organic matter and nutrients (nitrogen, phosphorus, and potassium), and improving the physical and chemical characteristics of the soil, thus enhancing the growth of olive groves. Organic matter is essential both for the nutrition of the olive trees and for an efficient production system [54]. As for the beneficial effect that the increased earthworm's density has on the yields of organic olive groves, this is probably due to the fact that earthworms decompose organic matter and recycle nutrients by enriching the surface soil through their feces. Thus, earthworm feces are rich in nutrients (carbon, nitrogen, and phosphorus) [55], which as mentioned above are important factors that determine the growth and fruiting of olive trees. According to the literature, earthworms are the most important component of soil fauna, which benefit the formation, conservation soil structure, and fertility [56].
The yields of conventional olive groves were observed to be favored, both by the organic matter of the soil and by the application of inorganic fertilization N. As mentioned above the organic matter is necessary for an efficient production system of the olive groves. Organic matter in the soil plays a central role in regulating the availability of N, P, and K and can also act as a chelating compound, making some micronutrients more available to the roots of olive trees in the form of complexes [57]. Finally, inorganic nitrogen fertilization has been shown to continuously increase yields of olive groves, but only when the N in the leaves is below the adequacy threshold [58].

Organic Olive Groves
In the research area it was observed that the Shannon diversity index of the herbaceous plants, the area of the field, the application of organic fertilization K and manure application were considered the best indicators of the herbaceous plant species richness in organic olive groves.
The above results for the herbaceous plant species richness are in line with the theory that claims that the correlations between herbaceous plant species richness and diversity are simple, positive and powerful [59]. Tuomisto and Ruokolainen (2005), Sulivan and Sulivan (2006), and Solomou and Sfougaris (2011) [60][61][62] found in their research a positive correlation between the species richness and the Shannon diversity index of herbaceous plants in orchards, olive groves and natural ecosystems.
It is documented that organic fertilizer is one of the factors affecting the herbaceous plant diversity [63]. In the organic olive groves of the research area the application of manure and organic fertilizer K increased the herbaceous plant species richness in the following ways: it increases the soil organic matter, it contributes to its fertility by adding nutrients such as nitrogen, it helps to retain its moisture, reduces temporary stress due to lack of moisture, and contains varying amounts of viable herbaceous seeds that promote the preservation and promotion of high herbaceous plant diversity [64,65].
Also, the application of manure in agro-ecosystems enriches the soil with organic matter that very easily affects the measurable functions and processes of the soil. Especially, it is a source of nutrients for plants helping to increase the herbaceous plant species richness and provides an energy substrate for soil organisms [66]. In addition, soil organic matter has a huge impact on various physical properties of the soil, such as the amount of water available for plant growth [67]. Pleasant and Schlater (1994) [68] found that organic fertilizer increases the herbaceous plant species richness by adding plant species, while Yang et al. (2009) [69] that soil parameters, such as soil organic matter, related to the diversity of species of wild fauna and flora. In contrast, Cook et al. (2007) [70] found that the application of manure only as fertilizer did not affect the richness and diversity of herbaceous plants during the first year after its application.
As for the relationship between the area of the field and the herbaceous plant species richness, it probably follows one of the rules of ecology according to which as the area increases, the species richness tends to increase, regardless of the classification group or type of the ecosystem [71,72].  [76] found a negative correlation between field area and herbaceous plant diversity in rural landscapes.

Conventional Olive Groves
Only one variable (inorganic fertilization N) had a significant effect on the herbaceous plant species richness in conventional olive groves. Thus, herbaceous plant species richness increases with increasing the application of inorganic nitrogen fertilization. Regarding the above role of total soil N, a possible interpretation is that of all the nutrients that are applied to the soil, this is the key element that has the most significant effect on the growth and development of cultivated and native plants, and which is the most important limiting factor of growth and yield. The vital role of nitrogen in plants is due to the results of the following facts: it is a structural component of the chlorophyll molecule that is an essential factor for the production and utilization of carbohydrates, is a component of enzymes, a stimulant of plant growth and function, a component of amino acids, which are the building blocks of proteins, and promotes the intake and utilization of other nutrients [77].
Due to the management practices of conventional agriculture and soil erosion which leads to the loss of nutrients (N, P, K) from the soil, it is necessary to apply inorganic fertilizers that aim to enrich the soil with these components. Therefore, in the present study the positive effect of inorganic N fertilization on the herbaceous plant species richness could be attributed to the theory which is based on the model of Al-Mufti et al. (1977) [78] and Grime (1979) ("humped-back curve") [79] and predicts high species richness when availability nutrients are medium. Under these conditions, a small number of competing species tend to dominate the vegetation, leading to the exclusion of slow-growing species. Grime (1979) [79] described the relationship between species richness and productivity as a curved curve with low species richness at very high and low productivity values. The shape of the curve appears like this, because only a few species adapt to these extremely poor nutrient conditions and a few dominant species prevail in all other conditions that have a high level of resources. Studies conducted on a wide range of plant communities have verified the shape of the curve and have shown not only the maximum species richness, but also the maximum species variability observed at medium productivity levels [80,81]. The results from the present study on potassium are inconsistent with the results of various researchers [82,83] who have shown that nutrients are an important factor in the herbaceous plant species richness in agro-ecosystems. However, many studies have shown that increasing productivity, through inorganic fertilizers N, P, and K, is accompanied by a decrease in the species richness of the meadows [84,85].

