Probability of Erosion Utilizing Google Earth Engine and the RUSLE Method in the Tuntang Watershed

The Tuntang Watershed is an important watershed in Central Java. Management of watersheds in the Tuntang stream is a priority for various parties to carry out. One of the things that threatens the sustainability of the Tuntang watershed is erosion. The erosion rate can lead to sediment accumulation and siltation in the Tuntang River reservoir, which can cause catastrophic flooding. Flood disaster mitigation caused by erosion needs to be done, one of which is by calculating the erosion rate per year that occurs in the Tuntang watershed. This study calcultated the predicted erosion rate (per year in the Tuntang watershed) using the Revised Universal Soil Loss Equation (RUSLE) method, processed using the Google Earth Engine (GEE). Google offers a cloud-storage technology called GEE. Programming in JavaScript is required to operate GEE. GEE is a petabyte-scale data-based tool that can be used to analyze and archive geospatial data that is open source. The computing environment is designed for the processing of geospatial data, including the depiction of spatial analysis of satellite imagery. Data for RUSLE is obtained from the database in GEE, and the results can be imaged on a map. According to the study’s findings, the degree of soil erosion throughout the Tuntang Watershed was essentially constant, with Moderate erosion predominating in the majority of locations. Senjoyo Sub Watershed, Rowopening Sub Watershed, and Tuntang Hilir Sub Watershed are the primary locations with severe erosion. Rowopening Sub Watershed is the region that is the worst.


Introduction
The Tuntang Watershed is an important watershed in Central Java.Watershed management in the Tuntang stream is a priority for various parties.One of the things that threatens the sustainability of Tuntang river basin is erosion.The percentage of erodibility can result in sediment buildup and siltation of the Tuntang River reservoir, which can cause flooding.Even though erosion of soil is one of several natural processes that occur in environment, it can have negative effects on drainage systems, farming, agroforestry, land degradation, and the removal of non-rechargeable land resources [1].In this case, preventing soil degradation and sustaining top-grade soil depend heavily on estimates of soil erosion rates as well as dispersion processes in time and space.Because a number of factors, including geography, climate, vegetation cover on the land, and soil, affect soil erosion, estimating it can be difficult [2].Many methods can be used to determine the rate of erosion, and all of them provide insight into the area experiencing erosion.A method that can be used to measure the level of erosion vulnerability is the RUSLE (Revised Universal Soil Loss Equation) method.In a number of environments, this formula has frequently been used to predict soil erosion [3].RUSLE uses multispectral satellite imagery and satellitederived terrain elevation models, along with rainfall and ground data [4].In contrast to previous soil loss methods, the RUSLE method has the benefits of a straightforward formula, few parameters, and accurate results with high estimates.This formula has developed into a widely accepted quantitative estimation model for soil erosion.RUSLE is the most common approach for calculating erosion rate, according to data from the Global Applications of Soil Erosion Modelling Tracker (GASEMT) [5].
In this study, the rate of annual erosion that occurred in the Tuntang watershed will be predicted using the RUSLE method and processed using Google Earth Engine (GEE) between 2016 to 2021.Google offers a cloud-based framework called GEE.GEE can be operated using the Java Script programming language.GEE is a petabyte-scale data-based tool that can be used to analyze and archive open-source geospatial data.The computing infrastructure is built for geospatial data processing both data processing and visualization of spatial analysis of satellite imagery.When handling the conservation measure components and plants cover of the RUSLE model, GEE provides a substantial advantage [6].GEE has also been utilized in a variety of applications, including as image processing, classification of landforms, climatological analysis, urban planning, hydrology, and the prediction of natural disasters [7].
The use of GEE to determine erosion rates has been the subject of numerous prior research.A paper by Papaiordanidis from 2020 titled "Soil Erosion Prediction Utilizing the Revised Universal Soil Loss Equation (RUSLE) in Google Earth Engine (GEE) Cloud-based Platform" became a standard reference.This study looked into seasonal spatiotemporal effects on soil loss for the Greek Pindos mountains using RUSLE calculated in GEE.The study also calculated the association between the average RUSLE value and the seasonal components of RUSLE (rainfall and vegetation) [8].The following study, entitled "Dynamic Variations of Soil Erosion in The Taohe River Basin Utilizing The RUSLE Model and Google Earth Engine," was carried out by Wang in 2020.Google Earth Engine was utilized to extract data in this investigation.a number of parameters from the improved RUSLE model.Based on data from many sources, The soil erosion modulus of the Taohe River Basin was calculated, and the features of soil erosion intensity's spatiotemporal change were looked at.[6].
Researchers want to carry out the study since they haven't discovered a model for applying the RUSLE and GEE methodologies to calculate erosion rates in Indonesia.Numerous research in Indonesia analyzed erosion rates using various simulation and calculation methods.Research by Lestari [9] is one of the studies on data analysis utilizing information systems.In this study, classification was done using the Normalized Difference Vegetation Index (NDVI), Normalized Difference Wetness Index (NDWI), Soil Adjusted Vegetation Index (SAVI), Standardized Precipitation Index (SPI), and Interpolation methods using Inverse Distance Weighted (IDW).After that, a ranking was done to examine areas, particularly villages that may be susceptible to flooding near the Tuntang River Basin.One of the studies approached in 2022 was the preparation of geoWEPP (geospatial interface for water erosion prediction project) model input data for erosion forecasting in the Cikeruh-Citarik sub-watershed, according to research from Amaru [10].
The Tuntang Watershed Research Location, the RUSLE Model Used, and the Use of GEE as the Primary Analytical Tool are the Features that Set This Research Apart from Others.The Annual Erosion Rate of the Tuntang Watershed and the Map of Erosion Rate Estimation Based on GEE are the study's final outputs.

