Analysis of changes in vegetation density with sentinel-2a image on Malino sub-watershed

Changes in vegetation density, especially in sub-watershed areas, will continue to occur over time. This is due to increasing human activities impacting land use change. The Malino sub-watershed one part of the Jeneberang watershed, almost all located in Tinggimoncong District. Monitoring changes in vegetation density in the Malino Sub-watershed is carried out by utilizing remote sensing technology in the form of sentinel-2A imagery and using the NDVI type of vegetation index method. This study aims to determine the distribution and changes in vegetation density in the Malino Sub-watershed in 2017, 2019, and 2021. From the analysis results conducted using the NDVI vegetation index, the distribution of vegetation density in the Malino Sub-watershed is divided into four classes, namely the non- vegetation, sparse vegetation, medium vegetation, and dense vegetation. Changes in vegetation density in the Malino sub-watershed from 2017 to 2019 occurred in the dense vegetation density class of 2.21% or 192.69 ha, while the smallest occurred in the non-vegetation density class of 0.24% or 20. 97 ha. Meanwhile, the most significant change in vegetation density in the Malino sub-watershed from 2019 to 2021 occurred in the dense vegetation density class of 11.64% or 1,016.60 ha, while the smallest occurred in the non-vegetation density class of 0.12% or 10.46 ha.


Introduction
Increased human growth and activity will continue over time.This increase will undoubtedly affect changes in an area, especially regarding land use.One tangible example regarding changes in land use is converting green land into residential or business land.As a result of this change, of course, it will also impact changing the vegetation density in the region.
The existence of vegetation in an area, especially in the sub-watershed area, has a very.Important role.One of the roles of vegetation is to reduce the risk of flooding during the rainy season.This is because the presence of vegetation will increase water absorption during the rainy season.And will then form groundwater flows.The greater the groundwater flow that is created, of course, This will reduce the occurrence of surface runoff, which can result in an increase in water discharge in The river [1].
The Malino sub-watershed an area of the Jeneberang watershed.The Malino sub-watershed is in the upper reaches of the Jeneberang watershed, and almost all of its territory is in Tinggimoncong District.Tinggimoncong is one of the sub-districts in Gowa Regency and is known as one of the sub-districts with many tourist attractions such as Parang Bugisi Waterfall, Takapala Waterfall, Malino Pine Forest Tourism, and many more.In addition, based on data from [2], the number of residents in Tinggimoncong District for 2010-2020 has increased from 22,157 people in 2010 to 23,332 people in 2020.Seeing this, 1230 (2023) 012140 IOP Publishing doi:10.1088/1755-1315/1230/1/012140 2 it is certainly not impossible that Tinggimoncong District is experiencing development in terms of development.As a result of these developments, it will undoubtedly impact changes in vegetation density in Tinggimoncong District, especially in the Malino Sub-watershed area.
Monitoring of changes in vegetation density can be carried out efficiently and cover large areas through remote sensing technology.Remote sensing is a science used to obtain information about an object (the earth's surface).With remote sensing, information about an object can be obtained without requiring direct contact with the thing to be studied.One of the remote sensing technologies that can be used to monitor changes in vegetation density is satellite [3].
Sentinel-2A is a type of high-resolution satellite that can be used to monitor changes in vegetation density.Europe's Copernicus program developed Sentinel-2A and aims to conduct earth observations, specifically for environmental monitoring [4].In addition to utilizing a sentinel-2A satellite image, the Normalized Difference Vegetation Index (NDVI) type vegetation index method is also needed to determine the high and low density of vegetation in the Malino Sub-watershed.NDVI utilizes the red and infrared bands in satellite images to determine the density level in vegetation.
Based on this description, research on the analysis of vegetation density with sentinel-2A image in the Malino Sub-watershed needs to be carried out so that it can provide information related to changes in vegetation density in the Malino Sub-watershed area.

The purpose of the research
The purpose of this research is as follows, 1. Generate distribution maps of vegetation density in the Malino sub-watershed in 2017, 2019, and 2021.2. Generate maps of changes in vegetation density in the Malino Sub-watershed from 2017 to 2019 and from 2019 to 2021.This research aims to provide information related to the distribution and changes in vegetation density in the Malino sub-watershed so that it can become input for the local government in supporting land use planning in the area.

Time and place
This research was conducted from January to March 2022.The research location was in the Malino subwatershed area, Gowa Regency, South Sulawesi Province (Figure 1).Research data was processed and analyzed at the Soil and Water Engineering Laboratory, Study Program of Agricultural Engineering, Faculty of Agriculture, Hasanuddin University, Makassar.

