3.4. Relationships between Growth Parameters

In spite of their variations, canopy diameter, tree height and trunk diameters were highly correlated (Figure 13). Canopy diameter (D_f, cm) increased proportionally to tree height with r = 0.98 and to trunk diameter with r = 0.94.

Similarly, tree height increased proportionally to the increases of trunk diameter with r = 0.99. The correlative matrix is given by the following table.

where ΔH is the annual increase of tree height, D_c, D_t and H are relative to measurements made for canopy diameter, trunk diameter and tree height, respectively determined for each year of monitoring from 1^st January to 31 December.

3.5. Root Development

3.5.1. Root Profils

Root profiles made for young olive trees aged one to six

years showed two or three types of roots following the tree age (Figure 14). During the first years of cultivation, trees developed fine roots in the upper layer (20 cm), which extended rapidly in lateral and vertical directions with inclination varying from 30˚ to 60˚ depending on their size. Older trees developed roots horizontally beyond the first 30 cm.

Root number and diameter emerging from the lateral face of the trench are given in Table 5.

Most roots (70%) develop in the first 60 cm of soil. Largest diameters evolve between 2 mm (one year old tree) and 32 mm (four year old tree) depending on tree age. Large roots developed horizontally with numerous ramifications (fine roots). Some roots developed at deeper strata and reached 1.0 m depth. Very few roots were found below this depth even for the oldest tree.

3.5.2. Root Extension

Results presented in Table 6 show that the main development of the root system occurred during the first 2 - 4 years of cultivation, horizontally and within the top layers (0.2 - 0.4 m). During this period, the soil volume explored by roots increased at regular rate of about 1.0 m³/year.

The volume explored by the 5th year-old-tree represents 47% of that reached by the older tree.

3.5.3. Root Density

Results relative to root density estimation are reported in Figure 18. A noticeable root concentration was observed for both East and South directions close to trunk in the top layers around each of the six trees. The average root densities (D_r, cm·cm⁻³) determined for the cultivar Chétoui are reported in Table 7.

Absolute values ranged between 0.001and 0.670

cm·cm⁻³ with averages varying from 0.070 to 0.303 cm·cm⁻³ following the depth, distance to trunk, direction and tree age. Greater values of up to 0.5 cm·cm⁻³ were recorded in the first sixteen centimeters and at 40 cm from trunk. These values decreased significantly as distance to trunk increased (except some measurements for two and three year old plants). Roots were less frequent at all depths outside the canopy limit and particularly for the deeper layers. At these depths, however, we should mention that root densities rarely exceed 0.040 cm·cm⁻³ for both directions.

3.5.4. Length of the Root System

The overall length of the root system varied from 1.0 km to 33.9 km following age (Table 8).

Significant increase of the root system length was observed for the six-year-old tree. It was 4.8 times greater than that recorded on the previous year. The lowest value

was recorded for the four-year-old tree. There was no any apparent cause which could explain this result.

3.5.5. Root Development and Soil Moisture

Simultaneous monitoring of soil moisture carried out during the campaign 2003 nearby the emitters and at the limit of the canopy, showed large variation with values ranging between 15% and 39% following the depth and the distance to trunk (Figure 16). Lower values were observed in the upper layers, while minimums were recorded within the superficial strata (0 - 20 cm) as a result of water evaporation and root absorption, confirming the concordance between root development and soil water depletion [72]. Results showed also large variation between measurements made at the limit of the canopy while low variation of soil moisture was observed nearby the emitters with values ranging between 32% and 38% following the depth.

3.5.6. Root Development and Canopy Growth

Canopy growth is correlated to root development. Results presented in Table 9 showed for trees aged one to four years larger extension of roots in comparaison to values observed for their canopies. Then, differences between canopy and root extension decreased and roots reached for the six-year-old tree maximum distance to trunk of 2.10 m, while the canopy limit was observed at 1.95 m.

The projected canopy area (S_c), derived from canopy diameter measurements, increased slowly after planting to reach 0.21 m² for the one-year-old tree and 11.94 m² for the 6-year-old-tree (Table 10), while the root area

progressed with a constant rate of 1.2 m² per year to reach 13.8 m² for the 6-year-old-tree.

