Spati al Anal ysi s of Chang es i n the Dem og raphi c Characteri sti cs of the Cres-Loši nj Area U si ng the G I S M od el

. The paper aims to carry out a spatial analysis of the changes in selected demographic characteristics of the Cres-Lošinj area using the GIS model. For this purpose, data from all censuses carried out so far (1 857–2021 ) were used, which were primarily analyzed cartographically by using Mean Center and the Standard Deviational Ellipse (Dir-ectional Distribution) spatial analysis methods, which are an integral part of the basic software package ArcGIS Desktop version 1 0.0 produced by ESRI. Tabular and graphical appendices were used to additionally complement the analyses. The analyses were carried out on three levels: the level of the Cres-Lošinj area as a whole, the individual level of the Town of Cres and the Town of Mali Lošinj, and the level of settlements. For the area of Cres-Lošinj as a whole, the results showed a tendency of the mean center of the population moving towards the north in the 1 857–1 948 period, then towards the southwest in the 1 948–1 991 period, while in subsequent censuses the tendency was again towards the north. Among the local self-government units, a common feature of the movement of the mean center of the population is that in the observed period (1 857–2021 ), and particularly the last observed year (2021 ), the mean centers of the population are very close to the center of the town/municipality. Using the standard devi-ational ellipse (directional distribution) method, it was determined that larger settlements (Mali Lošinj, Cres and Veli Lošinj) have an increasing influence on the population direction, and it was determined that the population is concentrated closer to the larger settlements, i.e. to the centers of local self-government, while at the same time the influence on the directional distribution of the population of a large number of small settlements in the northern part of the research area decreased due to an increasing reduction in the number of inhabitants. The GIS model used in this study is suitable for researching the complexity of the demographic dynamics of the so-called of small populations because it enables an analysis at multiple spatial levels and contains methods of spatial analysis that use absolute numbers.

The research area ofthis paper is the Cres-Lošinj area, and according to their distance from the mainland, these two islands belong to the outer islands (Nejašmić 1992, Podgorelec 1999. In addition to the larger islands ofCres and Lošinj, this area also includes several smaller islands: Ilovik, Susak, Unije, Male Srakane and Vela Srakane and a large number of uninhabited islets (Podgorelec 1999). According to the current administrative and territorial organization, the research area ofthis paper includes two local self-government units − the Town of Cres and the Town of Mali Lošinj, which are part of the Primorje-Gorski Kotar County. The Town ofCres comprises a total of26 settlements located on the Island of Cres. The Town of Mali Lošinj comprises a total of14 settlements, which are distributed on the area of the Island of Lošinj (5 settlements) and other islands: Cres (4 settlements), Male Srakane (1), Vele Srakane (1), Unije (1), Ilovik (1) and Susak (1) (Figure 1). The total surface ofthe research area is about 515.9 km 2 ,whereby the Town ofCres is slightly larger (291.6 km 2 ) than the Town ofMali Lošinj (224.3 km 2 ).
According to the last population census from 2021 3 , 10,303 inhabitants were recorded in the Cres-Lošinj area. The largest settlements are also the only urbantype settlements − Mali Lošinj (with 5,577 inhabitants, 2021) and Cres (2,205 inhabitants, 2021).