Area of Study
The Tuntang Watershed is located at a position of 110 ° 18' 26" -110 ° 51'01" E and between 6 ° 45' 31'' -7 ° 26' 55'' LS, with a total area of 130.  (12.400,263 Ha).Tuntang River as the main river has a length of 106.5 km with a river discharge of 722,4 m 3 / second.This region's river branches range from 4 (four) orders to 6 (six) orders.The Bancak, Senjoyo, Temuireng, and Tuk Bening Sub Watersheds have the highest order with 6 (six) orders, indicating that the river channel has a rapid rise in flood water levels as well as a rapid decline, whereas the Rowopening Sub Watershed has the lowest order with 4 orders, indicating that the river channel has an increase and decrease in flood water levels that is neither too fast nor too slow.The Tuntang watershed area has two types of river flow patterns: Rectangular dendritic (Jajar Hulu, Rowopening, and Tuntang Hilir Sub Watersheds) and Medium dendritic (Bancak, Blorong, Senjoyo, Temuireng, and Tuk Bening Sub Watersheds), which show limestone and clay soil characteristics [11].

RUSLE
For the Tuntang Watershed, Soil erosion was calculated using the RUSLE techniques.The RUSLE erosion model predicts the average annual soil loss over time brought on by flow from field slopes in specific crops and control regimes, as well as from rangeland.The following equations represent the RUSLE approach technically:

))
Where: A = Typical soil deterioration (ton hr -1.yr -1. ) R = Factor erosivity of rainfall-runoff (MJ mm ha -1.Hr -1 yr -1 ) K = Susceptibility of soil detached (ton ha hr MJ ha -1.mm -1 ) LS = Incline length along with degree C = Land cover value P = Support Practise component Then, in Google Earth Engine, these formulas are processed based on data obtained from various sources, namely: • R Factor using CHIRPS data • K Factor using OpenLandMap data • LS Factor using NASA SRTM data • C Factor using MSI's Sentinel-2 data • P Factor using MODIS Land Cover data

Factor for Erosivity of Rainfall-Runoff (R)
Precipitation is a major external force causing soil erosion and also serves as an index for assessing soil separation and transportation by rainfall.In areas where detailed rainfall data is not available, alternative equations are used to calculate rainfall erosivity instead of the standard RUSLE calculation.The longterm statistics that have at regular intervals sampling rate were used to create the R-factor equations employed in these investigations.In this study, the R-factor was obtained from Ram (cited in Jain and Das).[12] : Where: P = the region with yearly rainfall from 340 to 3500 mm (MJ mm ha -1 h -1 year -1 )