Tools and materials
The tools used in this study consisted of data collection and data analysis tools.The data collection tool is a mobile phone with Google Maps and Open Camera applications.At the same time, the data analysis tool is a laptop that includes QGIS and Microsoft Exel software.
The materials used in this study were sentinel-2A satellite image data for the Malino sub-watershed in 2017, 2019, and 2021, the data for the Malino sub-watershed, the data for village administrative maps of Gowa Regency, and field data (field photos and coordinate points).

Research procedure
The procedure in this study is as follows, 3.3.1.Preparation.At this stage, the selection of research locations and literature studies was carried out.The selected research location is the Malino Sub-watershed.This is because almost the entire area of the Malino Sub-watershed is located in the Tinggimoncong District, which is one of the many tourists visiting.At the same time, the literature study is carried out by looking for references in the form of journals, books, and others related to this research.

Data collection.
The data needed in this study consists of 2 types: primary and secondary.The preliminary data is in the form of field photos and coordinate points taken in the research area using the open camera application.While the secondary data is in the form of a sentinel-2A image which can be downloaded at https://earthexplorer.usgs.gov/, the data for the Malino sub-watershed to limit research locations and the data for village administration maps of Gowa Regency.The sentinel-2A images taken in this study are those for 2017, 2019, and 2021.The choice of the year is intended so that changes in vegetation density in the Malino Sub-watershed can be seen more clearly.

Data processing.
Data processing uses QGIS software, which begins with atmospheric correction on sentinel-2A images.This correction aims to increase the accuracy of the image data by eliminating atmospheric disturbances.The corrected image data will then be cropped using the data from the Malino sub-watershed.This cropping aims to obtain an image more specific to the study area, as for the sentinel-2A image data that is cropped, namely the red band image and the NIR band, which has a spatial resolution of 10 meters.
The cropped image data (red and NIR bands) will then be analyzed using the NDVI-type vegetation index.Analysis was conducted to obtain NDVI values (-1 to +1) in the study area to classify vegetation density.The unsupervised classification method is used to classify vegetation density in the Malino subwatershed.The classification of vegetation density in the Malino sub-watershed is based on Table 1.[5,6].  1.The field data obtained will then be compared with the results of image classification using the confusion matrix technique.After comparison using the confusion matrix technique, validation will then be carried out using the Kappa coefficient to test the accuracy of the vegetation density classification that has been made.

Map Layouts.
The processed vegetation density map in the Malino sub-watershed will then be layout.In addition to the map that has been made, the layout will also display other attributes as map descriptions such as map title, map legend, map scale, and others.

Results and discussion
4.1.Validation of using sentinel-2A image to identify vegetation density in the Malino sub-watershed.
In validation, the sentinel-2A image is a 2021 image, and the field photos used are field photos in 2022.
For this reason, the conditions at the research sites in 2021 and 2022 are assumed to be the same in this study.Validation begins by using the confusion matrix technique by comparing image classification results with objects in the field (refer to Table 1).After the comparison using the confusion matrix technique, the validation is continued using the Kappa coefficient.The number of samples used was 42 samples selected randomly.Based on Table 2, it can be seen that there are still errors in the classification results made.This can be seen in the User's Accuracy (UA) value, which reaches 100% only in the sparse vegetation density class.For Producer's Accuracy (PA), only the dense vegetation density class achieves a value of 100%.Based on the results of validation calculations using the Kappa coefficient, an accuracy of 72.46% is obtained.According to Landis and Koch (1977) in [7], the Kappa coefficient value between 40% to 80% (0.4 -0.8) is included in the medium category.This indicates that the density classification results carried out have less accuracy.This lack of accuracy can be caused by several factors, one of which is because the classification was carried out using the unsupervised classification method, which resulted in several misclassifications in each density class.

Dense Vegetation
Based on the comparison table (Table 3), it can be seen that if the resulting color in the NDVI image is darker (black), then the NDVI value will be closer to the value -1 and indicates that vegetation in the study area is getting rarer or even has no vegetation.Meanwhile, if the color produced in the NDVI image is brighter (white), then the NDVI value will be closer to +1, indicating that the vegetation in the study area is getting more and more dense.This aligns with the opinion of [8], stating that the NDVI value will be directly proportional to vegetation density (has a unidirectional relationship).