This study aims to carry out a spatial analysis ofthe changes in selected demographic characteristics of the Cres-Lošinj area by using the GIS model of population development of the Cres-Lošinj area. The model of this study used data from all censuses conducted so far (1857−2021), while a more detailed analysis was conducted for the period from the mid-20th century onwards (1948−2021). Among other methods, the model of this study used the Mean Center and Standard Deviational Ellipse (Directional Distribution) spatial analysis methods, which are an integral part of the basic software package ArcGIS Desktop version 10.0 produced by ESRI, and applied the methodology of O'Neill et al. (1989), whereby each analyzed phenomenon is observed on at least three levels. The aforementioned methodology was applied in this study of the development of demographic characteristics due to its hierarchical approach on which it is based. Thereby, a deeper analysis of demographic processes was enabled, because local selfgovernment units were observed at the focal level, settlements at a level below, and the Cres-Lošinj area as a whole at a level above. On the other hand, in previous research in which GIS technology was used, population dynamics in different contexts were analyzed on one level (e.g. Shi 2010, Nyussupova andRadionova 2011) or on two levels (e.g. Antipova et al. 2012, Ficior et al. 2021, Onyango et al. 2021, Wiatkowska et al. 2021. Various methods have been used in previous studies related to the demographic characteristics of the Croatian islands at one point in time, i.e. to changes in the demographic characteristics ofthe Croatian islands over a shorter or longer period oftime. In this sense, it can be said that in older studies, the demographic characteristics of the Croatian islands were conducted without the use ofGIS technology (e.g. Novosel-Žic 1976, Novosel-Žic 1987, Stražičić 1977, Crkvenčić 1992, Nejašmić 1992, Lajić, 1993, Lajić and Nejašmić 1994, Podgorelec 1999, Lajić and Mišetić 2013a. In more recent studies, GIS technology was used, whereby we can differentiate between studies that only used it for cartographic displays (e.g. Mišetić 2013b, Marinković 2020) and those in which spatial analysis methods were used (e.g. Jovanić 2011, Jovanić andTurk 2013).
The application ofthe model is not often represented in studies ofindividual Croatian islands. Lajić and Mišetić (2006) conducted a study on the state and contemporary demographic processes on the Croatian islands. Among other things, they constructed a multivariate model ofthe future demographic development of certain island groups. Šulc (2014, 2016) used the model to conduct a study related to the area ofthe Croatian islands. However, these models differ from the model used in this study in several segments. There are differences in the data used, which is in accordance with the differences in the purposes ofthe models. In the model applied in this study, it refers to current demography, while in other researches it is future demographic development (Lajić and Mišetić 2006) and tourism (Šulc 2014, 2016). Also, the above models do not contain methods of spatial analysis, while in this study, a GIS model was used that contains data and methods of spatial analysis along with the applied methodology of O'Neill et al. (1989) whereby each analyzed phenomenon is observed on at least three levels.
From the above, it is evident that this study stands out from previously conducted studies on the demography of the Croatian islands, primarily due to the application ofa GIS model, which is comprised ofselected demographic data and methods of spatial analyses, whereby a methodology was applied in which each analyzed phenomenon is observed on at least three levels. HAMZIĆ, M.: PROSTORNA ANALIZA PROMJENA DEMOGRAFSKIH OBILJEŽJA CRESKO-LOŠINJSKOG PODRUČJA PRIMJENOM GIS MODELA će u konačnici rezultirati zaostajanjem u odnosu na procese koji su se u skladu s prilikama odvijali na kopnu i onim otocima bliže kopnu (Stražičić 1977).
Cilj je ovog istraživanja prostorno analizirati promjene odabranih demografskih obilježja cresko-lošinjskog područja primjenom GIS modela razvoja naseljenosti cresko-lošinjskog područja. U primijenjenom su modelu rada korišteni podatci svih do sada provedenih popisa stanovništva (1857−2021), a detaljnija je analiza provedena od sredine 20. stoljeća (1948−2021). Između ostalog, pri modelu su korištene metode prostorne analize Mean Center i Standard Deviational Ellipse (Directional Distribution) koje su sastavni dio osnovnog Slika 1 . Cresko-lošinjsko područje -područje istraživanja. Izvor: DARH; SRPJ. Naselje / Settlement Središte grada/općine / Municipal center Granica grada/općine / Municipal border Creskološinjsko područje / CresLošinj area In this paper, the applied GIS model of population development in the Cres-Lošinj area consists ofthe data and spatial analyses used ( Figure 2). The data used in this study refer to the census data entered in the spatial databases, while spatial analysis was used to determine demographic processes, i.e. trends in the Cres-Lošinj area. Namely, in this study, a general or neutral objectoriented GIS model was applied, which means it contains data related to settlements. This allowed for the application ofcontent analysis on three levels − the aim ofthis study was to apply the methodology ofO 'Neill et al. (1989) as part ofwhich each analyzed phenomenon is observed on at least three levels. Although it was created on examples of ecological systems, in this paper it was applied to the analysis of demographic characteristics of the Cres-Lošinj area due to its hierarchical approach, i.e. hierarchical theory. This means that the focal level or level of analysis is in the center of attention. It is limited by the control conditions of the level above (+1 level), whereby it receives a new function, and the level below (−1 level) with the connection elements of the focal level. Therefore, it can be said that in the conceptual model ofthe database ofthis paper, the local self-government units are represented by grouped settlements within the corresponding local self-government units, whereby they become the basis for the analysis of the focal level or level of analysis (0 level). The level above (+1 level) is represented by the Cres-Lošinj area as a whole. The level below (−1 level) is represented by the settlements.
