Price trends in greenhouse tomato and pepper and choice of adoptable technology

Currently, intensive horticultural production in the Spanish Mediterranean region is strongly influenced by the liberalization of markets, which means farms must become more competitive and focus on consumer demands, such as product quality and health. The goals of the present work are to determine how the current economic situation of the tomato and pepper are reflected in the origin price, and to carry out an assessment of the technology available in the greenhouses. This study focuses on three principal Spanish regions of greenhouse horticultural production, located in Almería, Murcia and the South of Alicante (South-East Spain). Trend and seasonality of origin prices have been determined, and the ARIMA models have been used to predict pepper prices in Almería. Our analysis resulted in the following: the development of purchasing power, based on the price trends of the studied varieties; seasonal indices are developed to assist decision-making with respect to production schedules, choice of varieties and use of heating; finally, prediction is used to assist in planning the productive and commercial activity for the following season. There are two major determinants of technology adoption which include the fundamental elements of the equipment (heating and substrate) and the different production modes, like integrated production in different certified modes, and ecological production. Additional key words: ARIMA models, greenhouse equipment, intensive horticulture, price analysis, time series.


Introduction
The current situation of European horticulture is highly conditioned by the liberalization of markets, as well as by demands for quality, traceability and environmental sustainability factors imposed by big retail chains and consumers, which obliges farmers to keep competitive farms.
To respond adequately, farmers must increase the technological level of their farms, improving structure and equipment in the greenhouse, increasing levels of technical and biological management, extending the degree of environmental control, and even adopting fundamental changes like soilless culture.
The growth in horticultural production in countries in the Mediterranean region, including Morocco, Egypt or Turkey, and ease of export to European Union (EU) markets, oblige Spain to strengthen its competitiveness.The main way to do so is to increase the level of technology, given that the labour costs are much lower in the aforementioned countries.
Adopting technology is fundamentally a business decision, impacted by technical, social and economic factors.The main aim is to increase the potential profits and reduce risks.
The demand for technical changes includes modifying productive processes (for example, those that are imposed by demands like disinfection of soil and virosis) and the characteristics of technology and equipment (structure and greenhouse equipment, plant material, etc).Social aspects encompass numerous factors, which affect family issues and those of socio-economic character, like the variation in market structure and consumer preferences.Finally, among economic factors, prices are the most relevant.
The prices farmers receive for their production directly influence the technology they adopt on their farms.On one hand, they affect the choice of production techniques, varieties, schedules and, consequently, the most suitable equipment for each greenhouse.On the other hand, prices determine income and define income levels for agricultural producers as well as their incentives to invest in new technology.
The main objectives in this research are to determine how the current economic situation of the tomato and pepper are reflected in the origin price, and with respect to the information obtained about prices, make an assessment of the technology that can be adopted in Spanish greenhouses.
To develop these objectives, we analyze the price series over a wide time-period, which allows for reliable predictions to be made.Price analysis contributes to the rationalization of business decision-making, which is based on concepts, results and predictions about information concerning a long time-period, rather than precipitously and founded on conjunctural situations.
This study considers two main horticultural species grown in the greenhouse in Spain, namely: tomato (Solanum lycopersicon L.) and pepper (Capsicum annuum L.).The reference point of this information is the main areas with greenhouse production in the Mediterranean, located in Almería, Murcia and the South of Alicante.

