Multi-criteria decision system for greenhouse site selection ın Lower Euphrates Basin using geographic ınformation systems ( GIS )

Greenhouses have become quite a sought-after agricultural sector due to government support and private company preferences in recent years in Turkey. Greenhouse activities are only limited to Mediterranean and Aegean regions of Turkey and it is not much preferred in the inner regions. Due to these regions being preferred by people, mainly as tourism and settlement areas, it has become essential to find new areas for greenhouse operations. The climatic and topographical conditions in the Southeastern Anatolia Region, are appropriate regarding greenhouses. In this context, the Lower Euphrates Basin is selected as the study area. It covers four provinces (Adiyaman, Gaziantep, Kilis and Sanliurfa) and is one of the preferred regions for greenhouse operations in recent years. In this study, the climatic conditions and the production capacity of the Antalya province, which has the most greenhouse production areas of the Mediterranean Region and Turkey, are taken into consideration and the four provinces of the Lower Euphrates Basin were investigated for their compatibility with the desired conditions concerning green housing. Climate, soil, wind, altitude, slope, aspect, distance to rivers and lakes data are essential in choosing the greenhouse location. A Geographic Information System (GIS)-based multi-criteria evaluation for greenhouse site selection was applied in the study area by using these data. The validity of the results has been checked over with the field studies and also compared with the climatic conditions and production capacity of Antalya Province. The suitable, nonsuitable or partially suitable areas for greenhouse siting in the basin were determined in terms of GIS, for guidance to producers and researchers in the future.


INTRODUCTION
Greenhouse production is one of the most important income generating branches of vegetable production.The primary need of the people for food is increasing with the growing population.Whether the production from agricultural areas is low or inadequate, the application of modern agricultural techniques can not meet the expectations from agriculture.Therefore, about 795 million people are undernourished globally, below 167 million over the last decade, and 216 million less than in 1990-1992.The decline is more pronounced in developing regions, despite significant population growth.In recent years, progress has been hindered by slower and less inclusive economic growth as well as political instability in some developing regions, such as Central Africa and western Asia (Anonymous, 2015).The increase in crop production is possible only if the production period is continuous.The greenhouse is the most crucial vegetable or fruit production material of our time.In the next century, the impact of the work will increase and this sector will become more interesting because of climatic changes or too much production demand.Greenhouse farming is a vegetable production plant where climatic conditions can be controlled and production control can be carried out.However, if done in climatic conditions that are not suitable, the economic benefit will bring damage instead.
In recent decades, greenhouse area has risen worldwide, due mainly to the increased use of plastic greenhouses for growing vegetable crops.Site selection is a critical factor for profitable and sustainable greenhouse production.The main factors determining the location and site selection of a greenhouse production area are: the cost of production, quality of produced yield, and transportation cost to markets (Nelson, 1985).
A wide variety of greenhouse technological packages are available nowadays, from simple plastic houses to very sophisticated glasshouses.Information on the growth strategies that can be afforded to conveniently supply the markets with high-quality vegetables yearround and in the greenhouse packages commonly used for that purpose are included, with particular reference to the Mediterranean Basin (Castilla and Hernandez, 2007).
Turkey's fastest-growing population, the reduction of agricultural areas due to climate change and the increase of residential areas have led people to find different production areas or to increase the number of products received from the unit area.The production can be continued for 12 months by increasing the greenhouse areas.Greenhouses are the vegetable production facilities, which enables production in the desired periods by keeping under control in indoor conditions such as temperature, humidity, CO 2 , etc.
Since, air temperature and humidity are the two significant parameters affecting thermal comfort significantly, and only sensible load can be handled by an evaporative cooling system, the conventional evaporative cooling system is suitable for dry and temperate climate where the humidity is low (Costelloe and Finn, 2003;Heidarinejad et al., 2009).
During the production in greenhouses, internal environmental conditions need to be regular.Worldwide, studies have identified greenhouse and indoor conditions.The indicated species traditionally grown in the warm season, adapt to ambient temperatures ranging from 17 to 28°C and are sensitive to the cold.Temperatures persistently below 10 to 12°C over several days affect productivity, as do temperatures above 30 to 35°C.The daily variation between day and night average temperatures (thermal periodicity) is required for proper physiological functioning, with these thermal differences between 5 and 7°C.The minimum daily radiation requirements of these species are estimated at around 2.34 kWh m -2 day -1 which implies around 6 h of light per day, to a minimum total of 500 to 550 h of light during these three months.Other desirable climate parameters for these species would be the soil temperature of more than 14°C and ambient relative humidity of 70 to 90% (Nielsen et al., 1988).
Air temperature as well as solar radiation and air relative humidity is one of the most important variables of the greenhouse climate that can be controlled.It conditions not only crop development and production, but also energy requirements, which can account for up to 40% of the total production costs.The majority of plants grown in greenhouses are warm-season species, adapted to average temperatures in the range 17 to 27°C, with approximate lower and upper limits of 10 and 35°C.If the average minimum outside temperature is < 10°C, the greenhouse is likely to require heating, particularly at night.When the average maximum outside temperature is < 27°C, ventilation will prevent excessive internal temperatures during the day; however, if the average maximum temperature is > 27 to 28°C, artificial cooling may be necessary.The maximum greenhouse temperature should not exceed 30 to 35°C for prolonged periods (Kittas et al.,, 2013a).
In Turkey, there is no regular heating in the greenhouses.Generally, local heating is applied to protect the plants from frost.Failure to achieve the desired level of climate conditions and the lack of heating cause problems such as low yield, limitation in producing variety and the necessity of using hormones.The most considerable expense of the greenhouses is the electricity costs associated with heating and cooling.In the case of careful selection of the greenhouse site, it is thought that these costs can be considerably reduced and productivity can be increased.
In Mediterranean areas, less energy is used (500 to 1 600 MJ m-2 ); however, heating is increasingly adopted to achieve early production and a constant quantitative qualitative yield, leading to higher energy use.Improved environmental control (e.g. more CO 2 supply, additional lighting), intensified production schemes and use of cooling systems all increase energy consumption.Average energy use accounts for 10 to 30% of total production costs, depending on the region (Kittas et al., 2013b).
Direction and speed of the wind are fundamental in the orientation of greenhouses.Covering the short axis of the greenhouse against the formation of natural ventilation  (Saltuk et al., 2017).
Table 1.The data set used for MCDA modeling (Didan, 2015;Worldclim, 2018;GDRE, 2018;GTOPO30).facilities or winds at high speed will enable it serve for many years.Also, greenhouse positioning should be considered for maximum utilization of solar energy.

