The Utilization of Sentinel-2A and ASTER Imagery for Monitoring the Changes of Public Green Open Space and Oxygen Needs in Sukoharjo Regency in 2004-2019

High land requirements have an impact on land conversion. This study aims to calculate the accuracy of the results of mapping public green open space from ASTER and Sentinel-2A imagery, know the changes in green public space, calculate oxygen demand and the needs of green space in 2004 and 2019. The types of green open spaces that are interpreted visually include urban forests, river borders, cemeteries, fields, and city park. Oxygen demand is calculated by the gerrarkis method including livestock, industry, population, and motor vehicles. The mapping accuracy with the ASTER is 96% while the Sentinel-2A imagery is 90%. The mapping of changes in public green open space show that 17,62 km2 public green open space has not changed, increased 1,15 km2, and decreased 2,61 km2. Oxygen demand in 2004 was 1053531,92 kg/day with green open space needs covering 10,41 km2, while in 2019 it was 1923959,31 kg / day with Green Open Space needs covering 19 km2. The need for green space in 2004 has been fulfilled from public green space of 20,22 km2. In 2019 the area of public open green space is 18,77 km2, so that public open green space has not been able to fulfill the needs of overall green open space.


Introduction
Sukoharjo Regency as one of the pillars of economic growth in Central Java has a rapid economic development. This has an impact on increasing the population significantly. Data from BPS Sukoharjo Regency in 2019 states that the population of Sukoharjo Regency has continued to increase from 2004-2019. In 2018 the population density reached 1,897 people / km 2 with an area of 466.66 km 2 [1]. Economic development has an impact on increasing the need for space and land. High land requirements, not accompanied by adequate land availability. This has resulted in a tendency to change the function of green open space land into a built-up area. Change of land function is a change in land function from one function to another and generally occurs in the suburbs. Data from the Sukoharjo Regency Agriculture Office states that the area of paddy fields decreases every year while built-up land increases by 50-100 hectares per year [11].
The Minister of Public Works Regulation in 2008 states that every city area must provide a Green Open Space of 30% of the total area, namely 20% for Public Green Open Space and 10% for private Green Open Space. The main functions of green open space are as an ecological function, socio-cultural function, economic function, and aesthetic function. Green Open Space also plays a role in providing oxygen for living things and the combustion process [2]. The increase in population, livestock, IOP Conf. Series: Earth and Environmental Science 884 (2021) 012035 IOP Publishing doi: 10.1088/1755-1315/884/1/012035 2 motorized vehicles, and industry has an impact on increasing oxygen demand. Oxygen is very important to sustain ecological life. oxygen plays a fundamental role in preserving the life of living things in the world. With this very important role, the balance between consumption and production needs to be fulfilled [9].The increase in oxygen demand continues to increase while the availability of green open space is decreasing every year.
Previous research regarding the estimation of green open space needs has been carried out by D. Purba, S. Subiyanto, and Hani'ah, 2018 [3]. In this research, the calculation of the area of green open space was carried out with a visual interpretation of the 2015 Quickbird Image and updated using the Sentinel-2A image in 2018. Oxygen demand parameters based on this research used parameters of population, number of motorized vehicles, and number of livestock only. The research only focuses on estimating the need for green open space and its oxygen demand. While in this study the research included monitoring related to the availability of green open spaces and oxygen demand within 15 years. In addition, this study uses the parameters of the number of residents, the number of motorized vehicles, the number of livestock, and the number of industries for the calculation of oxygen demand.
Monitoring of green open space changes is carried out to determine the environmental degradation that has occurred. Remote sensing can be used as a tool in overcoming urban problems. This is because remote sensing has advantages in time, cost efficiency and is environmentally friendly.

Analysis
The two images have different resolutions. The ASTER image on the band appears to have a spatial resolution of 15 m while the Sentinel-2A image has a spatial resolution of 10 m. Resampling needs to be done on the image, to overcome the difference in spatial resolution. Image resampling was done using RStudio software. The resampling method used is bilinear interpolation. The advantage of this method is that the time required for processing tends to be fast with smooth image results [10] Based on this classification, not all types of green open space can be interpreted. Interpretation is only carried out on the type of Public Green Open Space, while Private Green Open Space such as office pages is not interpreted. This is due to the spatial resolution factor of the two images used. The types of green open space that were interpreted were river borders, fields, city parks, funeral, and urban forests. Interpreted appearance is one with a size exceeding the Minimum Mapping Unit (MMU) used. The determination of the MMU size is adjusted from the appearance of the object, where the appearance of the expanding area uses MMU 3mm x 3mm, while in areas that tend to lengthen using a minimum width of 1 mm.
The two interpreted images have different spectral characteristics. The VNIR band on the ASTER image starts with the green band (0.556 μm), red band (0.661 μm), and near NIR / Infrared band (0.807 μm). The difference with the Sentinel-2A image is that the blue band has a wavelength of 0.49 μm, the green band is 0.56 μm, the red band is 0.665 μm, and the NIR band is 0.842 μm. The interpretation was carried out by utilizing a standard false color composite (NIR, red, and green bands), namely

Preprocessing
In the resampling process, each band will experience a change in pixel size. The result of the resampling process is changing the Sentinel-2A image resolution from 10 m to 15 m. Image cutting is done on ENVI software using ROItool (Region of Interest). The study area includes 12 sub-districts in Sukoharjo Regency.

