Dynamics of land cover change, regional development, and its local dependence driving factors in Bojonegoro Regency

Bojonegoro is a regency in East Java province with a population of 1,322,474 (BPS, 2022) and high population density. Population growth is strong, and demand for land is increasing. Increased human activity leads to land conversion, which could impact regional development. The objectives of this study are to: 1) Analyze the dynamics of land cover change, 2) Compare the SDI (Sub-district Development Index) scores of 2008 and 2020 in Bojonegoro Regency, and 3) Analyze the factors affecting the level of Regional development in Bojonegoro Regency. The methods used are overlay (GIS analysis), scale plot analysis, and geographically weighted regression (GWR) analysis. The most significant land cover change occurred in forest cover, with the area decreasing from 99,815 hectares (2000) to 84,845 hectares (2020), while the built-up area increased from 12,665 hectares (2000) to 22,901 hectares ( 2020). Bojonegoro sub-district has the highest SDI of 44,043 (2008) and 44,917 (2020) because it has the largest population and become a capital district and central business district (CBD). Based on the results of the GWR analysis, the highest local R2 is located in the eastern part of the Regency. In contrast, the lowest local R2 is located in the western administrative district of Bojonegoro Regency. The results show that the driving forces influencing regional development in Bojonegoro Regency vary spatially.


Introduction
Bojonegoro Regency has a population of 1,322,474 [1] with a reasonably high population density.Rapid population growth leads to solid dynamics of land cover change.Land cover change is the increase of one type of land cover from one type to another over a certain period [2], where the change is irreversible (irreversible); for example, forests converted into built-up space (e.g.settlements) would be difficult to convert back to forestry [3] In addition, agricultural land is often converted to construction land [4][5].
Changes in land cover due to population growth are also influenced by the socioeconomic conditions of landowners and government policies [6].Another factor leading to land cover change is the decline in rice productivity.Every year, many rice field owners convert their rice fields to other uses to bring more economic benefits.[7].Increasing activities and population, number and type of facilities can increase regional growth and demand for land [8][9].Therefore, land use change is one of the impacts on the growth and development of the territory.According to research [10], increased demand for land use leads to land conversion.
Population growth and various activities will encourage an increase in the number and type of facilities that characterize the region's development.The region's development must stimulate economic growth and social welfare, further optimize the use of the region's potential and meet the aspirations of the community in developing the territory according to the region's goals.This is the goal of regional autonomy policy.The region's development is reflected in the availability of infrastructure, geographical conditions, population growth rate and socioeconomic changes.
Rapid development in Bojonegoro Regency since 2000 may have impacted some land cover changes.Rapid physical changes can also impact regional development, including facilities and infrastructure, as well as emerging trade and service sectors.Therefore, it is necessary to analyze land cover change dynamics [11].Based on this context, it is necessary to study the dynamics of land cover change and the level of regional development and identify the factors influencing regional development in Bojonegoro Regency that are considered essential.Critical in policy formulation for research planning and development areas.This study's novelty is identifying local spatial factors influencing regional development in Bojonegoro Regency using the GWR model, which has never been done in any previous study.The local regression coefficient generated by the parameter estimates in the GWR model can be applied to see the influence of each explanatory variable (X) on the regional development (Y) level in the local area.Bojonegoro rights.Therefore, the results of this analysis can be used to develop more targeted policy guidelines/recommendations to improve regional development in this region, as each variable's influence is different at each location.

Method
This study was conducted in Bojonegoro Regency, East Java, as shown in Figure 1.Bojonegoro Regency is a district in East Java Province consisting of 28 sub-districts, including 11 sub-districts and 419 villages.This district is at coordinates 60 59' -70 37' South latitude and 1120 25' -1120 09' East longitude.The administrative boundaries lie in the north of Tuban district, the south of Nganjuk, Madiun, and Jombang districts, the east of Lamongan district and the west of Ngawi district and Blora district [12].

The dynamics of land cover change
Land cover changes were analyzed using the overlap technique.The overlay technique used includes land cover maps of the Bojonegoro Regency for the years 2000, 2009 and 2020, along with the administrative boundaries of the area, creating maps of the land cover change types of the Bojonegoro Regency in 2000, 2009 and 2020.Then, the area per land area per year was calculated and compared with each time in the year of observation.

Level of regional development
Regional development analysis uses the scalogram method to determine the development index value of sub-districts.Using scalogram analysis, sub-areas' roles are determined based on each sub-area's ability to provide services to the community.It contains information about all regional center names, populations, types, and service facilities stored in raster format.[13][14].Scalogram analysis also provides insight into grouping facilities into service centers based on the comprehensiveness of their service functions [15].
This study uses two-year points, specifically 2008 and 2020.SDI is a method that Indonesian researchers often use as a proxy for gross regional domestic product (GRDP) because it compares the number of facilities with the number of residents.These include education, health, economics, and government services [16][17].Data was obtained from the Indonesian Central Statistics Agency.This index is calculated using a weighted histogram with three steps.First, determine the weight of the characteristic index using the formula: Where ij is the object index's weight, each j's weight in region i,  ij is the number of j instances in region i.Facilities include education, health and economics, both public and private.i = 1, 2, 3, . . ., n (region number); j = 1, 2, 3, , p (base number).Then, normalization of the new variable data is performed using weighted data (ij) with the formula: "K ij " is the raw value of the scalogram index, min  j is the minimum value of  j setting j, and  j is the standard deviation value of j.Finally, the  was calculated as follows:

Geographically weighted regression (GWR) analysis
The GWR model is a regression model converted into a weighted regression model [18].Variable Y is estimated using predictor variables, each with a regression coefficient that depends on where the data are observed.Variable of the dependent and independent in this research are presented in Table 1.The GWR model is a commonly used method of statistical to analyze spatial heterogeneity, in which the same independent variable will give different responses at different locations within the study area.[19].GWR analysis is used to find the spatial diversity of the independent variable concerning the dependent variable to provide insight into local regression parameters based on geographical area [20].
Several studies have applied the GWR model to identify regional differences in regional development [21], natural resources [22][23], social conditions [24][25], and urban expansion areas [10,20].The result of this analysis is a regression model whose parameter values apply only to observed locations and vary depending on location [26].GWR analysis was performed in 2020 with the following GWR formula [27]: Where  is the variable associated with observation j,  is the independent variable  at location j, ,  are the coordinates of observation location j, , (, ) are the coordinates d of the observation location observation j and  ( ,  ) are the coefficients of regression or local parameter estimates for the variable of independent  at location j.Optimal GWR bandwidth analysis is determined by minimizing the corrected Akaike information criterion using a limited sample correction.This research uses data from the Central Statistics Agency.

Dynamics of land cover change
Land cover is the physiological condition on the earth's surface as observed, and it describes the vegetation and artificial structures that cover the earth's surface [28].Moreover, it is also the result of the arrangement of human arrangements, activities, and ways of dealing with certain conditions types of land use to carry out production, conversion and maintenance of area [29].Meanwhile, land cover changes refer to the state of the earth; the conditions of the earth change over time due to human activities [30].Ground cover can be divided into nine types: Forests, construction lands, fields, plantations, mining, dry farming, paddy fields, shrubs and water bodies.Land cover was analyzed using three-year points, specifically 2000, 2009 and 2020.Land cover maps of Bojonegoro Regency for the years 2000, 2009 and 2020 are shown in Figure 2. The results of changes in land cover area are presented in Table 2.In 2000, 2009 and 2020, the ground cover was mainly forest, with an area of 99,815 hectares, 98,834 hectares and 84,845 hectares, respectively.Bojonegoro Regency is geographically 40% forested.The forests in Bojonegoro Regency often face forest fires almost yearly around July, a dry month [31].East Java province still relies heavily on forests as its river basin and agricultural land are the most important sector and is considered one of the national food basket [32].From 2000 to 2009, the decreased land area was forests, dry agriculture, rice fields and water bodies.Land cover increased in area to include construction land, bare land and shrubs.However, there is one type of land use that has a fixed surface area, which is planted forests.From 2009 to 2020, the land area also changed.Area decline occurs in forests, open fields, plantations and rice fields.At the same time, an increase in area occurred in construction areas, mines, dry farms, shrublands and water bodies.Mining activity is increasing because the mining industry's potential in East Java is good enough to increase income in the region and create jobs.Dry agriculture is increasing, but the area of rice fields is decreasing.The decrease in rice field area corresponds to an increase in buildable surface area [32].Four processes can cause land cover change: The expansion of city boundaries, the modernization of city centers, the expansion of infrastructure networks, and the development or loss of community activity centers [33].Factors influencing land cover change include topography, population, land value, accessibility, infrastructure, and environmental sustainability [34].Factors affecting land cover change, human behaviour and actions are the determinants of land cover change [35].One of the factors causing land conversion is the increasing population [36].This growing population leads to a significant need for space, so the conversion of rice fields to settlements (colonies) frequently occurs [37].
The impact of changing the use purpose from agricultural land to non-agricultural land and changes in people's livelihoods leads to weakening the intensity of community-based social relationships that are the basis of identity.Of agricultural society [38].Therefore, livelihood changes require adjustments to adapt to the situation to meet their needs [39].Furthermore, it leads to reduced rice yields and reduced food supplies [40], can damage soil conditions [41] and damage the environment [42].

Level of regional development
The level of regional development of Bojonegoro Regency was analyzed using central government infrastructure and accessibility indicators, such as educational, health and economic facilities.This analysis uses scores from two years, specifically 2008 and 2020.The results of the histogram analysis in Bojonegoro Regency can be presented in Table 3.  scores have populations that continue to grow.However, the number of amenities and accessibility is still low, so the development of common areas is still low.Differences in ownership of public service facilities between districts significantly impact a region's attractiveness as a growth center [43].The more extensive the economic and social facilities, the more attractive it is for residents to live there [44].
The development of a territory is influenced by population growth and the variety of service activities, which in turn increase the number and type of establishments in the area [45][46][47].The development of an area is influenced by many factors, including geography, amount of infrastructure [13,[48][49], population, distance from the city center [50], capacity for suitable access, climate and abundance of natural resources [51].Additionally, the number and type of settlements in an area are often closely related to population [52].
The analysis results can be seen on the spatial distribution map of SDI values for Bojonegoro Regency in Figure 3