MATERIALS AND METHODS
Euphrates River is the longest river in southwest Asia; it is 2,800 km long and is one of the two main constituents of the Tigris-Euphrates River system.The river rises in Turkey and flows southeast across Syria and through Iraq (Anonymous, 2016).Four provinces in the basin of the Euphrates in Turkey (Adiyaman, Gaziantep, Kilis, and Sanliurfa) have formed the research area (Figure 1).In the study, elevation, slope, and aspect data was derived from GTOPO30 global elevation model with a resolution of 30 arc seconds (1 km) (GTOPO, 2018).Climatic data layers representing annual precipitation, maximum temperature of the warmest month, minimum temperature of the coldest month with a spatial resolution of about 1 km 2 were collected from Worldclim global database (Worldclim, 2018).Wind data were taken from the Turkey wind energy potential map prepared by the General Directorate of Renewable Energy (GDRE) and resampled to 1 km 2 resolution.The land cover coverage was taken from the Modis MODIS/Terra Vegetation Indices 16-Day L3 Global 250 m SIN Grid V006 (MCD13Q1) with 1 km resolution (Didan, 2015).Study area soil types are taken from general soil map of Turkey prepared by the Ministry of Food, Agriculture and Livestock (MFAL) (Table 1).Distance to dams and rivers were achieved by the evaluation of the topographic map of the study area in the GIS environment.
The criteria, which are essential in the selection of the greenhouse location, recommended in Table 2 were obtained from FAO (2017) publication and had been used in the study.Accordingly, the following method was used in the grading of the values.If the value is above the average value, it is used to make the classification by decreasing the value.In the study, the assigned values between 1 and 10 are determined, and the most effective one is given the value 10.For example, in the study, the maximum wind speed for greenhouses was evaluated by assigning 10 for the areas that have wind speeds less than 4.5 m/s.Areas with a wind speed of 9 m/s are given the value 2. The same method was used in the other criteria that were effective in choosing greenhouse locations.
After transferring all data to GIS environment, Multi-Criteria Decision Analysis method supported with Geographic Information Systems (GIS) was used for the suitability maps of the study area.ArcGIS 10.2 software was used for this purpose.In the MCDA method, often accepted factors in the literature and field studies that are effective in choosing the greenhouse construction site were appointed as impact factors and weight values (Table 2).30 m × 30 m resolution factor value maps were created after assigning values to all the obtained factors, evaluated in the GIS environment.The obtained raster maps were analyzed with the weighted sum function of the Spatial Analyst Tool in the ArcGIS 10.2 software.As a result of the analysis, suitability maps for greenhouse siting in the study area were obtained (Figure 2).The validity of this map was checked by the land studies and also a comparison of the greenhouse suitability map produced for Antalya with the same factor values.
In this study, we have tried to find a common intersection point by correlating the information with the topographical structure of the basin and combining this data with the long-term climatic data.We aimed to evaluate all the data together with the most suitable areas as possible and to generate maps of compliance in the numerical sense.From here it is aimed to implement both investors' knowledge about the application areas of the investors and to encourage the producers in the right areas as possible in the investments and incentives of the state.