2.1 . Metodološke napomene uz korištene popisne podatke Unutar prostorne baze podataka u ovom radu primijenjenog GIS modela razvoja naseljenosti cresko-lošinjskog područja nalaze se popisni podatci uneseni u prostornu bazu podataka. Od geometrijskih podataka u level. Namely, the first modern general population census was conducted in Croatia in 1857, and 16 more censuses have been conducted since then 6 . This is why for the purpose ofa basic analysis ofdemographic dynamics, that is, to gain a better insight into demographic trends over a longer period oftime, all censuses conducted so far were observed. However, due to significant spatial changes in demographic trends − very strong emigration and then immigration processes conditioned by socio-economic processes that took place differently in the research area 7 , the time period from the mid-20th century onwards (1948−2021) was observed in more detail.
It should be noted that when analyzing the demographic indicators of the censuses considered, we should note the methodological differences of the censuses conditioned by different census conceptions and definitions. Namely, the censuses from 1857, 1948, 1953, 1961, 1971, 1981 and 1991 were conducted according to the concept ofa permanent population (the so-called de iure population), while the censuses from 1869,1880,1890,1900,1910,1921 and 1931 were conducted according to the concept of the present population (the so-called de facto population) (e.g. Nejašmić 1991, Lajić and Nejašmić 1994), whereas the last three censuses (2001,2011,2021) were conducted according to the adapted concept ofa usual place ofresidence. It follows from the above that due to methodological differences, census results are not fully comparable (e.g. Pokos 2003, Lajić andMišetić 2013a). Due to the lack of fully harmonized census indicators and for the purpose of presenting the general demographic trends as faithfully as possible, the official census results published on the official website ofthe CBS are used as factual results in this paper and are compared with each other. However, the described inconsistency of census indicators should be taken into account in the obtained results.

Methodological notes to the spatial analysis methods used
In accordance with the methodology ofO 'Neill et al. (1989) applied in this study (explained in more detail in Section 2 of this paper), the analyses were carried out for the Cres-Lošinj area as a whole, then separately for the Town ofCres and the Town ofMali Lošinj, and then for the individual settlements (particularly the centers of the local self-government units). In this study, the Mean Center method ofspatial analysis was used for the purpose of determining the central position of the settlement in relation to the position ofall settlements, and the Standard Deviational Ellipse (Directional Distribution) method of spatial analysis was used for the purpose of determining the ellipse ofsettlement distribution in relation to the position of all settlements. Also, for the purpose ofspatial analysis and determining the changes in demographic characteristics, when using the above methods, in addition to entering the location ofthe settlement, census data (number of inhabitants) of each settlement was also entered. This allowed for the spatial observation of the selected demographic characteristics. The methods used are an integral part of the basic software package ArcGIS Desktop version 10.0 produced by ESRI.
The Mean Center method is a method of spatial analysis applied for the purpose ofdetermining the location that is the center ofconcentration ofobserved objects in the research area. The result ofits application is a point with coordinate values that outline that location. In this study, two options ofthis method were used: the option in which only the position ofall settlements is observed, which makes it possible to obtain a central position in relation to all settlements in the Cres-Lošinj area, and the option ofthe weighted mean center in which, in addition to the position of all settlements, the number of inhabitants of a particular census year is observed, which makes it possible to obtain the mean center ofthe population for that particular census year.
