Analysis of land use/land cover change and erosion hazard levels in the social forestry area of KPH Ulubila, South Sulawesi Indonesia

Changes in land use/land cover are part of the dynamics of nature, both changes caused by natural disasters and human activities. The need for land by humans encourages the acceleration of land cover change, especially from forests to non-forests. The rate of change in land cover will affect the hydrological state so that the land becomes very critical. One of the things that causes land to become critical is erosion. One of the policies carried out to reduce the rate of change in land cover, reduction of erosion area and empowerment of communities around forests is the Social Forestry program. Social forestry refers to any forest utilization activity by the community in providing products for their own use or generating local income. The case study in this study is the social forestry of KPH Ulubila. The results of the analysis in 2013 and 2022 showed that the highest decline in the agricultural/crop land class with an area of 87.96 ha or equivalent to 89.20% and secondary second forest was 681.86 ha or equivalent to 27.50% of the area of social forestry. In other classes, there was an increase in agricultural land, namely in the plantation class of 267.83 ha or 467.19%, agricultural mixed of 160.73 ha or 7.20% and paddy area of 25.31 or 18.42%. In the condition of plantation forests, there was an increase of 35.43%, shrubland of 23.05% and open land of 400.94%. The rate of change in land cover has affected the magnitude of the change in the extent of erosion. This change leads to poor conditions because the area of the very low (<15 tons / ha / year) decreases by 45.94% and low (15-60 tons / ha / year) 2.19%. The addition was actually seen in the very heavy (>480 tons/ha/year) 21.57%, moderate (60-180 tons/ha/year) 28.29% and heavy (180-480 tons/ha/year) 6.45%.


Introduction
Changes in land use/land cover are part of the dynamics of nature, be it changes caused by disasters or human activities.According to Guan et al. (2011), that land cover change is the main environmental change responsible for global change [1].The need for land for human activities is one of the main consequences in influencing land cover change.Due to the increasing number and human activities, land as an increasing resource area is changing, especially the conversion of forest land to non-forest.This is referred to as gross deforestation.From data released by the Ministry of Environment and Forestry of Indonesia (2021), Indonesia's gross deforestation in 2019-2020 reached 119.1 thousand ha.The largest deforestation occurred in secondary forest cover covering an area of 104.6 thousand ha IOP Publishing doi:10.1088/1755-1315/1230/1/012047 2 (87.8%), followed by primary forest of 12.3 thousand ha (10.4%) and plantation forest of 2.2 thousand ha (1.8%) [2].
This level of land cover change affects biogeochemistry, hydrology and the world's climate [3].The consequences of hydrological changes can be seen from the rate of change that continues to increase without conservation activities, making land critical.According to Didu in Indrihastuti et al. (2017), the definition of critical land between an institution and another institution varies in each point of view [4].From an agricultural point of view, critical land is associated with production (production) while from a forestry point of view it views critical land as a function as a regulatory medium for water management, a medium for producing results and as a medium for flood protection and/or downstream sedimentation.One of the things that causes land criticality is erosion.Erosion is the event of removal or removal of soil or parts of the soil from one place to another by a natural medium.In the event of erosion, the soil or part of the soil in a place is eroded and deposited which is then deposited elsewhere.Erosion and transportation of soil occurs by natural media, namely water and wind [5].From this, it can be seen that land cover change plays an important role in the erosion process.According to Waltner et al. (2020), in conditions of closure of agricultural land can contribute to the reduction of soil organic carbon so as to affect the quality of soil nutrients [6].
One of the policies carried out by the government in reducing the rate of land cover change, reducing critical land area and empowering communities around forests is the Perhutanan Sosial or Social Forestry program.In Regulation of the Minister of Environment and Forestry P.83/2016 [7] explained that there are five forms of Social Forestry, namely Hutan Desa/HD (Village Forest), Hutan Kemasyarakatan/HKm (Community Forest), Hutan Tanaman Rakyat/HTR (People's Plantation Forest), Hutan Adat/HA (Customary Forest) and Kemitraan Kehutanan/KK (Forestry Partnership).The term social forestry refers to various activities related to trees or activities in the forest carried out by landowners and community groups in providing products for their own use or generating local income.Social forestry may also include governments or other groups planting trees on public lands to meet the needs of local villages [8].Efforts made through Social Forestry activities from the results of research by Wulandari et al. (2014), in Taman Hutan Raya Wan Abdul Rachman, Lampung Province that community involvement in optimizing forest function from degradation by implementing agroforestry systems is quite high and recommends the sustainability of this social forestry [9].Through this success, is the success the same in the social forestry area in South Sulawesi?.
Judging from data from the Ministry of Environment and Forestry [2], it shows gross deforestation of South Sulawesi Province of 920 ha.When compared to other province in Indonesia, this figure is very low.However, this needs to be considered because considering its considerable influence, it is no exception in the research area, namely the KPH Ulubila Social Forestry area.KPH Ulubila is a newly formed KPH based on the decree of the Minister of Environment and Forestry: SK 371/MENLHK/SETJEN/PL.0/9/2020 [10] concering the Determination of the Area of the Protection Forest Management Unit of South Sulawesi on September 17, 2020 and entered the KPH Production type.The area managed by KPH is 64,625 ha.The agricultural and plantation sectors as the main livelihoods of the people in the UPT KPH Ulubila work area cause dependence on forests as a source of livelihood so that they are very vulnerable to forest encroachment.The tendency of community activities in managing forests that have an impact on changes in land cover in them.Broadly speaking, the Ulubila KPH Social Forestry area covers an area of 51,002 ha with the type of HD activities covering an area of 47,453 ha and HKm covering an area of 3,549 ha.Therefore, a form of research is needed on the quantity of land cover change and its relation to the amount of erosion in the Ulubila KPH Social Forestry area.