Technological development in greenhouses
The development of greenhouse structure and equipment has occurred gradually.The first changes involved the move from wooden structures to metallic profiles, followed by improvements focused in achieving good climatic control, which is essential to increase the yields and quality and vary the trading season.
The development of greenhouses in Spain was mainly based, during the initial years, on the installation of a type called «parral» employing a trellis hanging system, of low cost.The very rapid expansion of the covered surface area was founded on this type of flat or «shed» greenhouse (Caballero et al., 2001), which has undergone continuous modifications dictated by technological requirements (Molina et al., 2003).
From the 90s onward, this type of greenhouse, using a trellis hanging system, has been widely used in the Spanish Mediterranean, and has been developed and improved to correct its numerous drawbacks.The main modifications in shape and structure affected the roof slope, installation of roof vents and automated control of opening and closing.Different procedures are used to reduce high temperatures, as well as using whitewash, humidification by means of evaporative panels (cooling system) and misting, which can have two modes: high pressure (fog system) and low pressure.In addition to the aforementioned procedures shading screens are often employed.To guard against cold temperatures moveable thermal screens, double-layered covering and heating systems are used.The inflated double-layer covering reduces heat loss; however, its drawback is that it reduces solar radiation influx and, therefore, one must avoid condensation forming inside the air chamber (Castilla, 2005).
Additional elements and techniques can be added to the previously described modifications in structure that complete and perfect its functioning and the productive process in the greenhouse.The main ones are: heating, soilless culture, irrigation and ferti-irrigation, informatics applied to agriculture, carbonic fertilization and recirculation of nutrient solutions.
Heating can take two forms: hot air, to avoid momentary drops in temperature and avoid the risk of frost, and hot water, which maintains a stable temperature and enables the productive cycle to be brought forward.In response to the increase in cost of the most frequently used fuels (diesel, natural gas or propane), the current trends, which demand lower unitary costs, are moving toward a greater consideration of alternatives, like their substitution by bio-fuels, use of solar energy, and micro-tunnels inside the greenhouses, among others.
From the technical viewpoint, the adoption of cultivation in substrate represents a very drastic change for the farmer, who must have suitable training and make a strong investment in the necessary greenhouse equipment (Caballero and Fernández-Zamudio, 2006).As well as the issue of proper technical management of soilless culture, one must also add the economic and environmental attitudes that currently prevail in food production (Fernández-Zamudio et al., 2006).
Irrigation and ferti-irrigation are mainly applied by high-frequency drip-irrigation systems, with due planning based on electronic and informatics technology.Informatics is applied to agriculture to achieve the ideal management practice of the crop when it is necessary to use control systems and to help in decision-making and make the best choice of environmental conditions, growing techniques and crop planning (Baille and González-Real, 2001).Currently, there is the extra advantage of being able to connect these systems to Internet, which enables control at a distance.
Carbonic fertilizer application attempts to increase CO 2 levels in the greenhouse, thus increasing photosynthesis and consequently an increase in production.Opinions put forth by technicians do not coincide with respect to the economic advantages, particularly when it is necessary to frequently keep windows open given the need for ventilation.
Nutrient solution recirculation is a possibility with soilless culture, which is a great advantage from the ecological viewpoint and can also be economical.
As well as the modif ications to the greenhouse structures and to the elements that must be placed inside them, productive techniques or cultivation systems also play a decisive role.Their importance lies in the fact they do not require high fixed capital investment, rather they represent the substitution of capital by production techniques, but tend to have an important effect on the prices.Integrated production (IP) and ecological production (EP) should be highlighted as fundamental.Both of these production types require the installation of anti-insect netting, which are indispensable for the biological control.The reduction of pesticides as a consequence of IP techniques, or their almost total elimination in EP, have led to useful insects being widely used.This, as well as grafted plants, an aspect that was resolved some time ago in tomato cultivation and more recently for pepper.

