A comparison of three different cooling systems in parral type greenhouses in Almería

In warm climates, high temperature can limit growth and development of greenhouse crops and their product quality. Therefore greenhouse cooling has a high-priority to reduce these adverse effects. Together with natural ventilation, shade produced by whitewashing the greenhouse roof is the most usual cooling method in the whole of the Mediterranean area. However, this technique is not homogeneous; it is not selective and not adjustable. This inconvenience can be overcome by the use of other techniques such as shading with folding or rolling screens and/or evaporative cooling. This work evaluates three cooling techniques in parral greenhouses using a sweet pepper crop: whitening, shading with a mobile internal screen and cooling with a low pressure fog system. The latter gave the highest reduction in air temperature in the greenhouse in warm periods. However, final water consumption in the fog system was 319 mm, whereas irrigation water consumption was 520 mm. The evaporative cooling system, with 6.5 kg m of marketable production and the use of interior shade screens (with 6.7 kg m), did not increase the yield obtained by means of traditional whitening (7.1 kg m) used by growers in the area. Additional key words: refrigeration, shading screen, whitewashing.


Introduction
In Southern European countries, the production period of vegetables is flexible, as it mainly depends on the local climate and on economic factors (Nisen et al., 1990).Over the last few years, greenhouse pro-duction periods are being extended.This involves growing during the summer, when diurnal solar radiation can cause harmful temperature increases for crops developing in the greenhouse (Day and Bailey, 1999).This affects both crop yield and quality (Kittas et al., 1996).Greenhouse horticultural production in Almería is characterized by the use of simple low cost structures, with limited climate control (Lorenzo, 1998).In these greenhouses, temperature control is restricted
to management of natural ventilation to limit extreme values of humidity and temperature (Abreu and Meneses, 1994;Abreu et al., 1994).Natural ventilation is the most practical, economic and therefore the most used method to lower greenhouse temperatures during the day.Most of the greenhouses have manual ventilation systems, 90% of them have side vents which open and close by simple sliding of plastic film and 31% of the greenhouses do not have roof ventilation or have a low eff iciency (36%) sliding type (Fernández and Pérez Parra, 2005) et al., 2001).
The general use of low porosity insect screens on greenhouse vents to limit the entrance of small insects such as Bemisia tabaci and Frankliniella occidentalis, (mainly) which transmit viral diseases, decrease even more the ventilation area.Thus, natural ventilation is not sufficient to extract excess energy from the greenhouse during sunny, summer days (Baille, 1999).Summer conditions for crops in these greenhouses are far from optimal, especially in relation to temperature and vapour pressure deficit.For this reason, growers also use shade by whitewashing the greenhouse roof to reduce the amount of radiation entering the greenhouse.The combination of both «natural ventilation and whitewashing the roof» is the most common method used to cool greenhouses in summer.
However, whitening the roof is inconvenient.It can not be removed on cloudy days; it is not applied evenly, which results in different amounts of light reaching plants in different parts of the greenhouse; labour is required to apply it and to remove it; it is also not selective, as approximately the same percentage of photosynthetically active radiation (PAR) and near infrared radiation (NIR), responsible mainly of heat, is transmitted (Montero et al., 1998).For this reason, other cooling systems, such as forced ventilation, reduce that incident radiation by means of folding shade screens or use of evaporative cooling systems (pad and fan, fog systems) can, potentially, be more efficient alternatives to control high greenhouse temperatures.
The main objectives of this work were to evaluate the effect of a low pressure evaporative fogging system, a folding screen shade system and the common roof whitening on the greenhouse climate and their influence on the production of a bell pepper crop, in a parral type Almería greenhouse.