Social forestry
Social forestry is interpreted differently by various parties.In some countries social forestry is considered as an umbrella for various forms of community-based forest management or oriented towards improving community welfare [11].Hardcastle (1987) states that social forestry is a concept that is not new and has existed for centuries in continental European countries and in colonial times many countries under colonial rule (having forestry services that mainly served to establish forest protection and create "village forest areas" or urban firewood areas" [12].Social Forestry is stipulated in the Regulation of the Minister of Environment and Forestry P.83/2016 as a sustainable forest management system implemented in state forest areas or forest rights/customary forests implemented by local communities or customary law communities as the main actors to improve welfare, environmental balance and sociocultural dynamics in the form of Desa/HD (Village Forest), Hutan Kemasyarakatan/HKm (Community Forest), Hutan Tanaman Rakyat/HTR (People's Plantation Forest), Hutan Adat/HA (Customary Forest) and Kemitraan Kehutanan/KK (Forestry Partnership) [7].The five forms are as follows [13]: • Hutan Desa (HD) is a state forest managed by the village and utilized for the welfare of the village.
• Hutan Kemasyarakatan (Hkm) is a state forest whose main use is intended for community empowerment.• Hutan Tanaman Rakyat (HTR) is a plantation forest in production forest built by community groups to improve the potential and quality of production forests by applying silviculture in order to ensure the sustainability of forest resources.• Kemitraan Kehutanan (KK) is a collaboration between local communities and forest managers, holders of forest utilization business permits/forestry services, forest area loan and use permits, or holders of forest product primary industry business permits.• Hutan Adat (HA) in social forestry, the definition of customary forest is somewhat different from other 4 (four) social forestry schemes (HD, HTR, HKm, and KK) because this customary forest is not in a state forest area but is located in the right forest.HA are forests that are within the territory of indigenous peoples.

Land use/land cover (LULC)
Land is also seen as a biophysical characteristic whose appearance is often called land cover.Many definitions of land closure have been proposed by several researchers including Fithria et al. (2012) provides an understanding of land cover, which is defined as a biophysical object that has an area of coverage on the earth's surface [14] and according to Ellis in Roy and Roy (2010) LULC expansion refers to all physical and biological cover that exists above the surface soil including water, vegetation, vacant land and/or artificial structures [15].LULC affects the processes of infiltration, evapotration and magnitude of surface flow.Vegetation affects the volume of water entering rivers and lakes, into soil and groundwater reserves [16].

Erosion
Erosion is the process of eroding a soil or part of the soil by water or wind from one place to another.Soils eroded by surface flows will be deposited in places of water flow such as rivers, irrigation lines, reservoirs, lakes or river estuaries [5].Soil erosion affects dryland productivity which usually dominates upstream watersheds and will have a negative impact downstream [17].In erosion prediction, prediction technology began more than 80 years ago with the invention of the Universal Soil Loss Equation (USLE) method developed by Wischmeier and Smith Year 1987 [18].

