Application of Tourist Function Indicators in Tourism Development. Case Study of Tunisia

The article presents the assessment results of the tourism function development in Tunisian governorates. The evaluation was carried out with the index method based on two groups of function indicators – tourist traffi c intensity and tourism development indices. Additionally, the tourism function was analyzed taking into account the average length of tourist stays. The results indicate a varied degree of the tourism function development in the governorates – from a highly developed tourism function in the coastal governorates (Sousse, Nabeul, Monastir) to the initial stage in the mountain regions (Sidi Bouzid, Siliana, Zaghouan). In terms of the average length of stays, Tunisia turns out to be a short-term destination, in contrast to the general misconception. MONIKA WIDZ https://orcid.org/0000-0003-3264-927X Maria Curie-Skłodowska University in Lublin Faculty of Earth Sciences and Spatial Management Institute of Socio-Economic Geography and Spatial Management Department of Regional Geography and Tourism Aleja Kraśnicka 2cd, 20-718 Lublin, Poland monika.widz@poczta.umcs.lublin.pl


INTRODUCTION
The basic measures of tourism development in a given area are tourist traffi c and parallel development of tourist infrastructure. Determination of their size and structure indirectly indicates the touristic attractiveness of the region and helps to estimate the development of the tourism function. In the literature, it is referred to as the tourism-recreation function (Matczak 1989). It is defi ned as any socio-economic activity aimed at management of tourists in a specifi c spatial unit refl ecting the ability of the area to meet the specifi c tourism needs (e.g. Matczak 1989;Kowalczyk 2002;Kurek, Mika 2007). The tourism function, which is continuous and dynamic, is determined by various internal and external economic, social, and political factors (Gralak 2008).
Studies on the application of tourism function indicators in tourism have been reported by, e.g. Marković et al. (2017) in their study of the village of Zlakusa (Serbia), Lukić et al. (2018) in the Danube Region (Serbia), Štefko et al. (2018) in regions of Slovakia, and Wiskulski (2019) in Croatia. Investigations of the tourism function are quite an important part of studies of Polish tourism geography (e.g . Fischbach 1989;Liszewski (ed.) 1989;Matczak 1989;Derek 2008;Włodarczyk 2009;Durydiwka 2015;Zmyślony 2015;Krukowska, Świeca 2018). The authors carry out spatial analyses of rural and urban areas of Poland. A city, commune, county, and province (voivodeship) are the basic units of reference in their studies.
The aim of the study was to identify and characterize the tourism function in the administrative units of Tunisia. From the point of view of tourism development, the country offers highly attractive tourism resources, which are appreciated by domestic and international tourists. A signifi cant increase in the number of tourists in Tunisia was observed at the turn of the 20 th and 21 st centuries (Widz, Brzezińska-Wójcik 2020). It resulted in the development of tourism infrastructure, especially accommodation facilities, in the coastal locations in Tunis, Hammamet, Nabeul, Sousse, Monastir, Mahdia, and Djerba (Hellal 2017).
Tunisia is the smallest (163.61 thousand km 2 ) North African country with a population of approximately 11.44 million people. The average population density is 67 people per km 2 . The country is administratively divided into 24 governorates (delegations or districts) with a size in the range from 288 km 2 (Tunis) to 38,889 km 2 (Tataouine) (Institut National de la Statistique 2019). This information is important for the adopted research procedure.

MATERIAL AND METHODS
In the literature, the tourism function is determined with the use of the indicator method in two aspects: 1) tourism development resources (e.g. Defert 1960;Warszyńska 1985;Warszyńska, Jackowski 1979;Chudy-Hyski 2006;Szromek 2012) and 2) tourist traffi c (e.g. Defert 1988;Warszyńska 1985;Warszyńska, Jackowski 1979;Szromek 2012). The tourism development was assessed in this study with the use of the Baretje-Defert index and the so-called accommodation density index. In turn, the size and spatial differentiation of tourist traffi c in the Tunisian governorates were determined on the basis of the number of overnight stays. To this end, two indicators used commonly in tourism geography were considered: 1) tourist traffi c density, i.e. the Defert index, and 2) tourist traffi c intensity, i.e. the Schneider index. Additionally, an attempt was made to classify tourism in Tunisia in terms of the length of stay based on the indicator of the average length of tourist stays in the governorates.
The analysis of the tourism function was based on two types of data: 1) secondary resources (statistical data from 2017 provided by Offi ce National du Tourisme Tunisien -O.N.T.T., Institut National de la Statistique -I.N.S. and Commissariat Général au Développement Régional -C.G.D.R.) for calculation of the indices of the tourist traffi c intensity, tourism development, and average length of tourist stays; 2) primary resources (MerlinX reservation system in 2017-2019) for determination of the time ranges for short-, medium-, and long-term tourism.