Significant relationship was found between canopy (S_c, m²) and root (S_r, m²) areas, which can be approximated by a linear model with a correlation coefficient r of 0.94, illustrated by Figure 17, where:

The S_r/S_c ratio derived from both canopy and root areas decreased significantly from 20 to 0.9 following tree age. For the four, five and six year-old trees, this ratio approximated the unit.

The decrease of the S_r/S_c ratio implies a tendency to equilibrium between the under-ground and above-ground organs beginning from the fourth year after planting, as a result of the establishment of competition between shoots, roots and fruits. In fact, trees begin to produce olives

within the second year after planting and the first commercial crop arrived in year four (6.5 kg/tree). Results indicate also that young olive trees seem to be able to adjust their root systems to the larger above-ground development during the winter rest. This feature is well traduced by the root length/leaf canopy area ratio. We have found a value of 2.3 km·m⁻² of leaves for the 6- year-old tree. This value is considered as an optimum value for such conditions.

3.6. Crop Water Needs, Canopy and Root Development Relationships

Exploration of the usefulness of the developed approach is detailed in the following section.

In order to link the water supplied to trees to the evaporative demand, a supply ratio (K_supply) that takes into account only the tree-related quantities was defined as developed in a previous section. This ratio could be considered as a crop coefficient for young trees when reference evapotranspiration, rainfall and irrigation amounts are computed according to the previous equations and expressed in m³/tree. Adoption of such ratio allowed estimation of irrigation requirements for different rainfall and evapotranspiration regimes. The different water supply ratios, K_c-FAO, I/ET_o, P^* + I^*/ET^* and I^*/ET^*, determined for each of the six olive trees are given in Figure 18 to compare results.

Results show that the ratio of applied irrigation (I, mm) to reference evapotranspiration (ET_o, mm) calculated for the dry season from April to August was very low. It increased from 0.02 to 0.14 when trees grew from one to six years. When using the volume method to calculate the irrigation and precipitation values falling on the area covered by roots, K_supply becomes very close to the K_c-FAO.

Canopy development and root extension are essential for irrigation requirement estimation as they determine evapotranspiration and water availablity for the root system. Applied water didn’t seem to be limiting to vegetative growth since good development of trees was observed during the years of experiment. Application of

reduced amounts of water during the first 4 years, varying from 20 mm to 50 mm per irrigation with low frequency (4 irrigations annually except in 1998) induced large root development. During the following years, 2002 and 2003, root development was confined to the 40 - 60 cm layer as a result of application of drip irrigation. However, while canopy cover measurements are considered as easy, area covered by the roots is much more difficult to carry out. Our progress in the future will be measured by our capacity to integrate knowledge on water supply, evaporative demand and the soil volume explored by the root system for different locations and planting densities. Estimation of the effective rainfall remains however a big challenges for using the proposed method.

3.7. Relationships between Production and Growth Parameters

Picholine and Chétoui varieties produced fruits during the second year after plantation. These varieties behaved differently but they provided the same cumulative production and high water use efficiencies (Table 11). Low productions obtained for Meski cultivar with maximum yield of 0.5 tones resulted from its self-incompatibility.

Olive production was found to be highly correlated to the biomass produced by pruning during the low fruit years with r = 0.95. During high fruit yield years, the correlative coefficient was 0.69. Average yield was found to be correlated to the previous year’s canopy diameter gain, which was itself correlated to the biomass produced by pruning with r = 0.61 and r = 0.87, respectively as shown in the following Figure 19.

Canopy volume and fruit size were also found to be correlated to production with r = 0.90 and an optimum fruit size obtained with 5 - 6 kg of olives/tree, respectively.

In the other hand, wood water content was found to be correlated to the previous year rainfall amount received from July to December with r = 0.94 (Figure 20).

Wood water content ranged between 34% and 86% in average according to the year, with slight differences recorded between varieties (Figure 21). Values slow down in 2001 and 2002, but increased in 2003, improving changes between roots and canopy.

Low value was observed in 2002 for Picholine, a warm and high fruit load year, whereas the maximum value was observed for cv., Meski in 2000, the first pruning year.

