Influence of spacing on the initial production of hedgerow ‘ Arbequina ’ olive orchards

A new type of olive (Olea europaea L.) orchard, based on high-density hedgerows harvested by vineyard type straddleharvesters has appeared in recent years. This system greatly simplifies olive harvesting, which has a high labour demand in standard olive orchards. There are, however, no specific low-vigour olive cultivars or dwarfing rootstocks adapted to this system, as exist in other fruit crops. It is therefore crucial to select the optimum density for maximum production without causing serious competition problems. In this study we present initial results of the effect of densities, ranging from 780 to 2,580 olive trees ha, on the production of an ‘Arbequina’ olive hedgerow orchard during the first seven years after planting. There were no significant differences in fruit characteristics (oil content and moisture) among density treatments. Density had a linear negative influence on fruit yield tree. In contrast, accumulated yield ha, for both fruit and oil yield, showed a linear increase with increased density, with coefficients of determination ranging from 0.37 to 0.72. Accumulated oil yield at high density was up to 5,000 kg ha more than at the low density. Additional key words: high density, mechanical harvesting, Olea europaea L., planting density.


Introduction
In the last few years a new type of olive orchard, suited to fully mechanized harvesting, has appeared.
Trees are planted in rows at a very high density, of around 2,000 trees ha -1 , compared to 200-400 trees ha -1 in standard olive orchards.Due to the small distance between trees, a hedge is formed 2-3 yr after planting.Vineyard type straddle-harvesters are used to collect the fruit.The main advantage of these types of orchards is the lower cost of harvesting compared to standard olive groves, due to the marked reduction in the labour needed to harvest the crop.
Although there are no reports on the area cultivated with hedgerow olives, the practise is increasing rapidly in various countries such as Spain (Navarro and Parra, 2004;Tous et al., 2007), Italy (Godini et al., 2006), Morocco, Tunisia and the United States (Berenguer et al., 2006).However, very few experiments have been reported on the best conditions and viability for this type of olive orchard (Pastor et al., 2005;De la Rosa et al., 2007;Tous et al., 2007).
Many fruit crops such as cherry, pear and apple have been planted at high densities as low vigour plant material has been developed.This is especially true for apple, which has been studied the most (Brooks and Olmo, 1997;Dennis and Hull, 2003).Different olive cultivars exhibit different vigour levels, but a true dwarf cultivar or dwarfing rootstock does not exist.The primary cultivars used in high density systems tend to be very precocious, such as 'Arbequina', 'Arbosana' or 'Koroneiki' (De la Rosa et al., 2007;Tous et al., 2007), but they can not be considered to be low vigour cultivars.Tree vigour could be a problem in very favourable growing conditions, such as that in Andalusia, southern Spain, since hedgerows must be kept below certain dimensions to allow harvesters to pass over the top of the row.
Given the lack of specific low vigour cultivars choice of an optimum density could be a critical factor in the success of high-density olive orchards.High density trees should have their highest yield in the first years of growth but are likely to suffer from mutual shading problems later on.Vigour will be more difficult to control by pruning as reported in apple (Hampson et al., 2004).Recent studies on olive hedgerow orchards have shown the importance of planting design (orientation, row height, row width and canopy slope) to identify combinations that maximize incident solar radiation interception and, thus, productivity (Connor, 2006).However, there is little information on the influence of olive tree density on yield, and the few reported studies were conducted in standard olive orchards (Pastor and Humanes, 1990;Tous et al., 2005).
This work reports on the influence of olive tree spacing on production during the first seven years from planting of a high-density hedgerow olive orchard grown in the South of Spain (Andalusia).

Material and Methods
The trial was located in Pedro Abad, Cordoba, Spain (longitude 4°27' 37'' W, latitude 37°57' 43'' N and altitude 155 m) with an average rain fall of 477 mm year -1 .Average mean, maximum and minimum temperatures are 17.2, 24.5 and 10.2°C respectively (Gavilán et al., 2006).Ten densities treatments, ranging from 780 to 2,581 trees ha -1 , were evaluated in four individual blocks.Different densities treatments were obtained by varying the distance between rows from 5.7 to 3.1 m and between trees in the row from 2.25 to 1.25 m (Table 1).Each plot (per treatment and block) consisted of a 40 m long row, with 18 to 32 trees depending on density.The cultivar 'Arbequina' was planted in March 2000 in a north-south hedgerow orientation.Irrigation applied was 1,800-2,300 m 3 ha -1 yr -1 .Trees were trained to a monocone shape (central leader), eliminating vigorous lateral branches to allow unimpeded access of the harvesting machine every year.Phytosanitary treatments against Spilocaea oleagina and Margaronia unionalis were applied with a maximum of 6 yr -1 .
From 2002 to 2006 yield and fruit characteristics (oil content and moisture) were measured.This was three to seven years after planting.In 2003, unusually heavy rain prevented harvest of the trial and data collection (4 th year after planting).The fruit were subsequently removed to avoid interference with following year's production.The olives were harvested with a vineyard-type harvester model «Gregoire 120» (2.60 m high and 0.8 m wide) and were measured for each row of each treatment and block.Three 30 g random samples of olives were collected from the crop from each row to determine fruit characteristics.Samples were dried in a forced-air oven at 105°C for 42 h.Dried samples were weighed to determine moisture content and then oil percentage was measured in an NMR analyzer Minispec NMS100 (Bruker Optik GmbH, Ettlingen, Germany) following Del Río and Romero (1999).ANOVA and regression analyses were performed to determine the influence of density on production characteristics using Statistix (Analytical Software, Tallahassee, FL, USA).