Soil Erodibility Factor(K)
The term "soil erodibility" describes how easily soil can be eroded by forces of nature including wind, water, and gravity.It evaluates the soil's resistance to separation.By means of erosive forces, soil particles are transported and deposited.Soil erodibility is influenced by factors like soil composition, structure, content of organic material, slope gradient, and vegetation cover.Understanding soil erodibility is essential in developing effective soil conservation strategies and preventing soil degradation.The properties of soil, both physical and chemical, contribute to soil erosion.The vulnerability of the soil to runoff and erosion from rainfall is indicated by its erodibility.A soil's susceptibility to erosion is indicated by the K value, which varies from 0 to 1. Using the K factor values from David as mentioned in Benavidez, the K factor is determined [13] K = Soil Texture Class x K Where: K = Soil erodibility factor (tons of acre hours / hundred acres / foot / ton / inch)

Topography Factor (LS)
The factor of topographic is related to the slope distance and gradient, mentioned as LS factor.Erosion of soil will be severe in the watershed with a high LS.This study uses equation from David as cited in Benavidez [13] : where: a = 0,1 b = 0.21; and SL = Slope (%)

Land Cover Value (C)
The C factor has something to do with the management of land cover.The ratio of soil loss to vegetation cover uses the formula: The Vegetation Index of Normalized Differences (NDVI), which result is generated from satellite photos of the infrared plus red channels, is a measure of vegetation, that show the correlation between vegetation density and leaf chlorophyll concentration using spectral values at each pixel.If satellite imagery is available, researchers can use NDVI to identify sub-annual C factors, which can help them understand how cover affects seasonal soil erosion and pinpoint key times of year when soil erosion is a concern [14].

Support Practice Component (P)
The P factor refers to the proportion of soil loss in relation to the support strategy used.Implementing control practices, such as reducing run-off volume and velocity, altering The pattern of movement, and changing the surface runoff's flow, can effectively decrease the risk of erosion.The P factor's range of values is 0 to 1. Soil erosion can be reduced by using farming techniques like contour farming, crop rotation, and shelterbelts.The soil loss ratio for slope is obtained by the formula: P Factor = Land Cover and Slope x P (6)

Spatial Variation of Soil Erosion
Spatial distribution of the yearly mean soil degradation in the Tuntang Watershed for the years 2016 to 2021 was calculated using Formula (1).The computation was based on the RUSLE model's spatial variability values for every factor, and it has been plotted using Google Earth Engine.Soil erosion in the Tuntang Watershed remained consistent from 2016 to 2021, with Moderate erosion being the primary type in most areas.The highest concentration of severe erosion occurred in Senjoyo Sub Watershed, Rowopening Sub Watershed, and Tuntang Hilir Sub Watershed, with the most severe erosion in Rowopening.The most interesting results occurred in Tuntang Hilir Sub Watershed because of its location in an estuary and experiencing large erosion (approximately 40 ton/ha/yr), this could be caused by rising sea levels and abrasion, but the sediment cannot flow into the sea.
Figure 3 displays the changes that have occurred in the Tuntang Hilir Sub Watershed.Changes in land use can have an impact on the rate of erosion in downstream areas.Abrasion and probably climate change have a significant impact on the rate of erosion in this area.The pace of erosion in this area is greatly influenced by abrasion and climatic change.This can be seen in changes in coastlines using a time lapse series.The area of average erosion that occurs in Tuntang Hilir Sub Watershed is as follows; 2021: 9.75 km 2 ; 2020: 9.49 km 2 ; 2019 : 9.58 km 2 ; 2018: 11.32 km 2 ; 2017: 11.25 km 2 ; 2016: 9.55 km 2 .Between 2019-2021 there was an expansion of areas that experienced erosion, but in 2017-2018, the area that The erosion that occurs in Tuntang Hilir is part of the erosion rate process that occurs in the Tuntang watershed.The area that falls into this weight category must be a concern because it can cause blockage in the estuary.Although the area in 2017 is larger than in 2020, the map shows some potential for widespread erosion.This potential increases from low to very high category, even at some points showing erosion rates in the heavy category.In 2019 there was an anomaly where there was an area that experienced a decrease in erosion rates in the middle of the area that experienced heavy erosion rates.This can be a concern for further research.Changes in land use can affect the rate of erosion.In Figure 4, it can be seen that there has been a change in land utilization in the form of settlements in riverbank zone.In 2016, settlements in riverbank areas still did not appear massive.meanwhile in 2021, lots of housing will grow and develop along the riverbanks.This can cause an increase in the rate of erosion so that we can see in the Google Earth Engine mapping results in Figure 3, that the rate of erosion increases from year to year. Figure 4 can illustrate an increase in the average amount of land affected by erosion, thus the increase in erosion rate is closely related to human activities in making land use changes.This can be a concern for local governments to be able to reorganize watersheds to guard against threatening disasters in the future.In addition to changes in land use, changes in coastlines were also found that may be caused by 2016 abrasion and sea level rise.Although figure 4 is not yet clearly visible, this could be the beginning of research related to the relationship between coastal abrasion and watershed erosion.