Interpretation
Types of public green open space that can be interpreted are the field, river border, and urban forest, while the funeral and city park only appear to have more than MMU size which are interpreted. The dominant elements of interpretation are form, color, pattern and site.
The type of public green open space in the form of a funeral can be identified by looking at the visible shapes and patterns. The pattern is grouped with features that tend to be unique and distinctive from a collection of tombstones. Cemeteries are social facilities where this object is able to provide green open space reserves, as a water catchment area, as well as a microclimate creator. There are two burial objects that have been successfully interpreted, namely the funeral in Kartasura District, where this funeral has a very large size of 0.13 km 2 .
The field is built with its use that is open, public, and capable of being a water catchment area. The key interpretations of this type of green space are shape, color, and site. Most of the forms of Green Open Space are square or rectangular. The color that appears in both images is clearly visible from the reflection of the soil and the surrounding vegetation. In addition, field interpretation can be related to its location in the middle or near the settlement. This is because the social function of the field is to accommodate community activities.
City parks are in the form of parks open to the public that function as micro-climate creators. This type of green space is difficult to interpret without the help of local knowledge from the manufacturer. This is because most of the city parks are located in the middle of dense settlements with a park area that is not too large. City parks that can be interpreted are Satya Negara Square Sukoharjo Regency and Ananda City Park. The two city parks can be interpreted due to their size which is more than the specified MMU.
River border in the form of green lines on either side of the river to protect the river and maintain its sustainability. This green space interpretation uses key interpretations in the form of patterns, colors, and sites. This green space has an elongated, linear pattern along the river. In addition, their appearance is easily recognizable by the red, clustered reflection of the vegetation. This appearance can also be related to its location on the right and left of the river. The interpretation carried out in Sukoharjo Regency shows that the form of urban forest in this area tends to be clustered and spread. The clustered form of urban forest can be found in production forest in Polokarto District, while urban forest that spreads out is found in protected forest in Bulu District. The process of interpreting urban forest is a little difficult to interpret, so assistance is needed in the form of a map of the spatial pattern of Sukoharjo Regency 2011-2031 [7]. The map of the spatial pattern of Sukoharjo Regency 2011-2031 is shown in Figure 1  The urban forest spreads out in a protected area. The morphology tends to be unique and has certain ecological functions with an area of up to 3 km 2 . In addition to being seen from the spatial pattern map of Sukoharjo Regency in 2011-2031, urban forest is clustered in shape, it can also be identified from the appearance of vegetation with clustering patterns and tree shapes that tend to be uniform. This type of urban forest that can be interpreted is rubber and teak production forest.

Determination of samples and field checks
The total sample used was 50 samples that were evenly distributed in each study area. The results of the field check in 2004 showed two wrong interpretation results, namely the type of green open space for city parks and the field, while in the field survey in 2019 it was found that there were five wrong interpretation results, namely on the river and field boundaries. Field green open space has been mistaken because this object is similar to unplanted agricultural land or dry agricultural land. In addition, this green open space is also similar to a garden owned by residents. Borders of rivers and city parks experience errors in interpretation because of their proximity to each other so that they are difficult to distinguish from one another. Urban forest samples located in mountainous areas are difficult to reach. This results in a shift in the sample point to a place that is more easily accessible, without changing the type of green space to be checked for conditions in the field.