2) The interval of demographic regression with a predominant but discontinuous decline in the total number ofinhabitants (1910−2021). 1) Between 1857 and 1910, the total number of inhabitants of the Cres-Lošinj area increased from 19,085 to 21,260 inhabitants, which is the highest recorded number of inhabitants of this area so far. In this 53year period, the number ofinhabitants increased by 2,175 inhabitants, i.e. by an average of 41 inhabitants per year. The demographic increase, despite emigration, was a consequence of the entry of the population of the Cres-Lošinj area into the early sub-stage of the demographic transition. This means that a relatively high birth rate and a sharp reduction in death-rate have resulted in a very high natural increase. Emigration affected both the surrounding islands (e.g. the Island ofKrk) and the Cres-Lošinj area, and was conditioned by the demise of sailing ships and vine disease, which is why the population emigrated to overseas countries (Novosel-Žic 1987) and to Rijeka (Klempić-Bogadi and Podgorelec 2009). The predominance of emigration against demographic growth affected Cres in the late 19th century, while in the early 20th century it also affected Lošinj and the islands ofthe Lošinj archipelago. 2) Between 1910 and 2021, the research area recorded a double decrease in the total number of inhabitants (from 21,260 to 10,303), that is, by 10,957 inhabitants, i.e. by an average of99 inhabitants per year. Figure 3 shows that from 1910 to 1971 and from 1991 to 2021, each census recorded a decline in the total number of inhabitants of the entire Cres-Lošinj area, as well as the towns ofCres and Mali Lošinj. The emergence and strengthening ofsuch continuous depopulation was influenced by the demographic and other consequences of the world wars, the emigration of the predominantly Italian population due to the change in the jurisdiction over the Cres-Lošinj area, followed by the strengthening of migration from the countryside to the cities and the departure of the population to so-called temporary work abroad, particularly of the younger vital and fertile population, as well as demographic aging, natural population decline, etc.
Obrat od snažnih iseljavanja prema useljavanjima započinje šezdesetih godina kada je zabilježeno useljavanje na otok Lošinj, a sedamdesetih na Cres. To je u kretanjima ukupnog broja stanovnika vidljivo u porastu broja stanovnika prvo Grada Malog Lošinja (međupopis 1971−1981), a potom Grada Cresa (međupopis 1981−1991). Razlog tome je izgradnja modernih hotela zbog čega je bila potreba za is the inter-census period of 1890−1900, during which a smaller demographic decline was recorded (−1.2%). On the other hand, the second interval is represented by the intercensal periods of 1971-1981 and 1981-1991, during which a significant demographic increase was recorded for the Cres-Lošinj area as a whole (3.7% and 13.9%). However, it should be emphasized that in the intercensal period of 1971−1981, the recorded increase is the result of the increase in the number of inhabitants in the area ofthe Town ofMali Lošinj (specifically only in the settlements of Mali Lošinj and Veli Lošinj). The reason for this is the reduction of emigration on the one hand and immigration due to the development oftourism on the other hand, which manifested itselfas a positive migration balance and a positive natural change (Lajić 1993). In the area of the Town of Cres, during this intercensal period, a decrease in the total number ofinhabitants was recorded, because there was no significant development oftourism, and there was a further decrease in the number of inhabitants in most settlements (except for the settlements ofCres, Filozići, Miholaščica and Loznati, where an increase in the number ofinhabitants was recorded). In the second indicated intercensal period (1981−1991), a significant increase in the total number ofinhabitants was recorded (13.9%), and the result was an increase in the population of the areas of the Town of Cres and the Town of Lošinj.

Census trends (1 948-2021 )
Census trends during the 1948−2021 period are observed separately in this study. Namely, after World War II, this area was affected by the process ofdeagrarization, i.e. the abandonment of farming and the transition to other activities (Crkvenčić 1992), primarily related to tourism in this area. As a result, significant spatial changes in demographic trends were recorded in this period − very strong emigration and then immigration processes, which took place differently in the research area.
In the first post-war years, the process of decline in the total number of inhabitants continued in the Cres-Lošinj area as a whole, as well as in the area ofthe Town ofCres and the Town ofMali Lošinj (Table 1). It was conditioned primarily by emigration due to relatively high agrarian density, and then by the emigration of the mainly Italian population due to the return of the islands ofCres and Lošinj to the territory ofCroatia. In addition, at the end ofthe 1960s, just as people emigrated from other parts ofCroatia, part ofthe Cres-Lošinj population also emigrated to Western European countries, and emigration from the small islands ofthe Lošinj area to overseas countries was also noticeable (e.g. from the Island ofSusak to the USA) (Podgorelec 1999).