Mean Soil Loss
Google Earth Engine can calculate Mean Soil Loss from formula (1).Table 1 shows the result of Google Earth Engine Calculation.The highest mean soil loss was happened in 2020 and the lowest was happened in 2016 In 2017 there was a significant increase of 64% from 2016.Meanwhile, in 2018 and 2019 it decreased by 11.9% and 28.5% respectively.In 2020 there was a significant increase and exceeded the rate of erosion in 2017.This increase is likely due to various factors that occur along the Tuntang watershed.However, in 2021, there was a decrease although not too significant compared to 2020.The average Mean Soil loss in 6 years is 39.63588007 ton/ha/yr and it is categorized very high and very close to heavy.This can cause problems such as flooding and silting if not handled properly.The results obtained represent interesting data, with soil fluctuations that were eroded in that period and locations upstream and downstream of the Tuntang watershed, researchers provide suggestions to be able to calibrate and validate factual data so that data elaboration can be carried out.Various factors that cause erosion can be further investigated, and also the potential for erosion that has begun to appear in 2021 can be controlled and overcome.The increase in erosion rates in 2019 and 2020 also needs further investigation because it has a significant difference, which is 41%.

Conclusion
In this study, Tuntang Watershed soil loss in the years 2016, 2017, 2018, 2019, 2020, and 2021 was calculated using GEE and the RUSLE procedures.The regional variation of soil loss parameters severity grade was then investigated.The following are the primary conclusions: 1. Between 2016 and 2021, there was no significant change in the Tuntang Watershed's scope of soil erosion.Most of the areas experienced moderate erosion, with serious erosion being mostly confined in the Senjoyo, Rowopening, and Tuntang Hilir Sub Watersheds.The Rowopening Sub Watershed had the highest level of erosion among them.2. In the Tuntang Hilir Sub Watershed, the average area of erosion is as follows: 9.75 km2 in 2021; 9.49 km2 in 2020; 9.58 km2 in 2019; 11.32 km2 in 2018; 11.25 km2 in 2017; and 9.55 km2 in 2016.3. The mean soil loss was highest in 2020 and lowest in 2016.The average soil loss over 6 years was 39.63588007 tons per hectare per year, categorized as very high and close to heavy.The biggest difference occurred in 2019 to 2020, which was 41%. 4. Climate change can be a factor that exacerbates erosion.Further research regarding the relationship between climate change and erosion rates using Google Earth Engine can be carried out by adding and simulating various factors that cause erosion.So, predictions of erosion rates can be determined.
wider, this deserves further research, because there are other possible factors that can cause changes in sub watersheds in the upstream area or due to land change factors in the downstream area.While in 2016, the area that experienced erosion was smaller than in 2017 and 2018.