Habitat suitability modelling of Pasak Bumi (Eurycoma longifolia Jack.) in Riam Kanan conservation forest zone using Sentinel-2 biophysical parameters

Pasak bumi usually grow under the tree canopies. So that the character of its habitat is assumed to be estimated using several biophysical parameters of the tree canopy around it. The purpose of this research was to model the suitability of the pasak bumi habitat in the Riam Kanan conservation forest zone, using a number of biophysical parameters extracted from Sentinel-2 MSI imagery. Those parameters are Leaf Area Index (LAI), Canopy Chlorophyll Content (CCC), Canopy Water Content (CWC), Fraction of Vegetation Cover (FVC), and Fraction of Absorbed Photosynthetically Active Radiation (FAPAR). Ground surveys were carried out to find the coordinates of pasak bumi using accidental sampling method. Pasak bumi coordinate points are overlaid with biophysical parameters. Statistical analysis was then applied to predict the range of population values from each biophysical parameter, using Confidence Interval (CI) 95%. The results of the research show that CI LAI 2.532-2.772, CI CCC 137.101-158.028 gr/cm2, CI CWC 0.05-0.057 gr/m2, CI FVC 0.698-0.737, and CI FAPAR 0.732-0.765. The values of these biophysical parameters directly describe the biophysical characteristics of the pasak bumi habitat in the research location. These CI values are then implemented using binary modelling to predict the habitat of pasak bumi. Based on the results of modelling, it was found that the area suitable for pasak bumi plants was an area of 1,807.91 hectares. This area has a proportion of 1.55% of the total area of the conservation zone. To improve accuracy, other biophysical parameters can be considered to be involved in modelling.


Introduction
Pasak bumi (Eurycoma longifolia Jack.) plant or commonly known as tongkat ali, is an endemic plant that is commonly found in dense forests in Indonesia and in Malaysia. For a long time, this plant was known by the traditional society for its health benefits [1]. Especially used as traditional medicine for male vitality, and other medicines. According to [2], pasak bumi is commonly found in primary forests. According to [3], pasak bumi plants are spread in various elevation, but are dominant at an elevation of 300 meters or more above sea level. While according to [4], the population of pasak bumi plants can reach 130 individuals per hectare.
The condition of the pasak bumi habitat is found in areas that have elevation of 320 -402 meters above sea level, found sporadically and in a group pattern, daily average temperature 25.6 o C, average daily relative humidity 73.6%, light intensity 0.9 klx , and red-yellow podsolic soils with textures ranging from clay to sandy clay. [5] and [6] states that pasak bumi are found at elevation of 170 -200 meters above sea level in secondary forests, average daily temperatures of 28 -41 o C, average daily relative humidity of 53-59%, and intensity light 2.10-3.91 klx.
Pasak bumi is a non-timber forest product with conservation status "not determined" and is commercially traded on the island of Borneo. At present, the population of pasak bumi plants in nature, such as on the island of Borneo, Indonesia, is indeed still very abundant. So even though large numbers of people have taken it, it is estimated that this plant will not run out for the next few years. However, the problem here is the land where it grows. Where the land or forest where the pasak bumi grows continues to be degraded. The area is continuously decreasing due to land conversion for residential or agricultural purposes. So that threatens the sustainability of the pasak bumi plant is not because it is harvested, but because of the disturbed habitat for its growth. So we really need to predict or model the suitability of the habitat where these earth pegs grow. Good for future development needs, if the population in nature has been reduced due to damaged habitat, or if there is a pharmaceutical industry that wants to make this pasak bumi as raw material for medicines.

Figure 1. Pasak bumi plant
Pasak bumi is a dwarf that usually grows under a tree canopy. Because the pasak bumi grows under the tree canopies, theoretically this plant is difficult to observe directly using remote sensing imagery. Even by using very high spatial resolution imagery such as the image from an Unmanned Aerial Vehicle (UAV), pasak bumi still cannot be directly observed visually. So that the only way to identify the location and mapping the spatial distribution of the pasak bumi accurately is by direct ground survey methods. By studying the data on the growth of pasak bumi obtained from the results of ground surveys, we can model the suitability of the pasak bumi plant habitat in the area we want.
Habitat where plants grow can be predicted using various parameters. Such as biophysical, topographic, edaphic, climatic, and so on. These parameters can be assessed directly based on ground data samples. Furthermore, these parameters can be involved in spatial modelling methods to extrapolate habitat suitability of certain plant species in the desired area. As mentioned earlier, that the pasak bumi grows under the canopy of other plants around it. So that the growth of the pasak bumi can be assumed to have a spatial correlation with the surrounding plants, especially the tree canopy parameters that shelter it.
Biophysical parameters such as tree canopy density, canopy moisture, chlorophyll content, plant health conditions, etc., can be extracted directly from remote sensing images. Biophysical parameters like this can be used directly to predict habitat suitability of a plant species, as do pasak bumi. Since its presence in 2015, the European Space Agency (ESA) Citra Sentinel-2 MSI (Multispectral Instrument) has attracted worldwide attention. Because in addition to its highly qualified spatial resolution, which is 10 meters, and its temporal resolution capability can record the same place every five days, this image can also be used by the public worldwide for free. of course, this multispectral image such as Sentinel-2 can be used as a base for extracting biophysical parameters, such as canopy density and so on, for modelling the suitability of pasak bumi plant habitat.
The purpose of this research was to model the suitability of the pasak bumi plants in the Riam Kanan conservation forest area, using a number of biophysical parameters extracted from Citra Sentinel-2 MSI. The biophysical parameters in question are Leaf Area Index (LAI), Canopy Chlorophyll Content (CCC), Canopy Water Content (CWC), Fraction of Vegetation Cover (FVC), and Fraction of Absorbed Photosynthetically Active Radiation (FAPAR). A ground survey was conducted in October 2018 to find the location of the coordinates of the discovery of pasak bumi plants in the ground. The sampling method used is accidental sampling. The sampling area is only limited to the KHDTK Forest of Education and Training of Lambung Mangkurat University. Based on the results of the ground survey, 32 locations of pasak bumi plants were found. As can be seen in Table 1.