Information and methodology
It is common to use historical price registers to construct a time series in economic studies, within a wide variety of analytical topics, which focus on evaluating previous periods and current time, or making predictions.
To analyse the historical development as well as the current situation, trend and seasonal indices are applied in this work.The autoregressive integrated moving average (ARIMA) models have been used to make predictions, from the series of origin prices.
An interesting treatment of trend analysis is that of Makridakis and Wheelwright (1978), whose method has been incorporated into calculus programs (Statgraphics).Given their accuracy and simplicity, the trend has been used to calculate the variation in agricultural income by means of the prices, within the period of years that the series spans.With this objective, the works by Caballero et al. (1992) andDe Miguel et al. (1994) regarding different fruit and vegetable species, and Caballero and Fernández-Zamudio (2003) applied to the loquat are remarked.Other applications of the trend can be found in Hinojosa et al. (2003), who evaluated the international markets of cocoa and organic banana and Arias (2002) who evaluated tractor sales in Spain.The methodology on seasonal indices has been developed by numerous authors, among whom Lange (1974) and Uriel (1995) stand out.The works of Benedicto and Caballero (1983), Caballero et al. (1992) and Noguera (1996) used the determination of seasonal indices of price series in fruit and vegetable growing; while Guigó (2004) applied it to the prices of barley.
ARIMA models are widely used in the field of forecasting prices and there are numerous publications applying this method and checking its validity.In terms of methodological contribution the works by Box and Jenkins (1976) and Peña (2005) are remarked.To apply this methodology to empirical studies of time series in agriculture, some highlighted works include studies by Albisu et al. (1984) in the wine sector, Romero and Balasch (1986) in the evaluation of the effectiveness of the struggle against hail, Cases (1994) in citriculture, Douvelis (1994) in predicting the prices of sunflower seeds in the USA.Within the production area, Arias and Alonso (2000) focus on predicting the surface area sown with beetroot in Castilla-León, while Chambers (2004) forecasts the price of grain for animal fodder and Najeeb et al. (2005) predicting the area and production of wheat in Pakistan.
The information gathered takes as reference the main regions of greenhouses on the Mediterranean, located in Almería, Murcia and the South of Alicante, and encompasses the records of origin prices that figure in the periodical publications of the Agricultural Councils of Andalucía, Murcia and Comunidad Valenciana (Junta de Andalucía, 2006; Gobierno de la Región de Murcia, 2006;Generalitat Valenciana, 2006).To fully develop this work, technical and economic information was collected in the regions under analysis, by means of interviews made at farms and agricultural cooperatives.
Prices were taken fortnightly at the periods indicated in Table 1.The types studied were ribbed and round tomatoes, which in Almería coincide with the long-life variety, as well as green and red peppers.
The computer software Statgraphics Plus version 5.1 has been used to carry out the calculations for the three concepts included in the concepts: trend, seasonal indices and prediction.
Once the time series were constructed, with the updated periods and dates indicated in Table 1, the following determinations were carried out: a) Trend analysis, which shows the economic behaviour of the different species with respect to the prices and the current perspective.When the trend is growing, this gives very sure perspectives; in the event it is decreasing, it is worth carefully checking whether it is due to a slow decline in the purchasing power of a particular variety, which although inevitable is bearable, or whether it is a sharp and unrecoverable drop given the technical and economic characteristics that affect it.Finally, a very common situation is one concerning new variety that begin at very high prices in the first years, giving rise to extraordinary profits, and drop to normal profit margins in the following years.
The trend expresses the general movement of a series in the long run.For this analysis, the movement is separated from the cyclical, seasonal and residual variations that, according to classical theory, are considered in a time series.The values of the series have been time-adjusted using a linear equation, which although being the simplest, satisfactorily covers the trend of the price series.
b) The seasonal indices show the relative level that prices have reached, with respect to an average value 100, in the seasons that are encompassed in the series.With these data, it is possible to optimize the income, taking into account the expected productions, and consequently, based on climatology, greenhouse characteristics and available technical means.
To obtain the seasonality indices, a seasonal break down has been carried out on the original price series according to the multiplicative method.Following this procedure, the seasonal variations (E ik ) can be isolated from trend components (T ik ), cyclical variations (C ik ) and residual variations (R ik ) (Pérez, 2002).This leads to the result that Y ik = T ik * E ik * C ik * R ik , where Y ik is Table 1.Periods corresponding to each series according to crop type and zone, indicated in fortnights (f.)
The ARIMA models, with prediction for one year, can be more useful for commercial horticultural crops, where short-term decisions are important.Likewise, these models offer the possibility of providing, for the coming season, planning regarding the techniques and production factors that are most advisable to use, like the choice of variety, production schedule, disinfections, grafting, etc.
In the identification and estimation of the ARIMA models, the methodology developed by Box and Jenkins (1976), which is used to analyse univariate models of time series, was followed.In these models, a dynamic variable that fluctuates over time (in the current case, the price), is explained using just one variable, the historical record and its past values (Uriel and Peiró, 2000).
To apply the ARIMA methodology, it is essential that the time series is stationary.To do this, it is necessary to know a series of terms, which are defined below.
A stochastic process is a succession of random ordered variables Y t , where t can have any value between -∞ and +∞, and represents the passing of time, when dealing with time series, as in the case.The stochastic processes can be defined specifying the average and the variance for each Y t .Of interest to us are: -The stochastic process with an average zero, constant variance and where the random variables are independent for different values of t (covariance null).This stochastic process is known as white noise.
-The stationary stochastic process, which occurs when the distribution functions together are invariate with respect to a time lag.
In the univariate ARIMA models one works from the initial consideration that the time series has been generated by a stochastic process and specifically the elaboration techniques of the ARIMA models try to identify the model generating the observations which, once estimated and validated, in an iterative process, are used to predict future values of the time series (Uriel and Peiró, 2000).
Below is the description for a general Box-Jenkins model with a stationary series: Being L the lag operator, so that, multipling L by the variable in t, the value of this same variable in t-1 is obtained; d is the number of differences to change the series into a stationary one; λ is the parameter of the Box-Cox transformation (they take values 0 and 1); (φ 1, φ 2,..., φ p ) are the autoregressive coefficients and (θ 1 , θ 2,..., θ q ) are the moving averages coefficients.
This equation is usually abbreviated using the ARIMA terminology (p,d,q), where p is the degree of the autoregressive part; d is the number of differentiations in the regular part and q is the degree of the part of moving averages.
A model is defined as autoregressive if the variable Y t can be expressed as a linear combination of its past values adding an error term.
The general expression of an autoregressive model would be as follows: where a t is the sequence of deviations distributed identically and uncorrelated, called white noise, which corresponds to the ready defined stochastic process with average zero and constant variance.
A model of moving averages explains the value of a specif ic variable in a period t in terms of an independent term and a succession of errors that correspond to the preceding periods, suitably weighted.Thus, a model with q error terms is expressed as follows: In time series with a seasonal component, the appropriate model is an ARIMA (p,d,q) × (P,D,Q) s with the general formula: where s is the seasonal period, P is the degree of the autoregressive part of the seasonal component, D is the number of differentiations in the seasonal component and Q is the degree of the part of moving averages of the seasonal component.