The turn from strong emigration to immigration began in the 1960s, when immigration to the island of Lošinj was recorded, and in the 1970s when immigration to the Island ofCres was recorded. When it comes to the changes in the total number of population, this is evident in the increase in the number ofinhabitants, first of the Town ofMali Lošinj (intercensal period of1971−1981) and then of the Town of Cres (intercensal period of 1981−1991). The reason for this is the construction of modern hotels, which created the need for labor in the fields ofconstruction and tourism, but there was also the emigration ofa large number ofthe local population because there were no jobs offered for other professions (Podgorelec 1999). In the last few censuses, a large share of the immigrant contingent has a "fictitious population", as is the case with other Kvarner islands (e.g. Lajić and Mišetić 2013b, Jovanić and Turk 2013), but despite this, since the census year 1991, a continuous population decline is recorded for the entire Cres-Lošinj area, as well as the towns of Cres and Mali Lošinj. Still, according to the latest population census (2021), on the island territory ofthe Republic ofCroatia, with 7,565 inhabitants, the Town ofMali Lošinj is the most populated local self-government unit, and the settlement Mali Lošinj (5,577 inhabitants) the most populated settlement.
Observed by the level ofsettlements, it is noticeable that according to the last census (2021), all settlements, except for the town of Mali Lošinj, have fewer inhabitants than in 1948 (Figure 4). The centers of local selfgovernment units and settlements closer to them have more favorable indicators, while settlements further away show less favorable indicators. Also, it is evident that the settlements in the area of the Town of Mali Lošinj have more favorable indicators than in the area of the Town of Cres, where according to the latest census, two settlements are without inhabitants (the settlements ofStanić and Važminec).
In the last three censuses in the Cres-Lošinj area as a whole, at the level of the self-governing unit and in almost all settlements ofthe Cres-Lošinj area, a continuous decrease in the number of inhabitants is recorded. This is recorded despite a slight increase in the number ofinhabitants in the intercensal period of 1971-1981 in the area of the Town of Mali Lošinj, and in the intercensal period of1981-1991 in the area ofthe Town ofCres and the Town ofMali Lošinj, conditioned primarily by the increase in the population oflarger, i.e. coastal settlements (e.g. Cres, Mali Lošinj and Veli Lošinj) related to the emergence and development oftourism, but also the socalled "fictitious" population census. Only in the last Slika 4. Indeks promjene broja stanovnika cresko-lošinjskog područja (2021 /1 948).  census can a slight increase in the number ofinhabitants be seen in several smaller settlements, probably due to the so-called "fictitious" census 8 .
The negative natural change that has been present on the Kvarner islands for several decades (Lajić and Mišetić 2013b) and the strong long-term emigration from the settlements in the interior ofthe island ofCres and from all the islands of the Lošinj archipelago have resulted in a large number of elderly, single and abandoned households (Podgorelec 1999) and a large number of very small settlements. According to the latest population census from 2021, only Cres and Mali Lošinj are settlements with over 2,000 inhabitants. The next largest towns (Veli Lošinj and Nerezine) have over 300 inhabitants, while the remaining settlements are smaller. This means that only a total ofeight settlements, or 20.0% of the settlements in the Cres-Lošinj area, had more than 100 inhabitants. There are a total of six of them in the area of the town of Mali Lošinj, while two are in the area ofthe town ofCres, where the disparity in the size ofthe settlements is particularly pronounced. Cres (2,205 inhabitants) has several times more inhabitants than the remaining 25 settlements, with only one settlement larger than 100 inhabitants − Martinšćica (106 inhabitants).

Shifts in mean centers of population
As described above (Section 2.2), different methods were used in this study in order to spatially analyze and determine the mean centers of settlements in the Cres-Lošinj area and changes in demographic characteristics. In addition to the method in which only the position of the settlement is entered, whereby a point is obtained that actually represents the central position in relation to all the observed settlements, this study also used a method ofspatial analysis which, in addition to entering the position of the settlement, also enters the census data (number ofinhabitants) ofeach settlement.
Taking into account only the location of the settlements in the Cres-Lošinj area as a whole, the results showed that the central location of the settlements is located in the central part of the island of Cres, within the local self-government unit ofthe Town ofCres (Figure 5). On the other hand, taking into account the location ofthe settlements and the number ofinhabitants of each settlement in each census year, it is evident that the mean center of the population in each observed census year is located south ofthe central location ofthe settlement and is located on the southern part ofthe island of Cres, within the local self-government unit of the Town of Mali Lošinj. This shows that a larger number of smaller settlements are located in the northern part ofthe research area, while larger settlements (with a larger number of inhabitants) are located in the southern part.