Study site and materials
The research sites are located throughout the Social Forestry area of KPH Ulubula, Bone Regency, South Sulawesi Indonesia.KPH Ulubila has an area of 64,625 ha in 8 districts in Bone Regency, namely Mare, Tonra, Salomekko, Kajuara, Kahu, Patimpeng, Libureng, and Bontocani.The following is the extensive data on the Social Forestry program at KPH Ulubila presented in Table 1 and the spatial distribution in Figure 1.The material used in the study consisted of vector, raster and tabular data which can be seen in Table 2.

Methods
In this study, there are two main stages, namely land cover change analysis and erosion analysis.The methods on each of the analyzes are described as follows.
Where: n is sample size; Z is the coefficient of reliability or the value of the standard normal variable.The use of a Z value with a 90% confidence level is 1.65; p(1p) is population variation is expressed in the form of proportions; E is margin of error and N is the size of the population or the area of each land use/land cover.Comparison of classification results and reference data is carried out statistically using a confusion matrix.Countable accuracy consists of manufacturer producer's accuracy, user's accuracy, and overall accuracy.Then the calculation results are continued on the calculation of the Kappa index.The matrix can be seen in Table 3 and the following formula.
user's accuracy = overall accuracy = Where: A, B, C is reference data; A`, B`, C` is data from image classification; Xii is tasted data; ∑Xi+ is number of data in rows; ∑X+i is number of data in columns and N is total data tasted.The final stage is to make observations of changes.Land cover changes can be identified by comparing land use/land cover maps for 2013 and 2022.The method used is with change detection.Change detection is a analysis carried out to determine the rate/rate of land change at any time using remote sensing technology in determining changes in a particular research object between two or more time periods.

Erosion analysis.
To determine the magnitude of erosion (Ea), erosion determinant analysis was carried out, namely R (erosivity Index), K (erodibility index), LS (Slope length and slope index), CP (Vegetation cover index and soil conservation practice) using the USLE (Universal Soil Loss Equation) method developed by Wischmeier and Smith [19], with the following equation: Where: Ea is total land loss (ton/ha/year); R is rein erosivity factor; K is soil erodibility factor; LS is slope length and slope factor and CP is vegetation cover factor and soil conservation practice.Where each variable specified is an equation and provision as follows: • Where: R si monthly rainfall erosovity and (Rain)b is monthly rainfall in cm.The calculation of erosivity is carried out in each analysis using ten-year rainfall data where the analysis in 2013 uses data from 2003-2013 and 2022 using data for 2013-2022.
• F actors of soil erodibility whose magnitude depends on the type of soil.In obtaining the K value, the data used are derived from the Landsystem RePPProt Soil Classification 1986.
• The topographic index factors L and S, respectively, represent the influence of the length slope and slope steepness on the magnitude of erosion with the formulas of Moore and Wilson (1992) as follows: LS = Power ("Aspect"/22.13, 0.4)*Power (Sin("SlopeDeg")/0.0896, 1.3) ( Where: L si length and angale of the slope; Aspect is aspect slope and SlopeDeg is slope Gradien in degree.
• Vegetation cover factor and soil conservation practice actions the comparison between the magnitude of erosion of an area and vegetation cover with the management of certain plants and soils to the magnitude of erosion from identical soils without plants is shown in Table 4 below.

Land use/land cover accuracy assessment and classification
Tables 6 and 7 show the results of the error matrix in 2013 and 2022.In each of the results of the land cover classification showed an overall accuracy of 98.36% and 96.74% with a kappa coefficient of 97.74% and 95.54%.The results showed that in 2013 the lowest user accuracy was shown in the plantation class with a value of 66.67% and the lowest producer accuracy in the agricultural/crop land class with a value of 75%.In 2022 the lowest user accuracy and producer accuracy was shown in the agricultural of 75%.The low value formed was caused by almost the same condition of land cover sightings at the study site.From the results obtained, the Kappa value belongs to the good category and is acceptable for further use in the analysis of changes [20].The use of Google Earth data helps in validating past data with a high degree of accuracy [21].The description on the table is Sf is a Secondary Forest; Pf is the Plantation Forest; O is Open Land; P is the Plantation; U is a Urban Land; Ag is Agriculture; Am is Mixed Agriculture; Pa is the Paddy Area; S is Schrubland.The nine classifications are the result of the identification of land use/land cocver classes in 2013 and 2022.Visually and the distribution of land use/land cover classification can be seen in Figure 2.