Applying the ARIMA model in the calculus
To look for an ARIMA model for a particular time series, the following steps are taken.

Identification
First of all, the series must be stationary.There are two forms of stationarity: -Stationarity on average: if the series is nonstationary on average, one must differentiate the series until the trend disappears.The number of times that it is differenced is the value of d.
-Stationarity in variance: by means of the observation of the graphic representation of the series, one can see whether or not the dispersion is maintained around the existing level at each moment.If one observes that the level increases in line with the increase in dispersion, it is said that there is stationarity in variance, and one must carry out the Box-Cox transformation.In this way, a natural logarithm must be applied to the series if α = 1, when the relationship between typical deviation and the average is: σ = k*mα , and thus establish variance.
The terms p and q are determined by using the functions of estimated autocorrelation and partial autocorrelation of the series.Next, we compare them with the functions of autocorrelation or correlograms of the following theoretical models: a) autoregressive model of order p [AR(p)]; b) moving average model of order q [MA(q)]; or c) autoregressive moving average model, with p autoregressive terms and q moving average terms [ARMA(p,q)].In this way, an ARIMA model can be geared towards the appropriate time series that most closely identifies the data generating process.

Estimate
If the process is identified, the parameters of the autoregressive part and of the moving averages part are estimated for both the regular and seasonality components.
By contrast, if the process is not well identified, a new estimation must be carried out.Therefore, this is an iterative system, where the analysts experience is fundamental, in order to carry out proper critical contrast of the results at each stage (Peña, 2005).

Validation
The objective of the validation phase is to demonstrate that the model found is the one that best represents the behaviour of the series under study, which, in turn, is that which fulfils the following requirements to the greatest extent: a) the coefficients of the model must be statistically signif icant; b) the residuals of the estimated model must approximate the behaviour of a white noise, to do this the coefficients of the functions of the estimated autocorrelation and the estimated partial autocorrelation must not differ significantly from zero; c) the estimated model is stationary; and d) the prediction is adequate.