Observing the shifts in the mean center ofthe population ofthe Cres-Lošinj area as a whole, it is evident that from the first population census (1857) to the mid-20th century (1948), the tendency was towards the north. The reason for this is primarily that in the southern area (within the local self-government unit ofthe City ofMali Lošinj) there was a strong decrease in the number ofinhabitants (for example, the settlement of Mali Lošinj decreased by 2,718 inhabitants, and Veli Lošinj decreased by 898 inhabitants). Namely, Sectioni 3.1 describes that until 1910, an increase in the number of inhabitants was recorded in the Cres-Lošinj area as a whole, as well as in both local self-government units, which was conditioned primarily by an increase in the birth rate. Figure 5 shows that during this period there was no major shift in the mean center ofthe population. However, after 1910 there was a decrease in the number of inhabitants, primarily due to emigration. From the movement of the mean center of the population towards the north from 1910 to 1948, it is evident that this decrease in the number ofinhabitants (emigration) was more pronounced in the southern part of the research area, that is, less pronounced in the northern part ofthe research area.
From the mid-20th century, more precisely from 1948 to 1991, there was a significant shift in the mean center towards the southern part of the research area. Namely, as described above (Sections 3.1 and 3.2), a decrease in the number ofinhabitants was recorded from 1948 to 1971, and from the movement ofthe mean center of the population towards the southwest, it is evident that this decrease was more pronounced in the northern part of the research area, that is, less pronounced in the southern part. In the following intercensal periods, from 1971 to 1981 and from 1981 to 1991, in conditions of population growth in the Cres-Lošinj area as a whole, the mean center ofthe population continued to move towards the southwest. In the intercensal period from 1971 to 1981, the movement of the radnom snagom iz područja građevine i turizma, ali uz iseljavanje velikog broja domaćeg stanovništva jer se ne nude radna mjesta za druge struke (Podgorelec 1999). U posljednjih nekoliko popisa velik udio u useljeničkom kontingentu ima "fiktivno stanovništvo", kao što je to slučaj i kod ostalih kvarnerskih otoka (npr. Lajić i Mišetić 2013b, Jovanić i Turk 2013), ali unatoč tome se od popisne godine 1991. bilježi kontinuirani pad broja stanovnika cresko-lošinjskog područja u cijelosti, kao i Grada Cresa i Grada Malog Lošinja. Ipak, prema posljednjem popisu stanovništva (2021.), na otočnom području Republike Hrvatske Grad Mali Lošinj je sa 7565 stanovnika najnaseljenija jedinica lokalne samouprave, a grad Mali Lošinj (5577 stanovnika) najnaseljenije naselje.
Promatrajući kretanje težišta naseljenosti creskološinjskog područja u cjelini, vidljivo je kako je od prvog popisa stanovništva (1857) do sredine 20. stoljeća (1948) tendencija pomicanja bila prema sjeveru. Razlog je tome prvenstveno što je na južnijem području (unutar jedinice lokalne samouprave Grada Malog Lošinja) došlo do snažnog pada broja stanovnika (npr. naselja Mali Lošinj pad za 2 718, a Veli Lošinj pad za 898 stanovnika). Naime, u prethodnom je poglavlju (poglavlje 3.1) opisano kako je do 1910. godine zabilježeno povećanje broja stanovnika na cresko-lošinjskom području u cijelosti, kao i u objema jedinicama lokalne samouprave, a uvjetovano je prvenstveno povećanjem nataliteta. Na slici 5 vidljivo je mean center ofthe population towards the south is the result of the increase in the number of inhabitants recorded in the area ofthe Town ofMali Lošinj (more precisely, only in the settlements of Mali Lošinj and Veli Lošinj) and the decrease in the total number of inhabitants in the area ofthe Town ofCres. In the second indicated intercensal period (1981−1991), an increase in the total number of inhabitants was recorded, both in the Town of Cres and the Town of Lošinj, as well as in the Cres-Lošinj area as a whole, and from the shift of the mean center of population towards the southwest, it is evident that the increase in the number of inhabitants was more significant in the southern part of the research area.