Test accuracy
The accuracy test is carried out by comparing the interpretation of ASTER images with google earth. The overall accuracy results in 2004 can be seen in Table 2. Mapping public green open space using the ASTER image, the accuracy is 96% with a kappa coefficient value of 0.94. Based on the accuracy of the manufacturer and the accuracy of the user, all types of public green open space can be identified properly.
The accuracy test of public green space mapping in 2019 was carried out by comparing the interpretation of Sentinel-2A images with field surveys. The results of the creator's accuracy and user accuracy are more varied, as shown in Table 3. The overall mapping using the Sentinel-2A image produces an overall accuracy of 90% with a kappa coefficient value of 0.86. Based on the accuracy of the maker and the user, there are four types of public green open space that can be identified properly.
The accuracy produced by the two images exceeds 85%, which according to Sutanto (1986) states that if the accuracy rate exceeds 85% then it is feasible for further analysis [8]. The kappa value produced by the two images ranges from 0.8 to 1, where according to Altman (1991) this value is in the very strong or very good category [5]. Based on this, the mapping of public green open space with both images can be further analyzed because the higher the overall accuracy value and the kappa coefficient value, the better the mapping results can be.      Table 4.  The reduction in the area of river borders has occurred in many river border areas adjacent to residential areas. This is due to the increasing population and increasingly varied community activities, so that more space is needed to live or do activities. Vegetation on the border of the river also turns into a dwelling for the community.
Monitoring In addition, kecamatan with flat relief tend to be used for industrial estates.
The western part of Sukoharjo Regency which includes Polokarto, Bendosari and Nguter Districts is bordered by Karanganyar Regency, so the relief tends to be bumpy to hilly. These three districts are still dominated by the use of agricultural land in the form of terraced rice fields and gardens owned by residents, and in some areas used for rubber plantations. This is different from the situation in Karanganyar Regency, which is dominated by standing trees.  Table 5. The most additional public green open spaces were added in Grogol District, but the highest reduction in public green open space was also found in Grogol District. This addition occurs because there is a development of urbanparks, while the reduction in the area of public green space occurs due to residential areas that are increasingly densely populated and also the development of industrial areas.  The total oxygen demand for livestock is not limited to the boundaries of each district, but is analyzed at the district level. This is because there is no specific data in each district. Oxygen requirements for large livestock have greater oxygen requirements than small livestock and poultry. According to Muis (2005) the oxygen demand for cows and buffaloes reaches 1.70 kg / day while 2.86 kg / day for horses. Small livestock types such as goats and sheep require oxygen of 0.314 kg / day while poultry is 0.17 kg / day [3]. The oxygen demand for pigs is based on the research of Hannon et al. (1989) which is equal to 1.24 kg / day [4]. The oxygen needs of each type of livestock and their numbers in 2004 and 2019 are detailed in Table 6. The increase in livestock numbers also has an impact on increasing the amount of oxygen demanded. This can be seen in the oxygen demand in 2019 which reached 1,014,770.48 kg / day while in 2004 it was only 279,325.70 kg / day. The calculation of oxygen demand based on the number of industries is not limited to the boundaries of each district, but is analyzed based on district administrative boundaries. This is because there is no specific data in each district. The division of industrial types is classified into the agro and forest products industry, the textile and various industries, and the electro-machine metal chemical industry. The calculation of oxygen demand based on industry uses a large number of industries. This is based on the assumption that large industries use a large number of machines and for a longer period of time than small and medium industries. In 2004, the total number of large industries was 25 industries, while in 2019 it grew to 129 industries. This development mostly occurred in the electro-machine metal chemical industry.
The use of large industries for the calculation of oxygen demand is based on research by Ryadi (1984). The research states that the average fuel requirement for the engine is 185.76 kg / day, while each kg of diesel motor fuel requires 2.86 kg of oxygen [6]. Based on this formula, the oxygen demand in 2004 was 13,281.77 kg / day and in 2019 was 68,533.92 kg / day.

d. Oxygen Needs for Motor Vehicle
The number of motorized vehicles in 2004 was 169,968 units and in 2019 801,581 units. Over the past 15 years, there has been an increase in motor vehicle consumption by 631,613 units of vehicles. The number of motorized vehicles is obtained from BPS data from Sukoharjo Regency which is derived from vehicle tax revenue data collection in Sukoharjo Regency. The assumption used in calculating oxygen demand is the number of vehicles originating from Sukoharjo Regency, while vehicles originating from outside the area are ignored. The total oxygen demand is not limited to the boundaries of each sub-district, but is calculated at the district level.
The largest number of motorized vehicles is dominated by motorbikes. Bus cars have the smallest number and the number continues to decrease every year. This is followed by an increase in the number of passenger cars and motorbikes, indicating that the use of private vehicles is increasing every year. The fuel and oxygen requirements for each type of vehicle are detailed in Table 7. The total oxygen demand of each type of motorized vehicle in each year can be seen in Table 8. The calculation results show that there is an increase in oxygen demand from motorized vehicles, namely in 2004 of 56,687.56 kg / day while in 2019 it was 70,042.94 kg / day.

Need for green open space
The need for green open space is calculated using the Gerrarchic approach. This method shows that green open space in every m 2 will produce 54 grams of dry weight, wherein one gram of dry weight is equivalent to oxygen production of 0.9375 grams / day [6].

Conclusion
The conclusions obtained from this study are ASTER imagery and Sentinel In this research, the interpretable type of green open space can only include public green open space. This is adjusted to the type of image used. The suggestion for further research is the use of remote sensing images with better spatial resolution and the same characteristics, so that it can simplify the process of interpreting types of public green open space in more detail and with more accurate results. In addition, data on the number of motorized vehicles in and out of Sukoharjo Regency can also be added so that the calculation of oxygen demand can be more accurate. Monitoring changes in green open space is very important to do. Future research is not only limited to monitoring changes that have occurred, but also estimating the need for green open space in relation to oxygen demand and the city's rapid development.