Results
The trend analysis has been carried out adjusting the series to a time function with a linear equation.After obtaining the trend equation, it is subjected to tests carried out on the residuals of the linear model to determine how well the model fits the data.These tests are to check the difference in the average between the 1 st half and the 2 nd half of the series and in the difference in the variance between the 1 st half and the 2 nd half of the series.If the model passes the tests, it infers that the difference is not statistically signif icant (p ≥ 0.05), and therefore, the equality between the 1 st half and the 2 nd half of the series is verified, both on average as in variance.
The trend was applied to evaluate the economic situation of the product in terms of price development over a large time span.The results are shown in Table 2 regarding the annual percentage variation.The series follow a trend that can be represented graphically.As an example, for the present work, Figures 1 and 2 have been included, in which the equations of the trend can be observed for the round tomato in Murcia and green pepper in Almería.For the other plant types the same analytical procedure is followed.
Figure 1 shows a very stable trend with a descending slope of 34% over 14 seasons (2% annually).
The linear model passes the test for the difference in variance between the 1 st half and the 2 nd half of the series and the test for the difference in the average between the 1 st half and the 2 nd half, as these differences are not statistically significant (p ≥ 0.05), thus verifying the equality of variance and average between the 1 st and 2 nd half of the series.Figure 2 gives an example of the graphic behaviour of the trend lines for the pepper and corresponds to the trend line followed by the prices of the green pepper in Almería.The results from this analysis show that the descending linear trend has a 37% drop over 14 seasons, which means 3% annually.In this linear model the equality of the average between the 1 st and 2 nd half of the series is verified.
Table 2 shows the trend equations obtained in the analysis of the origin prices, by production regions and product type.The first thing one might observe is that the slopes of the equations of the trend are, in most cases, highly significant, as indicated by the p-values obtained.On the other hand, there is a very low probability that the values estimated for the aforementioned slopes are given only for reasons of randomness.In other words, it is highly probable that the estimated values are true (probability of error below the level of significance).
With reference to the tomato, the ribbed type displayed an annual price increase both in Almería and Alicante, while the round variety followed a declining trend that oscillated between 1% annually in Almería and 2% annually in Murcia.The price of the pepper has also dropped in recent years with a decrease of between 1% for the green pepper in Alicante and 3% for the green pepper in Almería and the red pepper in Alicante.
With respect to the seasonality indices, these have been obtained on the basis of the series of data indicated in Table 1, and reflect the price oscillation over an average season in percentage terms.Table 3 shows all the indices calculated; moreover, it indicates the average price of the last season analysed for each product, to serve as the reference value.
Production cycles exhibit similar patterns in Murcia and Alicante but are somewhat different in Almería.The effect of seasonality on the prices can also be observed graphically.Thus, as an example, the figures are shown corresponding to the round tomato in Murcia and the green pepper in Almería (Figures 3 and 4). Figure 3 gives the values of the seasonal indices for the prices of the round tomato in Murcia which clearly displays that the highest prices are obtained in the autumn months and at the end of spring.
The seasonality indices, reach a maximum of 143.48 in the fortnight 14 (month of April).After this point, it drops to a minimum of 64.14 in fortnight number 19 (month of July).
In the case of the pepper, the example taken for the seasonal behaviour of prices is that of the green pepper in Almería.In Figure 4 one can see that all the seasonal indices exceed the value of 100 between the second fortnight of December and the first fortnight of April, reaching a maximum of 166.51 in the fortnight number 11, and then dropping to a minimum of 57.95 in the fortnight number 17 (month of June).
Table 4 presents the periods of the season where the highest values are obtained for the seasonal indices of prices.One can also observe the similarities between Murcia and Alicante and how in Almería the seasonal peaks come forward in the case of the pepper.
Finally, this work is completed with a prediction obtained by applying the ARIMA models to the time series of prices for the green and red pepper in Almería.In these series of fortnightly periods, the time period studied encompasses the years 1992 to 2006.First of all, the series were adjusted so they would be stationary on average, differentiating them once in the regular part (d = 1) and in the seasonal part (D = 1), and the transformation natural logarithm was applied to stationarize them in variance.
The autocorrelograms obtained are shown in Figures 5, 6, 7 and 8, in which the X axis represents the lags corresponding to time movement of a time unit (in this case, fortnights), which occurs when the autocorrelation is carried out.The analysis of these autocorrelations, reflects the consideration of a moving averages of degree two [MA(2)] and a seasonal moving averages of degree one [SMA(1)], therefore, the series were modelized using an ARIMA (0,1,2) × (0,1,1,) 20.Estimated coefficients for each model are shown in Table 5.The moving average coefficients for degrees 1 and 2 as well as the moving average from the seasonal component, SMA(1), are both statistically significant (p-value < 0.05).
To validate the model, the residuals were checked to ensure they followed a white noise process (distributed approximately as a normal, average zero and constant variance) as well as the following tests, for which statistically insignificant p-values were obtained (p-value ≤ 0.05): -RUNS = Test for excessive runs up and down.-RUNM = Test for excessive runs above and below the median.
-AUTO = Box-Pierce test for excessive autocorrelations.
-MEAN = To test the difference in average between the 1 st and 2 nd half.
-VAR = To test the difference in variance between the 1 st and 2 nd half.
Figure 9 shows the results of the price forecasts obtained using this model, for the green pepper and red pepper in Almería.Deviations between predictions and the real prices are small, and the behaviour of both series is very similar throughout the campaign.All the values fall within the limits defined by the lines of the confidence intervals obtained from the prediction.
The main deviation to be found, between the end of November and the beginning of January, may be due to a commercial disturbance caused by the appearance of non-authorised residues, which on being detected in some European markets caused a sharp drop in prices (Junta de Andalucía, 2007).On correcting this incident, the prices recovered after January, which is more noticeable in the green pepper.
Concerning the green pepper, the prediction values show, in accordance with those obtained in the seasonality analysis, that the highest prices would be recorded between the months of January and March.The greatest price drop occurred in the season spanning the 1 st fortnight in March and the 2 nd fortnight in April.
The predictions for the price series corresponding to the red pepper in Almería are shown in Figure 9B.In this case, also coinciding with seasonality, the forecast for the highest prices occurred between the months of February and March, while the greatest price drop in the season would take place between the 2 nd fortnight of March and the 1 st in June.
Together with the analysis of the origin price trends, an assessment has been made of the technology and equipment to be found in greenhouses, thus offering the possibility of providing economic orientation to help in decision-making.
One of the main strategies followed to achieve higher prices is to obtain special products like the ribbed tomato, on-the-vine tomato, cherry tomato, different coloured peppers or the Wonder type etc, and to do so  it is necessary to carry out improvements in greenhouse structure and equipment.When price trends do not give the farmers favourable perspectives, this leads to situations in which low-levels of f ixed capital are used, as happens in the case of greenhouses covered only by netting, particularly in tomato growing.
Another consequence of the decreasing price trend on the most common production, is that important modifications, like soilless culture, are being implemented slowly or even reaching a standstill.Thus, since the year 2000, the area dedicated to soilless culture has not exceeded 20% in Almería.
Despite this, it is necessary to adopt more technology if growers want to improve quality and adapt their production system to the seasonal maximum of the prices; in fact, when the greenhouse is better equipped, even with some form of heating system, it is easier to reach high price levels, and bring forward the natural market cycle that normally takes place in summer and autumn.
Besides the size and equipment in the greenhouse, there is a direct relationship between the product prices, other production modes, which include IP and, in recent years, EP.
On the other hand, cultivation using IP systems with biological control techniques, which always demand a higher level of technology, contribute to improving the commercial image of the product.Nowadays, it is still difficult to quantify the economic differences favouring IP, except the priority of choice in the marketing phase.