4. Discussion

Young olive trees of cultivars Chétoui, Manzanille, Meski and Picholine grew during the first years of cultivation with variable rates, providing typical but different models of growth. During the planting year (unproduc-

tive year) and the low fruit load years, vegetative growth was important for all growth parameters, producing high amounts of assimilates and low C/N ratio. The following years, trees set fruits and yield increased substantially, while annual growth rates decreased substantially, reaching their lower values six years after plantation. Similar patterns of growth were reported for almond and peach trees [31], showing optimum increases four years after plantation and high variations from year to year in response to the changing conditions and the degree of adaptation of varieties to the environmental factors and their changes. Low rates recorded for canopy and tree height after the optimum value seems to be inherent to the establishment of competition between the vegetative growth and the growing fruits, which is well-known for adult trees [3,4,7,9,24,30]. Apparently, competition for nutrients affects also young trees, making fruits competing strongly with shoots. It was reported that the partitioning of carbohydrates is mainly controlled by the developing sinks and mainly olives [5], which drive after fruit set high the nutrients provided by the nearby leaves and even those stored in the trunk to ensure their development, thus, limiting the growth of the surrounding organs and even the development of distant ones [3]. Such situation was observed during the two last years of our experiment, when high productions were recorded, while tree height, canopy and trunk diameters growths were substantially reduced. However, although fruits are considered as main sinks, their strength may vary consistently during the season depending on the nutrientavailability, fruit load and the stage of development, allowing growth of the other organs, with, however, variable rates. Some works [24] indicated that trunk growth is reduced following harvest of a heavy fruit load and increases progressively after full bloom and may reach high rates at the end of summer and in early autumn. Consequently, and accounting for the relationship existing between the trunk and the branches, fruits may reduce vegetative growth directly or indirectly [77]. Therefore, the distance between the active sinks and leaves should be considered as an influent parameter when investigating the sink effect, since individual leaves can supply any particular sinks. The influence of the sinks can be controlled by thinning and pruning. Correct training is needed in the intensive orchards to maximize light interception and to promote the photosynthetic activity of the canopy. It is essential for the tree to adapt its size to the planting density, to set fruits early and to get high yields [3]. In our orchard pruning was practiced beginning from the third year of cultivation. Different degrees of severity were applied depending on the year’s conditions and variety, removing variable amounts of wood. The canopy size was reduced consistently during the first pruning year (50%). This affected greatly the growth of the trees, which reactdifferently depending on cultivar and year. The responses of the trees were not observed immediately after pruning, but later in the season (2000) or on the following years (2001 and 2002), when important amount of biomass was produced. This was useful to enhance flowering and yield of cultivars Manzanille, Chétoui and Picholine. These varieties set fruits during the second year of cultivation and provided their highest yields three years later, reaching values as high as 3.1, 2.5 and 2.3 tons/ha of olives, respectively. Adversely, compact canopies, like that of cv., Meski, leaded to low productions and fruit size at harvest due to the limitedinterception of light, resulting in reduced number of potential sites of fructification.

In addition to carbohydrate-availability, temperature is reported as an essential driving factor, controlling bud development and shoot elongation [4]. The increase of temperature recorded during the years of experiment (2000-2003) and particularly that registered in 2001, shortly before the spring active period of growth, enhanced consistently tree height growth. However, although this result corroborates previous works [34,45], reporting high correlation between early spring growth and temperature, it can’t be attributed to temperature increases only, because growth is also controlled by some endogenous factors and particularly during the “growth waves” [3-30]. Water affects also tree growth. Limited water resources during the early spring reduce substantially the photosynthetic activity of leaves, shoot length and cell turgor, leading eventually to a growth pattern dominated essentially by an autumnal flush, as it was observed in 2002. On that year, growth rates decreased substantially after the spring period and the summer flush was absent. Water shortage contributed also to slowing down the growth of trunk, which is reported to be more responsive to water shortage than other organs [7-24]. Therefore, trunk shrinkages can be measured at field level and used as indicator to manage efficiently irrigation of the orchard. With such practice, the negative effects of water shortage can be avoided and particularly during the critical stages of shoot and fruit development. These stages were identified through the establishment of the growth curves and the study of their dynamic.