RESULTS
In our study, we have investigated the suitability for greenhouse siting both in the four provinces of the Lower Euphrates Basin and in Antalya.Therefore, eleven criteria that could be effective in the construction of greenhouses were considered and the results compared to the province of Antalya, that is, one of the cities where greenhouse farming is mostly preferred in Turkey.
Because of the climatic conditions in the coastal areas of Antalya, the Serik district located to the east and the districts (Kumluca, Demre) in the south of the province has the most production potential.A large part of the total surface area of these districts is mainly covered with undergrowth production and citrus.Because the income of greenhouse cultivation is higher than conventional farming and it takes less time, citrus farming has shifted to the inner regions, and coastal regions have become areas where greenhouse cultivation is mostly applied.There are large-scale bazaars and commissions for vegetable growers in these territories.These three districts, mostly fulfill Turkey's potential for season vegetable farming.The fact that the total produced products reach the sellers and distribute them nationwide is like these commissions, and the producers sell all the goods in their hands, and producers are encouraging the establishment of new greenhouse areas.Antalya is surrounded by the western Taurus Mountains that limits the areas suitable for the greenhouse in the city.For this reason, it has been attempted to theoretically determine whether the study area will be an alternative to the Antalya province.The suitability values in MCDA calculations in the Antalya province were found in the range of 3.59 to 6.20.The same MCDA values are in the range of 5.16 to 6.20 and 4.69 to 6.20 for the coastal provinces (Kumluca, Demre) and Serik respectively.According to these values, the greenhouse existence of Antalya province also meets in these regions.Sanliurfa province is considered the closest province of Antalya concerning the appropriateness of greenhouse due to climatic and topographic conditions.Therefore, regarding light transmission, it is recommended to build the greenhouse with an E-W orientation.Nevertheless, light uniformity is better in N-S greenhouses since the gutter and ridge shadows change their position during the day as the sun moves.In some Mediterranean areas, greenhouses are E-W oriented, but the crop rows are N-S for greater crop uniformity.About ventilation, it is advisable to build the roof ventilators perpendicular to the prevailing winds to enhance the air exchange (Montero, 2013).
The soil structure is significant for greenhouses.Although soilless farming options are used in some regions of our country to prevent the diseases caused by the soil and to form suitable growing areas, most of our producers still prefer grounded agriculture.
The nutrients dissolved in the soil's water are mainly nitrate, magnesium, potassium and sulfates.The quantity of these nutrients is ordinarily low and not sufficient for greenhouse crops which have high nutrient requirements (Sonneveld and Voogt, 2009).
The ideal EC is specific for each crop and dependent on environmental conditions (Sonneveld and Voogt, 2009); however, the EC values for hydroponic systems range from 1.5 to 2.5 ds m -1 .Higher EC hinders nutrient uptake by increasing osmotic pressure, whereas lower EC may severely affect plant health and yield (Samarakoon et al., 2006).The decrease in water uptake is strongly and linearly correlated to EC (Dalton et al., 1997).
Geographic information systems, which are now widely used in many areas, offer a wide variety of possibilities, especially in planning work.These systems which allows the collection, arrangement, questioning, and analysis of geographical data in a digital environment, accelerates geographically based works and to bring out the results with a more synthesist approach.It is quite easy to obtain the net result by comparison of the collected geographical data with the use of these systems in a short period (Bayar, 2005).The multifunctional applications of GIS make it possible to analyze spatial data, consolidate the information obtained from these analyzes, and store them.GIS helps to decide on the processing and distribution of agricultural land and can be used for administrative purposes (Sonmez et al., 2004).
Multi-criteria Decision Analysis (MCDA) is a solution applied in situations where more than one criteria must be evaluated together.The basic method of solving the problem is to divide the problem into small, understandable and straightforward pieces so that a meaningful result can be obtained from these pieces (Malczewski, 1999).On the other hand, MCDA provides a rich collection of techniques and procedures for structuring decision problems, along with designing, evaluating and prioritizing alternative decisions.At the most rudimentary level, GIS-MCDA can be thought of as a process that transforms and combines geographical data and value judgments (the preferences of the decision-maker) to obtain information for decision making.It is in the context of the synergetic capabilities of GIS and MCDA that one can see the benefit of advancing theoretical and applied research on GIS-MCDA (Malczewski, 2006).
In the study, Geographic Information Systems (GIS) were used for analyses and querying, and Multi-Critical Decision Analysis (MCDA) was used for interpreting and solution.MCDA is a solution applied in situations where more than one criteria must be evaluated at the same time.It is known that, in the selection of the greenhouse location, if the climatic and topographic conditions are planned, there is a benefit in terms of stability and usability as well as economic benefit.Our study, especially encompasses Turkey's Lower Euphrates Basin in the Southeastern Anatolia Region, which is a potential area of greenhouse farm.It is aimed to determine the appropriateness of 4 provinces in the basin based on the province of Antalya and to determine the suitable locations for greenhouses in the basin.In this context, topographic characteristics (Elevation, slope, aspect), climate (maximum temperature of the warmest month, the minimum temperature of the coldest month and annual precipitation, wind direction and speed), land cover data and soil types of the study area have been used.This data is the primary material of the study.
The dataset mentioned above is layered in the ArcGIS environment, and Multi-criteria Decision Analysis is performed with this data.Besides, it is aimed to remind the producers of the importance of greenhouse farming in the region and to determine the greenhouse potential in the basin.