Local spatially driving factors of regional development in Bojonegoro Regency using GWR analysis
The results of the GWR analysis are presented in Figure 4.The local value of R2 is obtained in the range of 0.932585-0.932814;this value indicates that more than 90% of the independent variables can describe  The variables educational establishments (X1), number of medical establishments (X2), number of economic establishments (X3) and number of occupations (X4) all show a positive relationship with regional development.At the same time, the number of floods is changing (X5), negatively affecting the region's development.A positive value of the coefficient indicates a higher value of the variable, the greater the influence of the level of regional development.At the same time, a negative value of the coefficient means that the value of the variable will be inversely proportional to the level of development of the area.The more facilities/infrastructure there are, the better the level of development of the area.On the contrary, the area's development will be poorer if floods occur frequently.The impact of the variables can be seen from the darkness of the colours.The darker the colour, the greater the impact on the area's development.The coefficient of determination (R2) has a range from 0 to 1 (0 < R2 < 1).Suppose R2 has a higher value (close to 1).In this case, this means that the simultaneous influence of the independent variables is considered necessary, and if (R2) is close to zero (0), the degree of influence of the independent variables on the characteristics of the dependent variable is considered to have low concurrency [53] .The regression coefficient value of the independent variable are presented in Figure 5.The value of the regression coefficient with the highest positive variable is the number of economic establishments ranging from 0.36336 to 0.3643.This coefficient shows that an increase in regional economic bases will increase the region's development.The highest local parameter estimates (regression coefficient values) are distributed across the Baureno, Kepohbaru, Sumberrejo and Kanor sub-districts.At the same time, the lowest value of the regression coefficient is related to the number of industries.There are still some industrial parks in Bojonegoro.Based on the results of GWR, the growing number of industries in Bojonegoro Regency is low, affecting the region's development.Industrial development will impact and change communities' socioeconomic conditions [54][55][56].
The negative value of the regression coefficient is -0.008733 to -0.008313 in the flood variable.If more flooding occurs in Margomulyo, Ngraho, Tambakrejo and Sekar sub-districts, it will affect the decline in regional development of the areas.The incidence of flash floods in the area is due to deforestation.

Conclusion
Bojonegoro Regency continued to experience land cover changes from 2000 to 2020.The most significant land cover change occurred in forest cover, with an area decreasing from 99,815 hectares (2000) down to 84,845 hectares (2020), while the area of construction land increased from 12,665 hectares (2000) to 22,901 hectares (2020).The level of regional development in Bojonegoro Regency has also increased over time.Judging by the Sub-district Development Index (SDI) scores at the subdistrict level, the Bojonegoro sub-district has the highest SDI scores, namely 44,043 (2008) and 44,917 (2020), as this sub-district also has a large population.This is the CBD and the center of government and community activities.The highest local R2 value produced by GWR analysis is 0.932769 to 0.932814, distributed in the Eastern region, especially Baureno, Kepohbaru, Sumberrejo and Kanor subdistricts, while the local lowest of R2 is 0.932585 to 0.932631.widespread in the western region of Margomulyo, Ngraho and Tambakrejo sub-districts.The distribution pattern of local R2 values shows that the area is well developed, shown in dark red.The GWR analysis results show that each variable's influence level in each location is different.Variables with coefficients that have a positive impact on regional development are the number of educational establishments, the number of medical establishments, the number of economic establishments and the number of industries.Meanwhile, floods, truly manufactured disasters, harm the region's development.

Figure 2 .
Figure 2. Maps of the land cover in Bojonegoro Regency in 2000, 2009, and 2020.

Figure 3 .
Figure 3. Maps of spatial distribution SDI values Bojonegoro Regency in 2008 and 2020.
affecting regional development in Bojonegoro Regency.The highest local R2 0.932769-0.932814is spread in the eastern region, namely the sub-districts of Baureno, Kepohbaru, Sumberrejo and Kanor.While the lowest local R2 is 0.932585-0.932631,spread in the western region of Margomulyo, Ngraho and Tambakrejo sub-districts.The regional development of the eastern part of Bojonegoro Regency is better than the western part due to the continuous change in land cover from rice fields to built-up land, causing the area to develop more.The distribution pattern of the local R2 value can be seen in that the area is well-developed, shown in dark red.

Figure 5 .
Figure 5. Local parameter estimates of independent variable (Cn) by GWR model.

Table 1 .
Independent variables (X) included in GWR models.

Table 3 .
The SDI value for the Bojonegoro Regency area in 2008-2020.
SDI, which is presented in Table3.Table3shows that several sub-districts experienced changes in the SDI score.In 2008, the minimum, maximum, and average values were 6.202, 44.043, and 16.768.In 2020, SDI values increased to 6.118, 44.917, and 17.895.The analysis results can be seen in the map of the spatial distribution of SDI values for Bojonegoro Regency in Figure3.