The average growth pattern of tree height and shoot elongation showed seven distinct periods of growth. During the first stage, bud-break occurs. This process depends on the reserves of the tree, and may be inhibited if the amount of carbohydrates is not sufficient to ensure good development of buds [7-9]. It was reported that at bud-break, carbohydrate reserves provide the carbon needed for buds to grow until the leaf area of the tree provides enough assimilates to meet the sink demand. Shoots did not begin to export carbon to the rest of the plant until 15 - 25 days after bud break. Damages occurring during this first period of the growing pattern are essentially due to hot temperatures and water shortage. Then, shoots begun to grow, producing higher amounts of carbohydrates. Their elongation depends on the prevailing conditions. The first rapid period of growth (Period 2) occurred during the two first weeks of April for tree height, and some weeks later for lateral shoots, providing the most important elongation of the year. Indeed, the winter-spring growth produced in average 49% of the annual increase, while the spring contribution approximated 33% in average with important differences between years. This result corroborates previous works indicating that 50% to 80% of the vegetative growth occurs before flowering in response to the increasing temperature [33,45]. The rates achieved by the vegetative organs depend at this stage on the available assimilates, but also on root dynamic. It is generally thought that roots and shoots are in competition for carbohydrate resources during this period. Some works indicate that roots have priority for growth because the first noticeable development occurs in the roots with a flush of growth prior to any visible foliage development [3]. In general a bimodal periodicity of root growth is observed for many fruit crops, amongst olive. This model is characterized by a flush of growth occurring in the spring prior to the up ground vegetative growth, then, another flush of growth happens, just after the terminal bud setting.

The summer growth refers to growth observed from end of May to early September, occurring with variable rates. It was highly influenced by the watering conditions. The average gain reached 39% of the annual increase. This growth seems to characterize the irrigated orchards, only. Previous works [30] reported an important growth for Picholine of Morocco during the summer season when it is cultivated under irrigated conditions, while rain fed plantations showed only two active periods of growth [19], in the spring from March to end of May and in the autumn during the months of September and October. Summer growth was also influenced by tree load and fruit development stage. Although similar growth patterns were observed for productive and less productive cultivars and for bearing and non-bearing years, the rates achieved were higher for the fruited years.