In the following census years (2001, 2011 and 2021), a decrease in the total number of inhabitants was recorded both in the Cres-Lošinj area and in both local self-government units. From the slight movement of the mean center ofthe population towards the north, it is evident that this decrease in the total number of inhabitants is less pronounced in the northern part, and more pronounced in the southern part, where the majority ofthe population is located.
Među jedinicama lokalne samouprave zamjetne su razlike kretanja težišta naseljenosti (1857−2021) koje se odnose na udaljenost i smjer pomicanja (slika 6). Tako je za Grad Cres to kretanje težišta naseljenosti na vrlo maloj udaljenosti (oko 2,5 km) i pomiče se u smjeru zapadistok, a za Grad Mali Lošinj je na nešto većoj udaljenosti (oko 5 km) i pomiče se u smjeru sjeverozapad-jugoistok. Središte grada/općine / Municipal center Granica grada/općine / Municipal border population is that in the observed period (1857−2021), and particularly the last observed year (2021), the mean centers ofthe population are very close to the center of the town/municipality. Also, compared to the shifts in the mean center ofpopulation in the Cres-Lošinj area as a whole (Figure 5), where it is evident that the mean center of population of each observed census year is located at a greater distance (10 to 20 km) from the central location ofthe settlement, for the Town ofCres and the Town ofMali Lošinj it is evident that the mean center of population is closer to the central location of the settlement.

Population direction trends
In addition to the mean center method of spatial analysis, for the purpose of determining and spatially analyzing the demographic characteristics of the Cres-Lošinj area, this paper also applied the standard deviation ellipse (directional distribution) method ofspatial analysis. As described above (Section 2.2), the result of the application ofthis method is an ellipse ofsettlement distribution and an ellipse for each observed census year, which differ according to position, shape and direction ofextension.
The results showed (Figure 7) that the ellipse of settlement distribution is located in the northern part of the research area, because a larger number of settlements are located here, that is, a smaller number ofsettlements are located in the southern research area.
Observing the ellipses of the observed census years, the results showed that they are located further south than the ellipse of settlement distribution. The reason for this, as with the previously obtained results related to the mean center of the population, is the smaller number oflarger settlements located in the southern part of the research area. Due to these, the ellipses for the census years are located further south, i.e. the population ellipse is located further north due to the greater number of smaller settlements in the northern part of the research area. Also, as with the previously obtained results related to the mean center of the population, of the observed census years, the ellipse for 1948 is the most northern, while the ellipse for 2021 is the most southern. This is in accordance with the trend in the number ofinhabitants ofthe town ofMali Lošinj, where in 1948 the lowest number of inhabitants was recorded until then, which was followed by a strong growth (by 2,669 inhabitants). Observing the shape, it is evident that the ellipse for 2021 has the most elongated shape, which speaks in favor of the fact that larger settlements (Mali Lošinj, Cres and Veli Lošinj) have an increasing influence on the direction ofthe population. This is also confirmed by the ellipses for 1948 and 2021 at the level ofthe local self-government units ofthe Town ofCres and the Town of Mali Lošinj, where it is evident that the population is concentrated closer to the larger settlements, i.e. the centers of the local self-government units. When observing the remaining settlements, it can be said that due to the smaller increase, that is, due to the larger decrease in the number of inhabitants, their influence on the direction ofthe population is decreasing.

Discussion and Concluding Remarks
This paper presents the complex issue of the demography of the Cres-Lošinj area, which was carried out through a spatial analysis of demographic characteristics by using a GIS model that is comprised of the data and spatial analyses used. In this paper, the methodology was used, as part ofwhich each analyzed phenomenon is observed on at least three levels With regard to the changes in the total number of inhabitants of the Cres-Lošinj area as a whole, but also at the level of the local self-government units (the Town of Cres and the Town of Mali Lošinj) in the period from the mid-19th century to the present day, the existence of two time intervals was established: the interval of demographic expansion with a predominant but discontinuous increase in the total number ofinhabitants (1857−1910) and an interval of demographic regression with a predominant but discontinuous decline in the total number of inhabitants (1910−2021). In the last population census (2021), the number ofinhabitants was determined to be more than twice as small as in 1910, when the highest recorded number ofinhabitants was recorded both at the level of the Cres-Lošinj area as a whole, and at the levels oflocal self-government units and most settlements. The only exception to this is the Town ofMali Lošinj, which recorded more inhabitants in the last census than in 1910 (5,530 inhabitants in 1910; 5,567 inhabitants in 2021).