Distribution of vegetation density in the Malino sub-watershed 4.2.1. Distribution of vegetation density in the Malino sub-watershed in 2017.
The distribution of vegetation density and the distribution map of vegetation density in 2017 in the Malino sub-watershed was obtained from the processing of sentinel-2A image in 2017 using the NDVI type vegetation index method (refer to Table 1).The distribution of vegetation density is classified into four density classes: non-vegetation, sparse vegetation, medium vegetation, and dense vegetation.The distribution of vegetation density and a map of the distribution of vegetation density in the Malino Sub-watershed in 2017 can be seen in Table 4 and Figure 2.

No
Vegetation Density Area (ha) % Area Based on Table 4, it can be seen that in 2017 the most extensive density distribution was in the dense vegetation density class of 77.75% of the total area of the Malino Sub-watershed.Meanwhile, the minor density distribution was in the non-vegetation density class of 0.81% of the total area of the Malino Subwatershed.We can also see this on the distribution map of vegetation density (Figure 2), where the dominant color on the map is green, which is classified as an area with dense vegetation density.

Distribution of vegetation density in the Malino sub-watershed in 2019.
The distribution of vegetation density and the distribution map of vegetation density in 2019 in the Malino Sub-watershed were obtained from the results of sentinel-2A image processing in 2019 using the NDVI type vegetation index method (refer to Table 1).The distribution of vegetation density is classified into four density classes: non-vegetation, sparse vegetation, medium vegetation, and dense vegetation.The distribution of vegetation density and a map of the distribution of vegetation density in the Malino Sub-watershed in 2019 can be seen in Table 5 and Figure 3.  Based on Table 5, it can be seen that in 2019 the greatest density distribution was in the dense vegetation density class of 75.55% of the total area of the Malino Sub-watershed.Meanwhile, the smallest density distribution is in the non-vegetation density class of 0.57% of the total area of the Malino Sub-watershed.We can also see this on the distribution map of vegetation density (Figure 3), where the dominant color on the map is green which is classified as an area with dense vegetation density.1).The distribution of vegetation density is classified into four density classes: non-vegetation, sparse vegetation, medium vegetation, and dense vegetation, the distribution of vegetation density and a map of the distribution of vegetation density in the Malino Subwatershed in 2021 can be seen in Table 6 and Figure 4.  Based on Table 6, it can be seen that in 2021 the greatest density distribution will be in the dense vegetation density class of 87.19% of the total area of the Malino Sub-watershed.Meanwhile, the smallest density distribution was in the non-vegetation density class of 0.69% of the total area of the Malino Sub-watershed.We can also see this on the distribution map of vegetation density (Figure 4), where the dominant color on the map is green, which is classified as an area with dense vegetation density.

Distribution of vegetation density in the
An explanation regarding the distribution of vegetation density in the Malino Sub-watershed in 2017, 2019, and 2021 shows that in these three years, the density distribution that dominates is the dense vegetation density class.The dense vegetation density is inseparable from the number of vegetated areas in the region.Based on the results of the ground check conducted at the research location, the location still has a lot of forests and agricultural land.In addition, based on the explanation regarding the distribution of vegetation density in these three years, it can be seen that the NDVI-type vegetation index method can be used to determine the distribution of vegetation density in an area.This is in line with the opinion of [9].Various vegetation indices, particularly the NDVI, have proven helpful in generating information about vegetation density on vegetated lands.a decrease in the area include the non-vegetation density class and the dense vegetation density class.In comparison, the two density classes that have experienced an increase include the sparse and medium vegetation density classes.Based on Table 7, it can also be seen that the largest change in area from 2017 to 2019 was in the dense vegetation density class of 2.21%.Meanwhile, the smallest change in area was in the non-vegetation density class of 0.24%.9.The area of change in the table is obtained from the difference between the distribution area of vegetation density in 2021 and the area of distribution of vegetation density in 2019.This difference shows that two density classes have decreased and two categories have increased area.The two density classes that have experienced a decrease in the area include the sparse vegetation density class and the medium vegetation density class.In comparison, the two density classes that have experienced an increase include the nonvegetation and dense vegetation density classes.Based on Table 9, it can also be seen that the largest change in area from 2019 to 2021 is in the dense vegetation density class of 11.64%.At the same time, the smallest area change was in the non-vegetation density class of 0.12%.6, it can be seen that there are 16 (sixteen) areas of vegetation density from 2019 to 2021, where four conditions are still the same (no change) from the previous year.The four conditions include non-vegetation to non-vegetation, sparse vegetation to sparse vegetation, moderate vegetation to moderate vegetation, and dense vegetation to dense vegetation.Meanwhile, apart from these four conditions, (12 other states) underwent a change to a different density class from the previous year.The total area of the 16 (sixteen) conditions can be seen in Table 10.Table 11 is a combination of changes in vegetation density that occurred from 2017 to 2021.Where the table shows a decrease and increase for each vegetation density class, the decrease and growth in the area of each density class can be caused by land conversions such as the conversion of forest land into plantations, paddy fields, or even settlements.This is based on the increasing number and human need for land over time.According to [8], one of the things that can affect changes in vegetation density is the increase in population which causes changes in land use.