Fruits grew following distinct periods of growth, showing typical response to increasing light and temperature with maximum development observed from May to mid July and from end of August to end of October. Changes in fruit size are caused by cell division and cell enlargement in varying proportions at different times. Fruit development is the result of a long fructification process, beginning during the previous summer season around pit hardening [24] toward the following autumn, and thus covering some vegetative growth stages. During these periods, different fructification processes occurs, mainly flower bud induction, flowering, fruit division and enlargement and affect significantly the vegetative growth [18,24,27]. Previous studies indicate that in fruit crops, there is a strong inverse relationship between vigor and flower bud initiation [24,30,32]. Under conditions that produce high vigor, vegetative bud development has precedence over reproductive bud development and any factor that decrease vigor promotes bud induction. This agrees with our finding when monitoring simultaneously shoot and fruit growths under high and low fruit load productions, since slow rates were recorded for lateral shoots during the rapid fruit growth period and particularly during the first stage of cell division, due to competition for nutrients and water. Indeed, as fruits are strong sinks, draining the nearby available substances as well as for those located in other sites, they may reduce greatly the translocation of assimilates to any other organs, thus, limiting their development [7]. Then, when fruit growth rates slow down during and after pit hardening, shoot rates increased again. However, beginning from mid July, fruit growth as well as growth rates of vegetative organs decreased significantly. This period, characterizes the presumed “bud induction” process which seems to affect all the physiological processes of growth [7,30,32]. However, the beginning of the bud induction process and its duration were not clear if we referred only to the low rates observed on the graphic model. Additional analyses should be made to precise the time at which it occurs. Here, contrasted hypothesis were given. Some of them claiming flower bud induction happened in winter [78], while others reported two different induction periods [4,5]; the first is reversible, taking place in summer, and the second, which is not reversible, happens on winter. According to this last hypothesis, buds may remain in a condition of reversibility until February. Under high fruit load or/and restricted water conditions, buds may return to their initial state (vegetative) or abort [32,44]. Flower buds differenciate in February and flowers open in April. Flowering was observed in our case, first on Manzanille and Chétoui cvs., and then on Picholine. Full bloom duration ranged between 21 and 24 days-period depending on year and variety. In 1999 and 2000, full bloom was observed during the usually reported period for the Mediterranean region, on 12 May and 4 May, respectively. Under warmer conditions, flowering occurred earlier around 4 May in 2001 and 17 April in 2002. It was reported in early works that warm climate accelerate the flowering process which may be also inhibited under excessive hot temperatures [35,36]. Fruit set occurred in May and was enhanced by cross pollination thanks to the presence of different varieties. During this period, fruits develop rapidly and high amounts of assimilates are used to ensure cell division, resulting in significant reduction in shoot elongation [3]. This process is reported to be highly depending on water availability and determines fruit size at harvest. Fruits continue to grow with high rates for about one month period to reach the stage of pit hardening by the end of June. During this period of endocarp scarification (stage II of fruit growth), cells of the stony endocarp increased normally by cell division, but accumulation of dry-matter within the tree is, at this stage, at its lowest level, explaining the low vegetative rates. Indeed, from bud break to pit hardening, large amounts of assimilates are needed to supply shoots, roots, trunk, buds and the developing fruits [3,9,24,32]. A great reduction of the current vegetative growth is always recorded because the new biomass produced is mostly driven to fruits, while the previously existing shoots did not increase in biomass [4,7,31]. After this period of pit hardening, fruit size increased, while shoot development decreased noticeably. The carbon partitioning between individual organs, reproductive and vegetative, become dependent not only upon the photo-assimilate supplies provided by photosynthesis but also upon the stored carbohydrates as well as the ability of the translocation system to deliver these resources to sinks [79]. Here, we note that leaf age and position had a major influence on the initiation of export, but one export began it was indicative of greater carbohydrate production than the leave can utilize. Assimilate translocation is greatest during the final swell of development (stage III), which was observed around the autumn active vegetative growth period beginning by mid of September. During this period, fruits have priority for carbon. Duration of this stage is reported to be correlated positively with night temperature, while duration of the previous stages, are related inversely to night temperature [6]. The ultimate development of fruits during the autumn and early winter depend on the available amount of assimilates [6,40]. Vegetative growth may be reduced substantially during this period because fruits compete strongly with sinks as it was observed in 2002 and 2003. The growing fruits may also inhibit at early stage of development, the new flower buds, which were induced two months earlier [3,32]. Consequently, the number of potential sites of fructification will be reduced as well as the following production, which was found to be correlated to shoot length (r = 0.9) [2,4,46]. This period was followed by harvest. If heavy production is obtained, shoot growth may be limited during the following spring flush and may increase when new assimilates are produced during the months following harvest. Inversely, after a low fruiting year, more water and assimilates are available, enhancing the spring, but also the autumn growth. At the end of the growing season, if temperatures during the dormant period are high enough to increase the metabolic activity, slow growth may occur during the winter months, characterizing the quiescence period.

However, we thought that application of adequate water and nutrient amounts, associated with correct pruning, can reduce this competition for nutrients and stimulate shoot elongation even during the critical periods of pit hardening and flower bud induction stage. But, the response of the trees may differ depending on variety adaptation and its ability to react face to the exogenous changes. Manzanille provided low production during the warmer years, due to stomata closure [7]. This, limit water depletion, but reduce at the same time the diurnal photosynthesis activity, leading to a more negative water potential, less assimilate production and translocation and lower growth rates. Picholine was more sensitive to hot conditions than Chétoui which presented an evident adaptation to the environmental changes. Picholine wasalso more responsive to water supply and can be considered as the best water user, since it produced high water use efficiencies. It presented good adaptation to high densities, since it assured autumn development even under high fruit load, giving at the end of the experimental period, the highest diameter canopy (4.4 m), leaf area (18 m²) and cumulative yields. Chétoui cv., yielded only 2.1 tons/ha and was affected by a strict alternate bearing which is mainly controlled by genetic factors. Meski was sensitive to severe pruning and gave very low productions due to its self-incompatibility, but an excellent quality of fruit was obtained, reaching an average weight per fruit of 6 g with 20% of fruits exceeding a weight of 8 g. Similar results were reported for cvs., Chétoui and Picholine cultivated in the same area but under different watering conditions, whereas Meski cv., behaved differently, presenting lower axis growth but a higher number of leaves [80]. This indicates that growth of vegetative organs is controlled by other factors like hormones, which may modify the response of the tree. The effect of endogenous factors was evidenced when correlating the annual increases of tree height to the climatic parameters [38]. Correlations established considering all observations were significantly different from those made with the same series but by removing the spring growth data. These aspects would be probed deeper as more data is collected on the presently mature trees.