Material and Methods
The experiment was performed at the Cajamar «Las Palmerillas» Experimental Station at El Ejido (Almería) at an altitude of 155 m, 36º 47 ' 40" N and 2º 43' 10" W, during autumn 2002-2003.Three analogue multispan parral greenhouses were used for the experiment (Pérez Parra et al., 2004).Each greenhouse had 5 spans with their ridge oriented northsouth, ridge height was 4.2 m, gutter height 3.3 m and a roof area of 882 m 2 ; automated roof and side vents were protected with 20 × 10 thread cm -2 insect screen.Greenhouse cladding was colourless three-layer plastic film with a thickness of 200 µm.
The crop was a red bell pepper (Capsicum annuum L.) cv.Vergasa (Syngenta Seeds).Seed was sown in a nursery on 11 June 2002 and seedlings were transplanted to the greenhouse on 15 July 2002.The growing cycle f inished on 6 March 2003.The total duration of the growing cycle was 232 days.In Almería California type sweet peppers are transplanted to greenhouses between early June and early August, and the crop cycle ends between late January and late February.
Crop rows were 1.9 m apart with 0.25 m between plants.Final plant density was 2.1 plants m -2 .Plants were pruned to leave three main stems per plant, giving a density of 6.3 stems m -2 («Dutch» type trelling).
The crop was grown in B-12 (0-5 mm granule size) perlite in 40-L bags laid over a polystyrene channel to collect drainage.
The following cooling treatments were compared: -T 1 : a low pressure fogging system comprising a pump unit and a water distribution net.The pump unit included filters and a pump (giving a pressure of 4 atm).Water was from a rainfall reservoir.There were 5 fogging lines in each greenhouse, N-S oriented, each line was 4 m from the other.Lines were polyethylene pipe with fogging nozzles with an average flow of 7 L h -1 separated 1.5 m (0.16 nozzles m -2 ).A vapour pressure deficit (VPD) set point of 1 kPa was set.
-T 2 : aluminium internal folding shade screen (ULS 15 F, Ludvig Svensson), 50% shade and 20% energy saving, made of aluminium sheets with open spaces held together with strong polyester filament yarn.The special aluminium gives superior reflection and transmission efficiency, while the open spaces allowed sufficient airflow to give a considerable reduction in air temperature.The screen was 2.8 m above the green-house floor.The temperature set for screen activation was 27ºC.
-T 3 : whitening the greenhouse roof by applying calcium carbonate dissolved in water (Blanco de España) at 25 kg for every 100 L of water.The lime was applied using the normal local technique.It was applied on 14 July 2002 and washed off on 7 October 2002.
Dry and wet bulb temperatures were measured inside and outside the greenhouse with ventilated psychrometers with Pt-100 sensors (Fig. 1).The VPD was calculated from these measurements.Each greenhouse had two ventilated psychrometers, one at 1.5 m and the other 3.5 m above the ground.Thus in the folding screen treatment greenhouse there was one psychrometer above the screen and one below (Fig. 1).
Climate control and management was done using 30 s measurements averaged every 5 min.
The transmissivity of the covering material to PAR radiation was determined as the ratio between incoming radiation inside the greenhouse and the outdoor radiation, from an average of 5 measurements in an East-West direction in each greenhouse, using a linear sensor (LICOR Inc, Lincoln, Nebraska, USA).These measurements were taken on sunny days at noon [12:00 Greenwich Mean Time (GMT)].
To determine treatment effect on crop production, a one-factor experimental design with three treatments (T 1 , T 2 and T 3 ), five repetitions per treatment, and 16 plants per repetition was used.
Both marketable and non-marketable yield was determined at each harvest.Fruits were also classified into different categories, using precision scales (mod.Metler Toledo deviation of ± 1 g), according to the sweet pepper quality standard (OJ, 2000).