Discussion
The price series analyses have included the concepts of trend and seasonal indices, within the period encompassed in each series.For the pepper, also included is a price forecast for the following year.
Bearing in mind the wide range covered by the series and the fortnightly period of the prices, a high number of data are included.Therefore, the resulting trend gives very solid information as an expression of average price behaviour.As the values are updated to the Consumer Prices Index (CPI) corresponding to the last date of each series, one can easily calculate the increase or decrease in purchasing power for a given product and deduce the influence this has on the farmer's income.
According to the results of the trend analysis of the prices it can be concluded that: -There is a standstill or slight decline in price of the round tomato in the main producing regions, like Almería and Murcia.However, in Alicante a positive trend is recorded for this product, which could correspond to the differences in the marketing process.
-Concerning the differentiate type, the ribbed tomato follows a positive price development, both in Almería and in Alicante.This is because it is a special type of tomato and it is therefore necessary to find a niche in the market where higher prices can be obtained, which is possible with differentiated products.
-For the red and green pepper, and in the three growing regions studied, the development of the price series displays a downward trend that is not very great, common to products that started off at very high price levels, and as production has increased, the prices have followed a slight downward trend or stabilized.The results obtained using the seasonal indices provide highly valid information regarding the relationship between the production schedules and prices, and what should be done in terms of choice of variety, production techniques, greenhouse equipment and decisive aspects like heating.
The ARIMA models are shown to be valid for shortterm forecasts of one season.Nevertheless, the results provided are of great interest in crops like the pepper grown in the greenhouse, where the response obtained in production and prices is evident with respect to the techniques and means employed.
Heating has an effect on the production schedule and seasonality of the products, improved quality and, in its turn, affords greater security against the risk of low temperatures.However, there are certain limitations to its use due, to a great extent, to the increasing price trend of commonly used fuel.For its temporary and permanent use, strict technical and economic controls are demanded.
In contrast to what often happens in the sale of products with IP techniques, when products of ecological origin are marketed there is a favourable discrimination in the prices.This is a consequence, on one hand, of the current consumer demands with respect to food safety and health, and on the other, the guarantees that are attributed to the Ecological Production label.

Figure 3 .Figure 4 .
Figure 3. Seasonality of the round tomato prices in Murcia.
Figure 1.Round tomato prices series in Murcia.2 Figure 2. Green pepper prices series in Almería.

Table 3 .
Seasonality indices of the series of origin prices in percentage

Table 4 .
Period of the season with the highest seasonality indices of the origin prices
A B