Tourism development indicators
The Baretje-Defert (I BD ) index, also referred to as the tourism function index, is regarded as a universal measure of tourism function in relation to the characteristics of tourism development. It is expressed as the number of tourist accommodation facilities per 100 inhabitants of an analyzed area (Defert 1988): uest beds number of inhabitants 100 There are varying interpretations of the I BD index in the literature. As proposed by Warszyńska (1985), the tourism function can be defi ned in the following fi ve-grade scale: 0 -the process of tourism function development has not commenced (index value < 0.78), 1 -initial stage of development (index value 0.78-6.25), 2 -additional function (index value 6.25-25.00), 3 -equal or supplementary function (index value 25.00-50.00), 4 -basic or one of the basic functions (index value > 50.00). Warszyńska and Jackowski (1979), Kowalczyk (2002), and Szromek (2007) indicate that an area serves a real tourism function at an I BD index value of 100 (accommodation capacity equal to the number of permanent residents). Pearce (1995) mentions the six-grade interpretation of the I BD index proposed by Boyer (1972). According to this classifi cation, the index with the value of 40 denotes an area with a dominant tourist function. In the present study, the interpretation proposed by Warszyńska (1985), where the threshold index value of 50 indicates well-developed tourism areas, has been adopted.
The other measure of tourism development used in the study is the accommodation density index (I GBN ), which determines the density of accommodation facilities in the studied area (Warszyńska, Jackowski 1979): uest beds area of the nalyzed region in km a 2 100 The interpretation proposed by Warszyńska (1985) and adopted in the present study suggests a threshold value of 50 for areas with well-developed tourism. The other ranges of values and the degrees of tourist function development are interpreted in accordance with the Baretje-Defert index (I BD ).

Indices of tourist traffi c intensity
Two indices of tourist traffi c intensity, named after their authors Defert and Schneider, were used in the study. The Defert index (I D ), expressing the number of overnight guests per 1 km² of the area, facilitates an assessment of tourism density in the studied area (Defert 1988): number of accommodation users a area of the nalyzed region in km 2 As proposed by Warszyńska (1985), depending on the size of the index, the tourism function can be defi ned in the following fi ve-grade scale: 0 -the process of tourism function development has not commenced (index value < 15.6), 1 -initial development stage (index value 15.6-125.0), 2 -additional function (index value 125.0-500.0), 3 -equal or supplementary function (index value 250.0-500.0), 4 -basic or one of the basic functions (index value > 500.0). In the interpretation of the results of the I D index following Warszyńska (1985), an area with the I D index value exceeding 1,000 is well developed in terms of tourism.
The Schneider index (I Sh ) is similar to the Defert index, as it is based on the same principal variable, i.e. the number of overnight guests. This indicator shows the number of overnight visitors per 100 permanent residents of the area. Schneider combined some features of the Baretje-Defert and Defert indices and proposed an indicator comparing the number of tourists to the number of permanent residents in a given area: umber of accommodation users inhabitants number of 100 The value of I Sh facilitates classifi cation of the tourism function in the fi ve-grade scale from 0 -the process of tourism function development has not commenced (index value < 7.8) to 4 -basic or one of the basic functions (index value > 500.0). An index value of 500 or higher indicates well-developed tourism in an analyzed area (Warszyńska 1985).

Indicator of the average length of tourist stays
The tourism function in the administrative units of Tunisia was also analyzed in terms of the indicator of the average length of tourist stays, which indirectly provides information about the characteristics of tourist stays. The importance of this type of data from a tourist area has been emphasized by many authors (e.g. Doxey 1975;Raymond, Brown 2007;Durydiwka 2015). The length of stays may infl uence the degree of tourist function development.
The classifi cation of tourism in terms of the duration of stays is not explicit in the literature. Kruczek (2009) distinguishes short-term (up to three overnight stays) and long-term (over three overnight stays) tourism. The author additionally distinguishes no-overnight stay and weekend tourism in the former group. Więckowski (2010) classifi es tourist stays into short-term (up to two nights) and medium-term (from two to four nights) types. In turn, Buczak et al. (2015) underline that, in addition to the duration specifi ed in the Regulation of the European Parliament and the Council (EU) (2011), it is possible to introduce an additional classifi cation, e.g. 4-6 overnight stays, 7-13 overnight stays, etc., depending on the needs associated with the tourist stays.
Since international visits organized by tour operators as the so-called tourist packages dominate in Tunisia (Brzezińska-Wójcik, Widz 2017), it is important to use a classifi cation of stays corresponding to the duration of package holidays offered by travel agencies. Therefore, stay and tour packages in Tunisia offered in 2017-2019 were analyzed using the MerlinX reservation platform. In total, 10,211 packages were analyzed, including 9,882 stay packages, 194 tour packages, and 135 optional packages.