Climate
The average temperatures during the day (Table 1 and Fig. 2), measured by the psychrometer at 1.5 m above the ground, were 22.4 ± 4.4ºC for T 1 , 23.2 ± 5.6ºC for T 2 and 22.6 ± 4.9ºC for T 3 ; all values were higher than the average temperature outdoors (T out ) of 20.9 ± 5.4ºC.The maximum temperatures, at this height, were reached during summer.The hottest day was 5 August, with the following maximum temperatures: T out : 35.5 ± 4.4ºC; T 1 : 36 ± 4.1ºC; T 2 : 43.3 ± 6.2ºC and T 3 : 36.8 ± 4.5ºC.
During the period when the fogging system was used to maintain the VPD set point (until 90 days after transplanting or DAT), the average relative humidity during the day for T 1 (74 ± 7.4%) was higher than in T 2 (56.3 ± 10.3%) and T 3 (61.9± 9.7%) (Table 1 and Fig. 3).
Figure 4 shows change in temperature and VPD on a typical summer day.The temperature reached at the middle of the day for T 1 was 1.5ºC and 3.3ºC lower than in T 3 and T 2 , respectively.Something similar occurred with VPD values, T 1 kept values below 2 kPa.In the other two treatments values above 3 kPa and even up to 4 kPa, in T 2 , were reached.These values occurred at the start of the crop cycle, when the plants had a very low leaf area index and crop transpiration was very limited.
Table 2 shows the average temperature jumps in relation to the outdoor temperature at heights of 1.5 and 3.5 m above the ground, and for two outdoor wind conditions (Ve = 0-2 and Ve = 5-8 m s -1 , where Ve is the outside wind velocity) for days with different wind velocities at noon (GMT) for the period with whitewash (0-83 DAT).At both 1.5 m and 3.5 m the highest tem-  perature differences were always in the folding screen treatment, and ranged between 2.7ºC and 11.2ºC, respectively.Only the fogging treatment (T 1 ) decreased the greenhouse air temperature in relation to the temperature outdoors by up to -1.2ºC at 1.5 m.The whitening treatment results were intermediate between the other two treatments.The temperature difference between the inside and the outside (T i -T out ) was between 2ºC and 4.1ºC.
In relation to PAR during the time the whitening was on the roof (0-83 DAT), the transmissivity of T 1 was about 52%, in T 2 with 95% extended screen it was 21% and in T 3 it was 28%.

Production
The first peppers were harvested at 71 DAT and the last at 232 DAT.There were a total of 21 harvests.
The higher total yield, for the whole growing cycle, was from T 1 at 8.7 kg m -2 , followed by T 3, 8 kg m -2 , and T 2 at 7.4 kg m -2 .The differences were statistically signi-Table 1.Diurnal average temperature and relative humidity and standard deviation (±) for the daylight period inside greenhouses for three cooling treatments and outdoor over Period 1 (up to whitewashing, 0-83 DAT), over Period 2 (from whitewashing, 84-232 DAT) and over the full crop cycle (0-232 DAT) Figure 2. Change in the average diurnal temperature (ºC) for the three treatments: fogging system (T 1 ), shade screen (T 2 ), whitewashed greenhouse roof (T 3 ) and outdoors (T out ).
ficant (P < 0.05).However, there was no difference in final marketable pepper yield among the three treatments (mean 6.8 kg m -2 , Table 3 and Fig. 5).Differences in fruit quality favoured T 3 (60.6%I Category) and T 2 (61.1% I Category) in relation to T 1 (44.1% I Category).T 3 was significantly different to T 1 .The proportion of non-marketable was statistically higher (P < 0.05) under fogging (2.2 kg m -2 ) compared to whitening (0.9 kg m -2 ) or the shade screen (0.7 kg m -2 ).Plants from T 1 had higher early total and marketable production than in the other two treatments, 3.3 and 2.9 kg m -2 (at 120 DAT) of total and marketable peppers respectively, compared with 2.8 and 2.7 kg m -2 for T 3 and 2.1 kg m -2 for both total and marketable peppers in T 2 (Fig. 5).

Climate
Treatment T 1 was best at decreasing glasshouse temperature in relation to T e (∆T of 1.5ºC against 2.3ºC in T 2 and 1.7ºC in T 3 ).These results are lower than those of Perdigones et al. (2004), who obtained temperature differences (∆T) of -0.8ºC using a low pressure Hourly change in daytime of (A) temperature (ºC) and (B) Vapour Pressure Deficit (VPD) for the fogging system (T 1 ), shade screen (T 2 ), whitewashed greenhouse roof (T 3 ) and outdoors on a typical summer day (30 th July).Average wind velocity VeϷ1.8 m s -1 .