Sanliurfa province
When the study area is examined in terms of area size, the largest province in the study area is Sanliurfa.The total area of Sanliurfa province is 18584 km 2 .Sanliurfa province has the third most farmland in Turkey after Ankara and Konya provinces.64% of the province area is considered as agricultural land.In the study conducted, Sanliurfa center, southern and southeastern part of Urfa province has been determined to be suitable for greenhouse cultivation.It is seen that in Sanliurfa province, the districts of Viransehir, Ceylanpinar, Harran, and Akcakale consist of areas that are partially suitable or suitable for greenhouse farming.Nearly all lands of the Viransehir district can be defined as partially suitable and suitable areas for greenhouse farming.In our study, seven of the thirteen districts of Sanliurfa province were found partially suitable and suitable for greenhouse farming.In Sanliurfa, the percentage of areas suitable for greenhouse siting, partially suitable and unsuitable were determined as 12.03, 51.10 and 36.87%,respectively (Figure 3a).

Kilis province
In Kilis province, it is seen that the areas suitable for greenhouse construction were mostly collected in Elbeyli district and its vicinity.This district is followed by Kilis center, Polateli, and Musabeyli districts respectively.When the province of Kilis was examined, it was seen that the percentage of areas suitable for greenhouse siting, partially suitable and unsuitable were 7.21, 52.85 and 39.94%, respectively (Figure 3b).

Adiyaman province
In Adiyaman, the percentage of areas suitable for greenhouse siting, partially suitable and unsuitable were determined as 19.98, 37.24 and 42.78%, respectively.The central district and the eastern part of the province of Adıyaman consist of regions suitable for greenhouse cultivation.The most suitable regions for greenhouse cultivation are Samsat, Merkez and Besni districts.These districts are followed by Kahta district (Figure 3c).Currently, in Adıyaman province, 135000 m 2 greenhouse cultivation is performed and is mainly made of vegetables (tomato and cucumber).In Kahta district, benefiting from geothermal energy, production is carried out at 70000 m 2 .Besides, there are also 65000 m 2 of soilless greenhouse cultivation areas.The hot water extracted from Turkey Petroleum Corporation's (TPAO) oil wells in Kahta district, increase greenhouse investments in the region.The greenhouses heated by the hot water extracted from the oil wells in the Kahta district of Adıyaman provide employment and economic contribution.