Land use/land cover change
The influence of vegetation on the hydrological cycle is very important in the amount of volume of water that enters rivers, lakes and enters the soil as groundwater reserves [5].This influence will be felt when there is a massive change in an area.The results of the 2013 and 2022 land use/land cover classifications show nine classes with area per year and area of change seen in Table 8 below.The results of the analysis showed that the highest decline was shown in the agricultural class with an area of 87.96 ha or equivalent to 89.20% of the social forestry area.Similar conditions are shown in the secondary forest class which shows a decrease in area, namely 681.86 ha or equivalent to 27.50% of the social forestry area.These two land closures show a decrease in land area that is different from other classes which actually experienced an increase in agricultural land, namely in the plantation class covering an area of 267.83 ha or 467.19%, mixed agriculture 160.73 ha or 7.20% and paddy area 25.31 or 18.42%.In the condition of plantation forests there was an increase of 35.43%, shrubland by 23.05% and open land by 400.94%.
The condition of land change at the research site is quite significant and depends on the conditions of the activities of the managing community.There is a change in the form of cultivation of other commodity lands that triggers the decline of agricultural land is decreasing.In addition, the condition of secondary forests is reduced for arcancy purposes.This increase in interventions will affect the availability of different biophysical resources including soil, vegetation, water, animal feed and others [22].According to Sihalolo et al. (2007) over time, human behavior patterns undergo changes in various aspects including social, economic as well as views and utilization of natural resources [23].It can also be seen that the increase in residential area by 15.22% has implications for the increasing pressure of the population on land, due to the increasing need for land for their residences and land for other facilities to support it.In addition, there are violations in the development of residential areas in forest areas.
Other conditions will affect the hydrological cycle of vegetation, especially upstream, resulting in disruption of the natural function of the ecosystem (in this case the watershed ecosystem).This condition indicates that secondary forests are undergoing land conversion to non-forest.This change has an impact on changes in the hydrological cycle as a sign of an imbalance between components.This imbalance indirectly results in flooding and inundation downstream because it changes the rain pattern upstream.However, the results showed that there is a form of addition to the plantation forest class where there is a form of conservation efforts through rehabilitation activities.These results show that the state of land closure for agricultural activities is still very high.After the receipt of social forestry permits, agricultural activities in the area are getting higher and higher.This encourages less conducive social conditions in terms of forest conservation.Encroachment of forest areas resulting in biophysical degradation indicates that the development of social forestry is not optimal [24].The calculation of the length and slope steepness uses DEM data and equation 8 with consideration between the slope degree and the slope aspect.LS in the study area had a vulnerable of 4.01 between 64.17.

• Vegetation cover factor and soil conservation practice (CP)
This factor is based on the results of land use/land cover in 2013 and 2022 with the CP code in Table 4.

• Erosion Hazard Classification
Erosion map-forming variables based on the USLE method were combined using overlay analysis on ArcGIS.The results of these calculations are presented in Table 11 with the amount of erosion in 2013 and 2022.The results showed that there were changes in the wide distribution of each erosion hazard class.This change caused poor conditions because the area of the very low class (<15 tons/ha/year) was reduced by 45.94% and the low (15-60 tons/ha/year) was 2.19%.The addition is actually seen in the very heavy class (>480 tons / ha / year) 21.57%, moderate (60-180 tons/ha/year) 28.29% and heavy (180-480 tons/ha/year) 6.45%.The spatial spread can be seen in Figure 3.The development of erosion areas from 2013 to 2022 has increased quite massively where the area in the middle to very heavy class has increased.Arsyad (2010), said that erosion is strongly influenced by the type of land cover.The higher the value of factor C (plant management index), the greater the erosion produced [5].This is also in line with the results found by Damaneh et al. (2017), which show that land cover changes increase soil erosion in the areas studied [27].It can also be seen that changes in land cover to agriculture show an increase in erosion.If agricultural land expansion still occurs from forest and pasture cover conversion, and poor agricultural practices continue, soil erosion will increase significantly [28].A massive increase in soil erosion will have a negative impact on landslides, where from the results of research conducted by Hasnawir et al., (2017), the increase in landslides occurs along with changes in land cover from shrubs and forests to mixed dryland agriculture [29].