A B
fogging system and a ∆T of 1.7ºC using a mobile aluminium internal shade screen at 65% shade.Francescangeli et al. (1994), in tunnel type greenhouses, in which incident radiation was reduced by whitening and with a shade screen, reported differences of 2-3ºC compared to a control greenhouse.At high radiation levels, the temperature of the aerial part of the plant, directly exposed to the sun, can be up to 10ºC higher than the surrounding air temperature (Van Holsteijn, 1998).This can induce excess temperature stress damage, irregular fruit development and yield loss. Sáez (2005) compared leaf and air temperatures (T leaf -T air ) of a pepper crop using a high pressure fog system, forced ventilation and whitening, and obtained a positive difference in the fog system greenhouse.Leaf temperatures were higher than air temperatures, which could be interpreted as a stress symptom.Montero et al. (1981), working with a tomato crop, concluded that under sunny conditions leaf temperature was lower than greenhouse air temperature at humidity levels below 80%, but it was significantly higher at levels approaching saturation.The most probable explanation for this was that at high humidity, transpiration could have been limited and leaves were unable to be cooled as much as under a low humidity.
Recorded temperatures in T 2 were higher than in the other two treatments, due to limited air movement, and thus air renewal in the greenhouse, when the screen was extended.This was opposite to the desired effect, especially on days with low winds.Gómez (2001) determined that the aluminium screen shade systems when placed in a greenhouse were no more effective in decreasing daytime temperatures than traditional whitening.Fernández et al. (2003) also argued that the use of aluminium shade screens as to prevent heat stress in multispan parral type greenhouses to reduce air temperature was not as effective as whitening.
The influence of the wind on the temperature differences (T i -T e ) differed among the three treatments: -Whitening: low wind velocities gave a temperature gradient of 2.1ºC between a height of 1.5 and 3.5 m.When V e was higher (5-8 m s -1 ) the temperature gradient decreased to 1ºC.Sánchez (2002) also reported a vertical temperature gradient in a whitened greenhouse of 1ºC.
-Screen: with low wind velocities V e < 2 m s -1 , the vertical temperature gradient was 8.5ºC, due to decreased air renewal caused by the internal screen.When the V e ≈ 5-8 m s -1 the gradient was reduced to 2.5ºC.Sánchez (2002) found that internal shade screens caused a high vertical thermal gradient in the greenhouse with temperatures up to 4.6ºC higher above the screen than below it.
The fogging system was not able to maintain the VPD set point of 1 kPa in T 1 , during high demand periods and gave a f inal water consumption of the fogging system of 319 mm, whereas water consumed for irrigation was 520 mm.It was also observed that the low pressure fogging system, at times, wet the crop at certain moments, due to larger water drops.Montero et al. (2003) suggested that the VPD set point in a Mediterranean climate should not be below 1.5 kPa to avoid wetting plants.Fernández et al. (1998) reported a similar transmissivity value (c.31%) in a parral greenhouse with similar whitening.

Plant production
The T 1 treatment gave higher total and non-marketable pepper production than the other two treatments (Table 3).This was mainly due to the large number of deformed and parthenocarpic peppers.Treatment T 3 produced the most marketable peppers compared with T 1 and T 2 .Aroca (2003) also reported higher production of marketable pepper with whitening compared with a high pressure fogging system.Abreu and Meneses (2000) compared the effect of whitening on the yield of tomatoes in a greenhouse during the spring crop cycle in Portugal and found the control treatment (no whitening) gave the highest total production.
Treatment T 1 was the most precocious (2.9 kg m -2 at 120 DAT), while T 2 treatment only produced 2.1 kg m -2 .In many horticultural crops, flower retention by the plant, and fruit development are extremely sensitive to environmental stress.Aloni et al. (1996) showed that flower abscission in a pepper crop was increased under low light and high temperature conditions.This had a negative effect on production.

Conclusions
The fogging system was the most effective in controlling high greenhouse temperatures.The fogging system also efficiently maintained relative humidity and VPD values.
Interior folding screens were not efficient in controlling high temperatures (maximum ∆T 7.8ºC) and caused undesirable thermal stratification which greatly affected greenhouse roof natural ventilation.Plants grown under the fogging system were more precocious and had a higher final total yield but quality was negatively affected.This gave lower marketable pepper production than in the whitening treatment.Shading the greenhouse by folding screens gave the worst production.
. Moreover, the ventilation open area is insufficient (i.e.far below recommended literature values of 25-30% of open area in relation to the area covered by the greenhouse (Okhushima

Table 3 .
Total, marketable and separated production categories for the whole growing cycle (0-232 DAT).Values followed by a different letter are significantly different (p < 0.05)