Tree height, lateral shoot elongation, canopy and trunk diameters and roots are significantly correlated. The positive correlation observed between the area explored by roots and that of canopy (R² = 0.876) traduces the need of the tree to adequate the root system to a more vigorous canopy development. On the other hand, results showed that all parameters interfere with fruit development. Theses interferences were reported in previous works for some forest species [77]. So, and accounting for the anatomic relationships, growth of any organ may change substantially if other organs are affected, for example by pruning, in order to re-establish a new equilibrium between the canopy and the root system.

The study of the root system give us preliminary results on root distribution of young olive trees of cultivar Chétoui, which could be exploited to manage efficiently young olive orchards. Roots extended rapidly during the first 2 - 4 years of cultivation, horizontally and within the top layers. Most of them (70%) were localized in the first 60 cm of soil with only some roots developing at deeper strata, reaching 1.0 m depth. Large roots were observed beyond the first 30 cm of soil.Previous works reported similar results [31,33,53]. The root system was found to be confined to the top meter of soil, growing at depths between 0.15 m and 0.40 m and at maximum distance of 0.30 - 0.40 m from trunk. Roots of excavated trees aged 7-year were found to be contained in the shallow tillage (0.15 - 0.30 m) to approximately 0.3 m from trunk. Photographs and drawings clearly illustrate the horizontal development of the roots and the fact that roots of contiguous trees avoid competition by developing outwards from the tree row. For 12-year-old “Manzanilla” olive trees cultivated in south Spain, most roots were found to be localized in the outer layers at 0.45 m from trunk with diameter less than 0.5 mm [55]. In Create, “Kalamon” olive trees of 5-year-old cultivated showed maximum number of roots at depth of 0.4 m [57], developing numerous fine roots, which are very important because they represent the main absorbing surface [7,60]. Lateral fine roots are abundant and give rise to a fibrous root system which originate from the branching of a parent root and constitute their ramifications, generally at right angle. These studies allowed classification of roots into three main categories according to their ontogenesis: primary, adventitious and lateral [50]. In our case, the primary root constitutes the main root of the cutting. It was not dominated at the outset by a principal axis as it occurs in trees grown from seedlings. Rather, many adventitious roots were produced from the base of the cutting.