Conclusion
The conclusions obtained from the results of the study show a relationship between land use/land cover change and the rate of erosion.The land closure was formed based on the results of digitization on screens using Landsat 8 OLI imagery in 2013 and 2022.The classes formed are secondary forests, plantation forests, open land, plantations, urban land, agriculture, mixed agriculture, paddy area and shrubland.The results of the analysis showed that the highest decline was shown in the agricultural class with an area of 87.96 ha or equivalent to 89.20% and secondary forests with an area of 681.86 ha or equivalent to 27.50% of the social forestry area.In erosion areas, there are additions in the very heavy class (>480 tons/ha/year) 21.57%, moderate (60-180 tons/ha/year) 28.29% and heavy (180-480 tons/ha/year) 6.45%.The implementation of the social forestry program in the KPH Ulubila area is still not optimal in its implementation.It is hoped that the implementation of the social forestry program will have an impact on improving the welfare of rural communities and forest sustainability.
Rain erosivity is the ability of rain to erode the soil.The erosion value is obtained from the Lenvain (1975) equation by the formula: R = 2.22 (Rain)b1,36

Figure 2 .
Figure 2. Land use/land cover maps for 2013 and 2022.

Table 1 .
Social forestry area in KPH Ulubila.
RePPProt Land System Soil Data 1986 Erodibility data generation (erosion calculation variable builder data) 6. Peta Rupa Bumi Indonesia (RBI)/ Earth Map of Indonesia, scale 1: 50.000Research site administration data Postprocessing is the stage of processing image data into land use/land cover data through image interpretation activities.The method used in image interpretation is digitization on the screen including hue or color, texture, shape, pattern, size, shadow, association and site.The classification of land closures based on the "Technical Guidelines for the Interpretation of Medium Resolution Satellite Imagery for the 2020 National Land Clearing Data Update" issued by the Ministry of Environment and Forestry.After the two stages are completed, the accuracy rate on each resulting map is calculated.Accuracy tests were conducted to compare the digitization results with the state of the field in 2022 and the digitization results with the state of affairs in the field assisted by high-resolution GeoEye imagery in Google Earth Pro in 2013.The selection of accuracy test points is based on the Estok navitte Cowan formula calculated on each land cover class with the following formula: 3.2.1.Land use/land cover change.Analysis of land cover change is carried out in two stages, namely pre-processing and post-processing.Pre-processing is a raw data processing stage consisting of downloading Landsat 8 OLI Image data for 2013 and 2022 on the earthexplorer.usgs.govwebsite; geometric and radiometric corre; layer stacking using tapes 6, 5 and 4; and cropping image.

Table 4 .
[19]alue for various land use/land cover factors[17].Data making is carried out based on the year of analysis, namely 2013 and 2022 with a classification based on the level of erosion hazard based on the Regulation of the Minister of Forestry of the Republic of Indonesia Number: P.32/Menhut-II/2009[19].

Table 6 .
2013 land use/land cover accuracy assessment using confusion matrix.

Table 7 .
2022 land use/land cover accuracy assessment using confusion matrix.

Table 8 .
Land use/land cover change from 2013 to 2022.

Table 9 .
[25]osed land cover changes from 2013 to 2022.Erosion hazards Land change has implications for the amount of damage to natural landscapes.Agricultural practices change the state of land cover and therefore, can result in surface erosion of varying degrees or magnitudes.The calculation of the amount of rosion in the social forestry of KPH Ulubila was carried out using the USLE equation.The calculation of the magnitude of erosion uses the results of multiplication of rain erosion parameters, soil erodibility, slope length and slope steepness as well as land cover and conservation practices.•Erosivity(R)In the social forestry area of KPH Ulubila, the value of rainfall varies greatly.The use of rainfall data based on rainfall data of PERSIANN-CCS (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks -Cloud Classification System) satellite products from CHRS -University of Arizona is available starting in 2003 with a spatial resolution of 0.04 o and a temporal resolution of 1 hour[25].The rainfall used is carried out based on the last ten years of each analysis year i.e., 2003 to 2013 for the analysis of 2013 and 2013 to 2022 for the analysis of 2022.The distribution of station points used is the CSIRO station point with a resolution proximity of 4 km x 4 km with a number of points of 10.The erosion values at each station point are as follows.

Table 10 .
[26]ivity values in 2013 and 2022.Paleudult with a K value of 0.2; Rendolls with a value of 0.34 and Tropoquepts with a value of 0.28.This figure comes from the land data of the RePPProt Landsystem 1986.The highest value of the K valus indicates the greater the ability of the eroded soil[26].•Slope Length and Slope Steepness (LS)