Results
There were no significant differences in fruit characteristics among densities in any year of the study (Table 2).However, differences among years were highly significant (P < 0.05).
In all cases there were significant linear regressions between density and yield.Increased density negatively influenced yield tree -1 , with coefficients of determination ranging from 0.37 to 0.56 in the years under study (Table 3, Fig. 1).In contrast, yield ha -1 increased linearly with increased density with coefficients of determination ranging from 0.42 to 0.72 for fruit yield (Table 3, Fig. 2) and from 0.41 to 0.69 for oil yield (Table 3, Fig. 3).
Lower densities had higher cumulative fruit yields tree -1 (Table 4) but cumulative fruit yield ha -1 was higher at the higher densities.The relation between density and cumulative oil yield followed the same A Letters indicate significant differences between years for each characteristic (Tukey, P < 0.05).
Table 3. Regression analysis of the production response to density for the 3 rd , 5 th , 6 th and 7 th year from planting.Parameters of intercept and density (trees ha -1 ) are indicated, as well as coefficient of determination.For all traits and years the model was highly significant (P < 0.001)

Discussion
This paper reports the initial results of the first olive density trial in a hedgerow orchard.During the seven first years after planting, it seems that there was no strong effect of competition among trees in the higher density treatments so their lower yield tree -1 was compensated for the higher number of trees ha -1 .A significant linear increase in yield with increased density was maintained for seven years after initial planting.
Similar yield:density correlations have been obtained for nine years in apple (Hampson et al., 2004).In other medium-and long-term spacing trials in both apple and peach, high density treatments were also the most productive (Westwood et al., 1976;Marini and Sowers, 2000).The absence of a significant difference in fruit characteristics among densities was similar to results on apple fruit characters density trials (Widmer and Krebs, 2001).
The accumulated yield of 51 Mg ha -1 of olives, at the highest density, was higher than the approximately 36 Mg ha -1 of olives, for the same period and cultivar, grown at similar spacing in Catalonia (Tous et al., 2007).This may reflect more favourable growing conditions for olives in Andalusia, especially given that yield would have been higher still if the data for the 4 th year after planting was included.In other olive spacing trials, at lower densities, there was an increase in yield with density in the first 9-10 yr after planting (Pastor and Humanes, 1990;Tous et al., 2005).
In summary, seven years after planting, the expected stronger competition and mutual shading at higher densities had not negatively influenced olive production and oil content.It is, however, essential to verify if, in coming years, it is possible to control tree vigour, especially in high density treatments, to avoid problems produced by shading and such diseases as Spilocea olegina and Verticillium dahliae.An important question is what density is required to combine a high early yield with stability to maximize yield in the long term (15-20 yr in hedgerow orchards).The present data give an idea of the minimum number of plants ha -1 , and therefore the minimum investment for maximum  production in this new type of olive orchard.Future work will require further testing in blocks which are large enough to include the competitive relationships between trees in a commercial planting.

Figure 2 .
Figure2.Time trend of olive production ha -1 in response to tree density in the 3 rd , 5 th , 6 th and 7 th year after planting (regression lines are indicated).

Table 1 .
Olive tree densities used in the experiment

Table 2 .
Response of fruit characteristics to tree density of the 3 rd , 5 th , 6 th and 7 th year after planting.Differences among density treatments within years were not significant (P > 0.05) Influence of spacing on the initial production of hedgerow Arbequina olive orchards 557 Time trend of olive production tree -1 in response to tree density in the 3 rd , 5 th , 6 th and 7 th year after planting (regression lines are indicated).Time trend of oil production ha -1 in response to tree density in the 3 rd , 5 th , 6 th and 7 th year after planting (regression lines are indicated).