Conclusion
The distribution of vegetation density in the Malino Sub-watershed in 2017, 2019, and 2021 is dominated by dense vegetation density classes.The largest change in vegetation density in the Malino Sub-watershed from 2017 to 2019 and from 2019 to 2021 occurred in the dense vegetation density class, while the smallest occurred in the non-vegetation density class.Where for 2017 to 2019 these two changes have decreased from the previous year while for 2019 to 2021 these two changes have increased from the previous year.

Figure 1 .
Figure 1.Map of research location.

Table 4 .
Distribution of vegetation density in the malino sub-watershed in 2017.

Figure 2 .
Figure 2. Map of vegetation density distribution in Malino sub-watershed in 2017.

Table 5 .
Distribution of vegetation density in the Malino sub-watershed in 2019.

Figure 3 .
Figure 3. Map of vegetation density distribution in Malino sub-watershed in 2019.

Table 6 .
Distribution of vegetation density in the Malino sub-watershed in 2021.

Figure 4 .
Figure 4. Map of vegetation density distribution in Malino sub-watershed in 2021.

Figure 5 .
Figure 5. Map of changes in vegetation density in the Malino sub-watershed from 2017 to 2019.

Figure 5
Figure 5 is obtained by overlaying the distribution maps of vegetation density in 2017 and 2019.From Figure5, it can be seen that there were 16 (sixteen) areas of vegetation density from 2017 to 2019, where there are four conditions that are still the same (no change) from the previous year.The four conditions include non-vegetation to non-vegetation, sparse vegetation to sparse vegetation, moderate vegetation to moderate vegetation, and dense vegetation to dense vegetation.Meanwhile, apart from these four conditions, (12 other states) underwent a change to a different density class from the previous year.The total area of the 16 (sixteen) conditions can be seen in Table8.
Figure 5 is obtained by overlaying the distribution maps of vegetation density in 2017 and 2019.From Figure5, it can be seen that there were 16 (sixteen) areas of vegetation density from 2017 to 2019, where there are four conditions that are still the same (no change) from the previous year.The four conditions include non-vegetation to non-vegetation, sparse vegetation to sparse vegetation, moderate vegetation to moderate vegetation, and dense vegetation to dense vegetation.Meanwhile, apart from these four conditions, (12 other states) underwent a change to a different density class from the previous year.The total area of the 16 (sixteen) conditions can be seen in Table8.

4. 3 . 2 .
Changes in vegetation density in the Malino sub-watershed from 2019 to 2021.Changes in vegetation density in the Malino Sub-watershed from 2019 to 2021 can be seen in Table The data used at this stage is ground check data in the research area (field data).The results of the ground check itself are in the form of field photos showing the objects in the research area and the coordinates of the field.The division of things based on the range of NDVI values can be seen in Table

Table 2 .
Comparison between image calcification results and field photos (objects).

Table 3 .
Comparison of visualization between NDVI image and vegetation density maps and field photos.

734.33 100
Malino sub-watershed in 2021.The distribution of vegetation density and a map of the distribution of vegetation density in 2021 in the Malino Subwatershed were obtained from the processing of sentinel-2A image in 2021 using the NDVI type Malino sub-watershed 4.3.1.Changes in vegetation density in the Malino Sub-watershed from 2017 to 2019.Changes in vegetation density in the Malino Sub-watershed from 2017 to 2019 can be seen in Table7.The area of change in the table is obtained from the difference between the distribution area of vegetation density in 2019 and the area of distribution of vegetation density in 2017.This difference shows that two density classes have decreased and two types have increased area.The two density classes that have experienced

Table 7 .
Changes in the area of vegetation density in the Malino sub-watershed from 2017 to 2019.

Table 8 .
Conditions of vegetation density area in the Malino sub-watershed from 2017 to 2019.

Table 9 .
Changes in the area of vegetation density in the Malino sub-watershed from 2019 to 2021.

Table 10 .
Conditions of vegetation density area in the Malino sub-watershed from 2019 to 2021.

Table 11 .
Malino sub-watershed from 2017 to 2019 and from 2019 to 2021 Changes in vegetation density in the Malino sub-watershed from 2017 to 2019 and 2019 to 2021.