The period after World War II (1948−2021) was examined in more detail, where in the first decades depopulation was established primarily due to emigration, which occurred as a result of the abandonment of agriculture, i.e. the processes ofdeagrarization and deruralization. In the intercensal periods of 1971-1981 and 1981-1991, an increase in the total number of population was recorded due to immigration related to tourism, particularly in larger coastal settlements (e.g. the Town of Mali Lošinj, the Town of Cres), but with simultaneous emigration ofa part ofthe local population who could not be employed in various activities of the secondary and tertiary sectors in the Cres-Lošinj area. This Međutim, zajedničko je obilježje kretanja težišta naseljenosti to da su u promatranom razdoblju (1857−2021), a pogotovo posljednje promatrane godine (2021), vrlo blizu sjedištu grada/općine. Također, u usporedbi s kretanjem težišta naseljenosti na cresko-lošinjskom području u cjelini (slika 5), gdje je vidljivo da se težište naseljenosti svake promatrane popisne godine nalazi na većoj udaljenosti (10 do 20 km) od središnjeg položaja naselja, za Grad Cres i Grad Mali Lošinj je vidljivo da su bliže središnjem položaju naselja.
Promatrajući elipse promatranih popisnih godina, rezultati su pokazali da se one nalaze južnije od elipse razmještaja naselja. Razlog tome je, kao i kod prethodno Slika 7. Trend usmjerenosti naseljenosti cresko-lošinjskog područja u cjelini i na razini grad/općina (1 857-2021 ). Fig. 7 Population direction trends in the Cres-Lošinj area as a whole and at the town/municipality level (1 857-2021 ). means that the long-term depopulation in this area is brought into complex interrelationships with the processes of littoralization, deruralization, deagrarization, the affirmation ofnew directions ofEuropean migration, the expansion of tourism, the general aspirations of education and specialization, etc. (Lajić 1993). In the last three censuses, there is a decrease in the total number of inhabitants in the Cres-Lošinj area as a whole, at the level of the self-government units and in most settlements. Only in the last census can a slight increase in the number ofinhabitants be seen in several smaller settlements, probably due to the so-called "fictitious" census. The result ofall ofthe above is that in the last census, all settlements, except for the Town of Mali Lošinj, recorded fewer inhabitants than in the mid-20th century. As a result ofthe above demographic trends, this paper analyzed the spatial shift ofthe mean center ofthe population and the ellipse, which were observed in the period from the mid-19th century onwards. The analysis ofthe shift ofthe mean center ofthe population (carried out by using the Mean Center method of spatial analysis) determined for the Cres-Lošinj area as a whole that from the first population census (1857) to the mid-20th century (1948) the tendency was towards the north, the central position ofthe settlements. The reason for this is primarily that in the southern area (within the local self-government unit ofthe City ofMali Lošinj) there was a strong decrease in the number of inhabitants. From the mid-20th century, more precisely from 1948 to 1991, there was a significant shift in the mean center towards the southern part of the research area. In the following census years (2001, 2011 and 2021), a decrease in the total number of inhabitants was recorded both in the Cres-Lošinj area and in both local selfgovernment units. Given that there has been a moderate movement of the mean center of the population towards the north, it is evident that the decrease in the total number of inhabitants of the Cres-Lošinj area is less pronounced in the northern part, that is, it is more pronounced in the southern part. For the area of local self-government units, a similar shift ofthe mean center of population was also determined, whereby the mean center for 1948 was closest to the central location ofthe settlement, for 1857 it was a little further away, and for 2021 it was the furthest away from the central location ofthe settlement.
Using the standard deviational ellipse (directional distribution) method, it was determined that larger settlements (Mali Lošinj, Cres and Veli Lošinj) have an increasing influence on the population direction, and it was determined that the population is concentrated closer to the larger settlements, i.e. to the centers oflocal self-government, while at the same time the population of a large number of small settlements in the northern part ofthe research area is decreasing more and more.