Development of the tourism function in Tunisian governorates in relation to the value of tourism development indicators: The Baretje-Defert index (I BD ) and the accommodation density index (I GBN )
The values of the Baretje-Defert index calculated for the 24 areas range from 0.050 to 10.406 (Tab. 1). In accordance with the adopted classifi cation proposed by Warszyńska (1985), no governorate achieved the highest (4 th or 3 rd ) degree of tourism function development. An additional function (2 nd degree) was Tab  The values of the Baretje-Defert index and the accommodation density index clearly indicated the degree of the tourist function development only in some governorates. This was especially evident in the Sousse and Nabeul regions, where the additional function was assigned. The initial stage of development was recognized in another three governorates: Mahdia, Jendouba, and Tozeur. Signifi cant discrepancies between the values of both indicators were noted in the other governorates. An example of such differences is the governorate of Tunis, where the degree of tourist function development was estimated at 1 by the I BD index and at 3 by the I GBN index. Such a large variation in the tourist function development in this area is associated with the value of the second variable, i.e. the number of permanent residents and surface area of the region.

Development of the tourist function in Tunisian governorates in relation to the values of the tourist traffi c intensity indicators -Defert (I D ) and Schneider (I Sh ) indices
The results revealed varying values of the Defert index (I D ) in the range from 0.057 to 2018.184 in the 24 governorates of Tunisia (Tab. 3). The highest degree of development (4 th ) referred to as the basic or one of the basic functions was only identifi ed in the Tunis governorate. According to the interpretation proposed by Warszyńska (1985), this is an area with well-developed tourism, as the value of its index exceeds 1,000. Sousse was assessed as a governorate with highly important tourism function as well. The value of the I D indicator, i.e. 518.146 (3 rd degree), indicates the equal or supplementary function in relation to other economic functions (Tab. 3).
The second degree of tourist function development defi ned by Warszyńska (1985) . 3-4). The other delegations exhibited signifi cantly different values of both indicators. As in the case of the Baretje-Defert and accommodation density (I GBN ) indices, this is associated with the differences in the surface area and in the number of permanent residents. Examples of the differences in the Defert index are the Sousse and Tunis governorates. The number of overnight visitors in Sousse was 1.38 million, which is the highest number of all the governorates (2 nd place in the ranking and the 3 rd degree of tourism function development). In the Tunis governorate, there were only 581.24 thousand overnight tourists at the highest value of the index, i.e. 2018.184 (4 th degree). This discrepancy is related to the differences in the surface area of the governorates (Sousse -2,669 km 2 ; Tunis -only 288 km 2 ). The discrepancy between the values of the Schneider index can be illustrated by Kebili and Sousse. Kebili was ranked the 2 nd place despite the small number of overnight tourists, i.e. 330.69 thousand. In contrast, the leader among the governorates in terms of the number of overnight tourists, i.e. Sousse with 1.38 million overnight visitors, was only the 3 rd in the ranking. As in the case of the Defert index, this is associated with the different number of permanent residents in the area.

Comparison of the average length of tourist stays with tourist packages
The analysis of the offer from tour operators in Tunisia (stay and tour packages) helped to distinguish seven periods of short-term, medium-term, and long-term tourism (Fig. 1). The long-term tourism was predominant -92% (7-8 nights -36%, 14-15 nights -34%, 9-13 nights -29%, and over 16 nights -1%) in comparison with medium-term -6% and short-term tourism -2%.
The time intervals distinguished by the indicator of the average length of tourist stays (Fig. 1) facilitated determination of the type of tourism in terms of the duration of stays in each governorate (Tab. 5). Short-term stays were noted in 67% of the area of the country. In 13 out of the 24 governorates, the average length of tourist stays was 1-2 nights, i.e. these were weekend visits. The average length of stay of three nights was reported in three governorates, i.e. Ariana, Tunis, and Zaghouan. Medium-term stays (4-6 nights) were reported from 13% of all the governorates, namely Nabeul (4 nights), Bizerte (5), and Manouba (5). Long-term stays were reported from only 20% of the area of Tunisia. The longest stays (18 nights) were recorded in Ben Arous. 7-8 nights weekly stays were recorded in the Monastir, Sousse, Mahdia and Medenine delegations. There were no 9-13 and 14-15 overnight stays, i.e. the so-called two-week stays (Tab. 5). The results show that short-term tourism generally predominates in Tunisia. Long-term tourism is concentrated only on the east coast. This coincides with the offer from tour operators, i.e. long-term stay packages are available mostly in the governorates of Sousse, Monastir, Mahdia, and Medenine. In turn, the lack of records of 9-13-and 14-15-day stays is striking, especially since there is an extensive offer of tourist packages covering these periods in Tunisia (29% and 34%, respectively).