Gaziantep province
Areas which are not suitable for greenhouse farming are located in the province of Gaziantep as a percentage (53.58%)because of its presence in the north and its climate.Only in the southern part of the province Gaziantep that some areas are identified suitable areas for greenhouses.When the Gaziantep province was examined, it was seen that Karkamis district is the most suitable district for greenhouse cultivation.This district is followed by Oguzeli and Nizip districts.Islahiye district town center and its surroundings also have suitable areas for greenhouse farming.In Gaziantep, the percentage of areas suitable for greenhouse siting, partially suitable and unsuitable were determined as 10.31, 36.11 and 53.58%, respectively (Figure 3d).
In the study area, the climate is temperate in these areas.It is topographically flat, 400-600 m above sea level, with the areas near the ponds or that are bordered by water sources and water collecting basins such as the Ataturk Dam.This allows the possibility of greenhouse cultivation in the early and late season for crop Saltuk and Artun 2723 production.It is thought that this situation, which creates an important opportunity for vegetable production, especially for the Southeast Anatolian (GAP) Region market, can be supported by cheap and renewable energy sources such as biogas and geothermal, and this will further increase profitability.

DISCUSSION
In Turkey, greenhouse operations are limited only in the coastal areas of the Mediterranean and Aegean regions.
It is not preferred in other regions.Nowadays, in the Mediterranean and Aegean regions, the suitable areas for greenhouse cultivation are preferred for tourism and residential areas.For this reason, it has become mandatory to find new areas for greenhouse enterprises.For this purpose, Southeastern Anatolia Region is considered suitable for the climate and topographic conditions.In this study, the criteria that are effective in the selection of greenhouses as a settlement are chosen from the literature.Greenhouses need to be robust in terms of both static and strength.In particular, plant production structures such as greenhouse are produced from small diameter steel material due to the cost.Therefore, it cannot handle the wind load on it and demolitions occur.The air currents from the inner Anatolian wind pass to the Basra Bay and also show the research area as very risky in terms of wind load.In our study, the wind load coefficient of the long axis of the greenhouses that we sampled on the basis of provinces is higher than that of Antalya.Therefore, the wind factor was one of the most important factors in this study.Another important factor is the temperature.In particular, because the greenhouse indoor environment has only a layer of cover material thickness from the external environment, in terms of aquaculture, there has been a vast difference compared to Antalya.In the fall production period, 71% of the research area needs greenhouse heating.
The load combinations (wind, rain, and topography) of the effective loads in greenhouses were calculated both individually and together.All factors discussed in this study were classified according to their weight values, and the differences in Antalya were examined.
In the study area, the appropriate greenhouse siting areas were determined by MCDA analysis using the ArcMap 10.2 software.When the whole study area was examined, it was seen that suitable, less suitable and inappropriate areas for greenhouse siting are 13.23, 45.38, and 41.39%, respectively.Upon examining the provinces in the study area, Adiyaman (19.38%) and Sanliurfa (12.03%) provinces are those with the highest percentage of suitable areas for greenhouse siting in the basin.When the study area was examined on the basis of districts, it was seen that the most suitable areas for greenhouses are in Viransehir and Elbeyi districts.As a result, the use of this work in the greenhouse activities to be made in the Lower Euphrates Basin region in the coming years may help farmers use the resources properly.The Sanliurfa-Karaali region has a potential of the geothermal greenhouse, and the heating costs of this region can be considered an advantage.As a result of the questionnaires and evaluations made by public companies and private companies in the study area, it is stated that greenhouse activities should be among the priority policy areas.73% of representatives of public institutions, 67% of NGO representatives and 63% of university representatives stated that greenhouse should be among the priority policy areas (Anonymous, 2014).

Figure 2 .
Figure 2. The suitability map of the study area.

Table 2 .
The factors affecting the greenhouse site selection(FAO, 2017).