Sentinel-2 MSI Biophysical Parameters
Sentinel-2 MSI Biophysical Parameters are parameters that can be extracted directly from Sentinel-2 imagery, namely Leaf Area Index (LAI), Canopy Chlorophyll Content (CCC), Canopy Water Content (CWC), Fraction of Vegetation Cover (FVC), and the Fraction of Absorbed Photosynthetically Active Radiation (FAPAR). The extraction process is done automatically using the ESA SNAP Biophysical Processor, which is also integrated in the SNAP ESA software. ESA SNAP extracts Sentinel-2 MSI Biophysical Parameters using a machine learning algorithm, namely artificial neural network [9]. Sentinel-2 MSI Biophysical Parameters have been validated by several researchers, including the newest one by [10]. Of course, these five parameters represent various biophysical characters that will give the descriptions of where the pasak bumi grows. LAI is defined as half the developed area of photosynthetically active elements of the vegetation per unit horizontal ground area. It determines the size of the interface for exchange of energy (including radiation) and mass between the canopy and the atmosphere [9]. FAPAR corresponds to the fraction of photosynthetically active radiation absorbed by the canopy. The FAPAR value results directly from the radiative transfer model in the canopy which is computed instantaneously. It depends on canopy structure, vegetation element optical properties and illumination conditions [9]. FVC is used to separate vegetation and soil in energy balance processes, including temperature and evapotranspiration [9]. The chlorophyll content is a very good indicator of stresses including nitrogen deficiencies [9]. It is strongly related to leaf nitrogen content [11]. CCC represents vegetation health parameters. While CWC represents climate parameters, both temperature, rainfall, intensity of irradiation, and air humidity.

Statistical Analysis and Spatial Modelling
All the coordinates of the the pasak bumi plant are overlaid with the fifth Sentinel-2 MSI Biophysical Parameters. This aims to obtain quantitative biophysical data on the growth of the pasak bumi. Quantitative data from the overlapping of the pasak bumi and Sentinel-2 MSI Biophysical Parameters locations can be seen in Table 1. From this data, a number of statistical analyzes were conducted to look for certain values, namely mean, variance, standard deviation, and confidence interval. Confidence interval (CI) is calculated using a 95% confidence interval. Where CI itself is formulated as follows: Where: ̅ : mean z: the z value at confidence interval 95%, i.e. 1.96 S: standard deviation n: number of samples The minimum and maximum values of CI resulting from the calculation of equation (1) above are assumed to be the quantitative limiting factors for the growth of the pasak bumi for each Sentinel-2 MSI Biophysical Parameter. Furthermore, CI values for each Sentinel-2 MSI Biophysical This parameter will be included as a parameter in spatial modelling, where the spatial modelling method used is binary modelling. Binary modelling is used because the output of the modelling has only two categories, which are suitable for habitat for pasak bumi, and not suitable for habitat for pasak bumi.

Results and Discussion
LAI parameters represent leaf or canopy density, FVC represents the presence of vegetation, CCC and FAPAR represent the health of vegetation, while CWC and also FVC can be expressed as representing climate parameters, especially temperature, rainfall, intensity of irradiation, and air humidity. These five parameters represent the character where the pasak bumi grow. For as stated earlier, that the pasak bumi The Fifth International Conferences of Indonesian Society for Remote Sensing IOP Conf. Series: Earth and Environmental Science 500 (2020) 012020 IOP Publishing doi:10.1088/1755-1315/500/1/012020 6 grow under the canopy of the surrounding vegetation. So that the existence of the pasak bumi plant will have an association with the surrounding plants. Of course, in this study, the effects of other parameters that also determine the character of pasak bumi habitat, such as topography and so on, are ignored. This is a limitation in this research.  The extraction results of LAI, CCC, CWC, FVC, and FAPAR use the SNAP Biophysical Processor ESA on the Sentinel-2B imagery can be seen in Figure 4. Of course, the output of the SNAP ESA Biophysical Processor is raster as the data source, that is the Sentinel-2B imagery. The coordinates of each location found in the ground of pasak bumi are overlaid with each of these rasters. So that the pixel value of each parameter is obtained for each coordinate of earth pasak, as shown in Table 1.
The extraction of pasak bumi habitat suitability cannot be done on the entire research area, because some areas, especially on mountain peaks, are covered with dense clouds and cloud shadows in the imagery used. So that there is a possibility that the area of habitat suitability results of spatial modelling in this study is under estimate. Habitats that are suitable for pasak bumi in the ground may actually be wider than predictive results in this research.