Roots explored during the first five years of cultivation an additional volume of about 1.0 m³ annually. This rate was found to be lower than those observed for three years old trees cultivated on loamy soil in south Italy (670 mm/year of rainfall), reaching volumes of 8.6 m³ for the irrigated trees and 5.1 m³ for rain fed plants [65]. In our case and for the tree of the same age, roots explored a volume of 3.65 m³ only. Soil characteristics (clay) and its mechanical resistance may adversely affect this exploration, limiting the extension of roots, which represent the plants’ evolutionary response to the spatiotemporal variability [7,23,24,50]. This explains the lateral spread of roots and the depths they achieved. During the following years (5th and 6th year), application of drip irrigation led trees to concentrate their roots close to trunk, but with a substantial increase of their density, with average values varying between 0.07 to 0.303 cm cm^-3following the depth, the distance to trunk, direction and tree age. However, these densities were found to be lower than those presented in other works [24], reporting values between 0.1 and 1.0 cm·m⁻³. Greater values of up to 0.5 cm·cm⁻³ were recorded in the first sixteen’s centimeters and at 0.4 m from trunk indicating high root activity, since low water contents were found in this area. Similar results were observed for 12-year-old “Manzanilla” olive trees, with highest value of 0.7 cm·cm⁻³ recorded at 0.45 m from trunk [55]. For 7-year-old olive trees growing with only 150 mm mean annual rainfall in Egypte, maximum root densities were observed in the top layers at 0.15 - 0.30 m and up to 0.3 m from trunk [31]. Results showed also that root density decreased significantly as distance to trunk increased (except some measurements for two and three year old trees). Roots were less frequent at all depths outside the canopy limit. In most papers, and apart from the genetic factors inherent to the potentialities of the tree itself [57], the cultural practices [7], and particularly the heterogeneous distribution of water in the orchard (amounts and frequency) [33,60] are reported as the most influent factors that affect root density. Our results showed an important spatial variability of soil moisture under the canopy. Slight differences between measurements were observed nearby the emitters (values ranging between 32% and 38% following to depth), butlarger variations were noted at the limit of the canopy. This result was unexpected and may indicate that sites of maximum root density may have low root activity as a compensation mechanism, as it was reported in a previous study [7]. Temporal variations were also reported. Some works showed that adequate watering makes roots continue to grow during the dry season, thus, increasing the period of their activity and preventing their shrinking during this period [63]. Larger volumes of water would favor the existence of wider wet bulbs and could increase root length density. In opposite, low water availability can slow down root growth because roots are able to sense the soil dryness and order stomata to close; thereby reducing water losses and preventing excessive water stress. Water shortage may also increase mortality of fine roots even in the irrigated orchards; roots developed outside the wetted area during the rainy period may die. All these results show good concordance between soil profiles made for the six experimented trees, soil volumes occupied by roots, water content and root density distributions, which were used to determine length of the root system. This parameter traduces the capacity of the tree to meet the crop water needs. In our case values varied from 1 km to 33.9 km following to age and soil volume. These values were found to be lower than those found in Spain for non irrigated (42.4 km) and irrigated (53.6 km) Manzanille trees aged 12 years [63] and for six-year irrigated Coratina olive trees (3.7 km) planted at 6 m × 3 m spacing [65].

However, despite of these differences and variations, good equilibrium between the up ground and downground organs was observed since the 4th year after planting. It is well traduced by the under/above-ground ratio, which is anindicator of the capacity of the root system to absorb water. This ratio varied from 20 to 0.9 following tree age. In Spain, values of 0.34 and 0.50 were reported for irrigated and non-irrigated adult trees [65]. The high values observed during the first year after planting indicate greater availability of water per unit of leaf area.The decrease of this ratio was found to be a consequent to some water shortage. This may lead trees to operate several physiological modifications resulting in low yields. To explain the decrease of this ratio, other reasons are evocated such as the establishment of competition for nutrients between shoots, roots and fruits, which is important to insure a balanced development of the tree once it begins to set fruits; trees with large crops have less foliage development and smaller root increases. During periods of youth and first-fruit-sets, the tree re-orientates the mobilization of carbohydrates [3,60]. So, high amounts of assimilates are leaded to the growing olives in order to insure their development. These fruits will drain essential metabolites against the competing demands of the growing roots especially during the critical stages (flower formation and pit hardening periods). Thus, the number of roots and their length can be reduced and their growth will remain highly dependent on the available assimilates. It is important to mention here that trees of cv., Chétoui begin to produce olives within the second year after planting, and on the same year an important decrease of the under/above-ground ratio was observed; the first commercial crop arrives in year four (6.5 kg/tree).

The relationship between the up ground and down ground areas is also well traduced by the root length/leaf canopy ratio. A value of 2.3 km·m⁻² of leaves was observed for the six-year age tree and it is considered as an optimum value for intensive plantations, as it ranges between 2.2 and 2.9 km·m⁻² [24-60]. This result is of great importance and show that trees ensured during the period of experiment a good development. They were able to adjust their root system to the larger above-ground development to reach equilibrium between the two parts of the tree few years after plantation. This indicates that management of water in this orchard was adequately made, despite of the difficulties involved to fix the irrigation amounts with regard to the incomplete soil coverage and the restrained area concerned by tree transpiration, i.e., where roots are active, and which increased following a logistic-shaped curve [26]. This “constraint” was taken into account to develop a mathematical model allowing estimation of irrigation needs of young trees, based on the root/canopy long time study. In this model a supply ratio was determined in order to link the water supplied to the evaporative demand that takes into account only the tree-related quantities. Results show that the ratio of applied irrigation to reference evapotranspiration during the dry season from April to August was very low, increasing from 0.02 to 0.14 when trees grew from one to six years. When using the volume method to calculate the irrigation and precipitation falling on the area covered by roots, the supply ratio becomes very close to K_c-FAO. Estimation of effective precipitation remains however big challenges for using the proposed method.