Typical Herbaceous Plant Species
According to IndVal analysis, the herbaceous plant species Setaria verticillata, Medicago lupulina, Trifolium arvense, Hordeum bulbosum, Malva sylvestris, Papaver rhoeas, Trifolium campestre, Anagallis arvensis, Sinapis arvensis, and Matricaria recutita were recorded as "characteristic indicator species" in the organic olive groves. S. verticillata is a species found in orchards, olive groves and vines. It prefers warm, moist, and rich in organic matter, nutrients (e.g., potassium and magnesium) and relatively neutral pH, clay or sandy loam soils. It grows in well-lit but also in moderately shaded areas. M. lupulina grows in crops, in well-lit but also in moderately shaded moist areas with slightly acidic to weakly alkaline conditions and moderate amounts of nutrients in the soil. It also prefers clay or sandy loam soils. T. arvense grows in olive groves, vineyards and orchards. It needs light and moist sites with slightly acidic conditions and moderate amounts of nutrients and organic matter in the soil [86][87][88][89]. T. campestre is found in olive groves, ditches, and stony sites. It grows in well-lit and semi-shaded areas with slightly acidic, moist, and nutritious soils. H. bulbosum prefers sandy and loamy, nutrient-rich, and moderately moist slightly acidic to slightly alkaline soils. It grows in well-lit but also moderately shaded wet areas [86][87][88][89]. M. sylvestris is the most common type of mallow in agricultural ecosystems. It is often found in high density in olive groves, other orchards or vineyards. It prefers warm and moist soils, as well as soils rich in nutrients and organic matter. Also, it requires high light intensity, weak basic conditions (7.1-7.6) and high soil fertility. P. rhoeas prefers clay and loamy, moist and nutrient-rich soils. It is an indicator plant of non-acidic soils. A. arvensis is found in orchards and arboreal crops and is characterized by broad adaptability and moderate nutrient requirements and soil fertility. S. arvensis prefers well-ventilated, light-rich, nutrient-rich soils. It is an indicator plant of non-acidic soils. M. recutita is found both in cultivated fields and in uncultivated areas. It prefers clay, sandy loam, rich in nutrients but poor in calcium and moderate in moisture, acidic to alkaline [86][87][88][89].
In conventional olive groves the plant species S. oleraceus emerged as an indicator species. S. oleraceus is a plant that occurs in uncultivated areas, orchards, olive groves and areas with trees. It prefers clay and sandy soils with nitrogen adequacy. It grows in well-lit but also in moderately shaded areas and is a plant of warm areas (average annual temperature 14 • C) and soils with slightly acidic to alkaline conditions [86][87][88][89].

Conclusions
The findings of the present study provide useful information on the impact of certain biodiversity components and agro-environmental factors on olive grove yields and can help farmers adopt the right practices leading to more sustainable olive production. More specifically, the results suggest that yields increase with increased manure application and the density of earthworms in organic olive groves while the yields of conventional ones increase by increasing soil organic matter and the application of inorganic fertilizer N. Earthworms, thus, might be used to enhance indirect the olive groves production and are evaluated as possible indicators for monitoring in other similar Mediterranean ecosystems as well.
We must also consider the fact that the herbaceous plant species richness increases by increasing Shannon diversity index of herbaceous plants, the field area, the application of organic fertilizer K and manure in organic olive groves. Herbaceous plant diversity could be utilized as the best candidate to enhance biodiversity components such as herbaceous species richness and evaluated as possible indicator for environmental monitoring also in other similar ecosystems. As regard the conventional ones, herbaceous plants increase with the increase of the application of inorganic, nitrogen rich fertilizer.
Moreover, the «indicator plant species» that have emerged in olive grove management systems (S. verticillata, M. lupulina, T. arvense, H. bulbosum, M. sylvestris, P. rhoeas, T. campestre, A. arvensis, S. arvensis, and M. recutita were recorded as indicator species in the organic olive groves and S. oleraceus in conventional ones) can be used as indicators of environmental conditions, which would otherwise be very difficult, costly, and time consuming to measure. It is noteworthy that they can be monitoring tools in the management of olive grove ecosystems and provide useful information to farmers, agronomists, land managers and the scientific community. Furthermore, the findings revealed useful information towards understanding the functions of these ecosystems and the broader design of biodiversity management practices in managed ecosystems and lead to a more sustainable olive production.
Further research will focus on the evaluation of the environmental impacts of the oliveoil production chain. This effort will be made in order to identify the critical issues and suggest improvements for increasing environmental sustainability and competitiveness of olive cultivation in areas particularly susceptible to human pressure.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available as they were collected for the Research Funding Program: Heracleitus II and it will be used for further processing.

Acknowledgments:
We thank all the farmers of the western Magnesia Prefecture, for allowing us to conduct the experiments on their fields and C. Nakas of the Department of Agriculture, Crop Production and Rural Environment, University of Thessaly, Volos, Greece for his advice on the statistical analyses.

Conflicts of Interest:
The authors declare no conflict of interest Appendix A