Relationships between the values of tourism development indices (I BD and I GBN ) and the tourist traffi c intensity (I D and I S ) and the average length of tourist stays in the Tunisian governorates
The comparison of the size of the indicators of tourism development (I BD and I GBN ) and tourist traffi c intensity (I D and I S ) with the average length of tourist stays in the governorates does not show any signifi cant relationships (Fig. 2).
For example, the values of the tourism development measures calculated for the Tunis governorate indicate the 3 rd degree of tourism function development according to the accommodation density index (I GBN ). This result might be explained by the long-term leisure tourism. However, the analysis of the indicator of the average length of tourist stays shows that this is an area of short-term active stays. A similar conclusion is suggested by the comparison of the values of the tourist traffi c intensity indicators. For instance, the Kebili and Tozeur districts were assigned with the 3 rd degree of tourist function development according to the Schneider index, but these areas are characterized by short-term tourism, as indicated by the results of the analysis of the length of stays. Conversely, longterm tourism was indicated to prevail in the Ben Arous governorate, compared to the other districts, but this region is only characterized by the initial stage of development according to the values of the Defert and Schneider indices.

CONCLUSIONS
In terms of the degree of development of tourism function shown by the values of the function indicators (in accordance with the interpretation proposed by Warszyńska), the entire area of Tunisia exhibits the initial stage of tourism development or the additional function. This is related to the considerable differentiation of the governorates from highly developed tourism regions on the coast (Sousse, Nabeul, Monastir) to the initial stage of development in the mountain areas (Sidi Bouzid, Siliana, Zaghouan). The highest degree of tourist function development defi ned as the basic or one of the basic ones was achieved only by the Tunis governorate (as shown by the I D index). In turn, the process of tourism function development has not commenced in more than 10 regions (as shown by the I D , I Sh , I BD , and I GBN indices).
Tunisia is generally a short-term destination in terms of the average length of tourist stays. This is associated with the fact that long-term stays are concentrated only in the governorates on the east coast of the country accounting for merely 20% of its area. However, it should be noted that some governorates may exhibit a higher degree of tourist function development than that revealed by the indicator values. This may be related to several factors. First, there are some doubts as to the interpretation of the measures of the tourism function components in the following indicators: tourism development (I BD and I GBN ), tourist traffi c intensity (I D and I Sh ), and the average length of tourist stays. Additionally, the phenomenon of "second homes" is not taken into account in the total number of accommodation facilities included in the calculation of the Baretje-Defert and the accommodation density indices. In such a case, there are a number of non-permanent residents having their houses, fl ats, or apartments in the area. This phenomenon is relatively common among the older generation of tourists from France, Italy, and Belgium in Tunisia. The coastal governorates of Nabeul, Sousse, Monastir, Mahdia, and Medenine (in particular the island of Djerba) are most popular with foreigners (Hellal 2017). Moreover, in Tunisia, there are hundreds, if not thousands, of private accommodation facilities that are rented by tourists but not included in the statistics. This problem of the reliability of statistical data has been highlighted by Dryglas (2013).
The number of overnight tourists is used to calculate the Defert and Schneider indices. However, it does not include the number of tourists who come to the governorate only to visit a unique tourist attraction (e.g. the Great Mosque in Kairouan) and are not included in the statistics, as they spend the night in neighboring governorates (e.g. Sousse, Monastir). There are also certain doubts as to the accuracy of the data on the average length of tourist stays. As noted above, there are substantial discrepancies between secondary statistical data and trends in offers from tour operators.
Moreover, the results of the degree of tourist function development in the analyzed regions are highly diverse. This is associated with the different statistical variables used for the calculation of the indices. Therefore, the present results cannot be interpreted as an unambiguous indication of the degree of tourist function development. However, they constitute a good background for further considerations of these issues, which should include calculations and analysis of synthetic indicators, e.g. as in the procedure developed by Zioło (1973), the two-dimensional indicator of tourism function development (W W-Sz ), and the logistic indicator of the tourism function (W Lβ ) proposed by Szromek (2012).