This study give us preliminary results on shoot/fruit/ root interferences for young olive trees which could be exploited to manage more efficiently intensive orchards, particularly for water and fertilizer supplies which should be given at specific distances from trunks and precise amounts to guarantee their efficacity. However, these results should be enhanced by root activity observations. It would be necessary to develop a more detailed study on young trees to get more information on the relationship between root activity and root distribution because sites of heavy root density may present lower root activity even in young trees. In such study, the root system should be viewed as a population of roots with varying, although coordinated, morphological and physiological properties. Measurements of carbohydrates status at different stages of development at both root and canopy levels would also improve these results and give us more valuable information on the relationship between the rooting system distribution and the canopy development which are essential for irrigation requirement estimation as they determine evapotranspiration and water available for the root system. More knowledge is needed on the root growth of young trees because they are more vulnerable to water shortages. For such trees, and particularly plants obtained from rooted cuttings, water uptake remains highly dependent of the effective areas of transpiration and water absorption. So it is probably more convenient to consider evapotranspiration, rainfall and irrigation in terms of volume of water/tree instead of mm. Our progress in the future will be measured by our capacity to integrate knowledge on water supply, evaporative demand and the soil volume explored by the root system for different locations and planting densities.

5. Conclusion and Recommendations

The simultaneous monitoring of tree height growth, lateral shoot length, canopy, trunk and fruit diameter increases and root extension and distribution give preliminary results on the relationships existing between the up ground and the down ground organs and their changes with time during the first years of cultivation, where trees are subject to increasing fruit competition. The obtained models give precise information on the occurrence of the various phenophases of this species during the first years of cultivation. It could be used as a reference for a quantitative appraisal of the performance of olive tree varieties under a given environment. Active growth occurs basically in April, July and September. However, important variations in growth dynamic were observed according to the year and variety. Temperature, pruning, watering conditions, fruit load, fruiting interferences and the variety potentialities seem to be the driving factors controlling this growth. Competition for assimilates between the newly formed shoots and fruits is well known for adult trees and apparently it affects young trees similarly. Adequate watering, pruning and N-fertilizer supply can reduce this competition and improve yield subsequently but the response of the trees may not be observed on all cultivars similarly. Also, it is necessary to insure a certain vegetative growth every year to obtain suitable production. In addition, there was no doubt of the necessity of continuous water availability during the early spring to the early summer period for successful flowering, promoting adequate new vegetation and reducing alternate bearing. Irrigation should be, therefore, a regular practice that confines the root system within a projected canopy area. Adequate watering may also reduce the negative effects of temperatures exceeding 35˚C in summer. Symptoms of stress may be observed if water is not supplied correctly, and does not reach the entire root system. In this case, olive trees may also develop signals of stress coming from the non-wetted roots or portion of roots and may reduce growth.

REFERENCES

Abbreviations

ET_c (mm): Crop evapotranspiration as determined by the FAO method (Allen et al., 1998).

ET^* (m³): Evapotranspiration volume of an individual tree estimated from the root area of the tree.

K_c: Crop coefficient.

K_r: Coefficient introduced in the formulae of ET_c-FAO to take into account the soil coverage.

K_supply: a supply ratio determined in order to link the water supplied to trees to the evaporative demand, it takes into account only the tree-related quantities.

ET_o (mm): Reference evapotranspiration determined following to Penman-Monteith equation.

P_e (mm): Effective rainfall determined following to the USDA-SCS method (F.A.O., 1976).

I (mm): Irrigation amount along the irrigation period.

I^* (m³): irrigation amount supplied by localized system or in small basins around the trunk.

P (mm): Total rainfall.

P^* (m³): Effective rainfall for a single tree received around the trunk.

S_c (m²): Maximum projected canopy area assuming a circular shape of the trees.

S_r (m²): area concerned by tree transpiration i.e. where roots are active.

(t): is the number of years from planting.

Lo, Lx dimension of interest respectively at planting and at maximum growthα, β are adjustment parameters within the logistic root and canopy growth curve.