Study area

The Mt. Bawakaraeng Caldera is located in the province of South Sulawesi, Indonesia (Figure 1), at 2 830 masl. The morphology of Mt. Bawakaraeng is characterized by a high relief, extreme slopes, high degree of weathering, as well as events related to erosion such as soil movement and landslides (^{Hasnawir and Kubota, 2010}). The vegetation of the mountainous areas is composed of natural forest and forest plantations of Pinus merkusii Jungh. & de Vriese. The basement around the area of Mt. Bawakaraeng, consists of Tonasa formation, Camba formation, Baturape-Cindako volcanic rocks, extrusive and intrusive rocks and Lompobattang volcanic formations, whose origin dates from the Paleogene to the Quaternary period.

Figure 1 Study area of the Mt. Bawakareang Caldera, Indonesia. 

The soils in the study area are Andosols and are formed with volcanic materials; many have excellent internal drainage, due to their high porosity and their occurrence in high-elevation positions. The specific gravity of the soil is 2.70, the porosity is estimated between 66.7 % and 88.7 %, while the apparent density is low and not only on the surface of the land; it is typically> 900 kg m^-3, but can be as low as 300 kg m^-3, values that have been obtained in highly hydrated andosols (^{Hasnawir and Kubota, 2011}).

Methods

Erosion risk analysis requires an interdisciplinary approach involving disciplines such as hydrology, geomorphology and pedology (^{Ghosh and Bhattacharya, 2012}). In this case gully inventories related to debris flow disasters were used in the Mt. Bawakaraeng Caldera from 2007 to 2017. The inventory was based on information from the local government office in South Sulawesi, in field investigations and soil tests.

A standard laboratory analysis was performed according to the methods of the American Material and Testing Society (ASTM) to measure the particle size distribution. Three samples were taken per site (natural site, deposit of sediments in the upper part and deposit of sediments in the lower part). The difference between the particles (<75 μm) was determined with a No. 200 sieve. The standard method ASTM D 422 was applied to establish soil texture related to erosion. Soil permeability (k cm sec^-1) was calculated using equation 1 and simplified equation 2, which establish the permeability capacity of a sample without disturbance. As permeability values increase, there is an increased risk of erosion.

This method is widely used to calculate the coefficient of permeability of a sample without disturbance, due to its practicality in the field (^{Paudel et al., 2008}; ^{Aditian y Kubota, 2014}; ^{Sánchez-Castillo et al. 2017a}). To perform this method, two cylinders with the same dimensions are adjusted (Figure 2). One contains the soil sample, previously saturated, and the other is filled with water. Permeability is defined as the difference in time to drain all water from the upper can (^{Das, 1997}):

Where:

h1h2=2

a=A

L=5 cm

t2-t1 = Δt

a and A = Crosscutting area of the upper and lower cylinders

The equation was modified by replacing the values in equation 1, the permeability can be estimated as:

Figure 2 Method used to measure the permeability of the soil. 

The Universal Soil Loss Equation (USLE) is an empirical method to determine relevant parameters among the many physical variables such as climate, soil profiles, relief, vegetation and land use and practices of area management, respectively indicated by the factors R, K, LS, C and P. The USLE equation was used in this investigation to describe the level of soil erosion that occurs in the Caldera of Mt. Bawakaraeng.

It is pertinent to specify that this equation can not be used to establish the risk of erosion in gullies, as indicated by ^{Wischmeier (1976)} in his discussion of the uses and misuses of the USLE, when he indicates that the equation can be used to estimate the average floor loss in a particular segment of a slope represented by a selective topographic factor. The parameters in the GIS framework include erosion factors such as rainfall, erodability, topography and coverage (^{Duru, 2016}). The equation is expressed as follows:

Where:

A = Forecast of the average soil loss per unit area [ton ha^-1 year^-1]

R = Rain-runoff erosion factor (rain) in [MJ mm ha⁻¹ hr⁻¹ year^-1]

K = Soil erodibility factor [ton ha hr ha⁻¹·MJ⁻¹·mm⁻¹]

LS = Slope length and slope factor (without dimensions)

C = Coverage-management factor (radius of soil loss from a specific area)

P = Support factor in conservation (without dimensions)

Soil analysis

The sieve analysis method covers the quantitative determination of the size distribution of soil particles (Table 1).

Sediment deposits have a higher percentage of sand than those of natural sites. The higher sand content describes a soil more prone to erosion, compared to those with lower values. It even indicates the possibility of erosion in gullies, when there are major events of rain in the area of sediment deposits.

The soil in eroded sites could have been formed by the interaction of many factors such as parental material, climate, weather, topography or relief and microbial activity. The relative influence of each factor depends on the geographical conditions and the location, but the combination of all of them, in general, determines the type of soil and the thickness that develops in any place (^{Huggett, 1997}).

The results of soil permeability reached intervals from 0.0098 cm sec^-1 to 0.0362 cm sec^-1 (Table 2).

As the permeability value becomes higher, erosion is more likely to occur. Sediment deposits in the lower part of the Caldera showed higher permeability values (Table 2).

Risk of Soil Erosion

The various factor maps of the USLE model (R, K, LS, C and P) generated in a GIS environment using Arc GIS 10.1 and ERDAS Imagine 2017 and the associated GIS packages were integrated with the USLE model to compute the erosion rate and its severity, which resulted in the soil erosion risk map of the Mt. Bawakaraeng Caldera. This erosion was classified as: very low (≤15 tons ha^-1 year^-1), low (16 - 60 tons ha^-1 year^-1), moderate (61 - 180 tons ha^-1 year^-1), high (181-480 tons ha^-1 year^-1) and very high (> 480 tons ha^-1 year^-1). The spatial distribution of the erosion risk class revealed that 3.53 % of the Caldera is low, 12.87 % very low, 64.06 % moderate, 0.06 % high and 19.49 % of the area is under a great risk of erosion (Table 3 and Figure 3 ).

Figure 3 Map of erosion risk in the Caldera of Mt. Bawakaraeng, Indonesia. 

The highest values of risk of soil erosion are observed mainly on steep slopes adjoining drainage lines and rivers.

According to the results of the data collection of debris flow events from the government office of the province of South Sulawesi (Hazama-Brantas JO and Central River Region Pompengan-Jeneberang) there are 22 disasters related to erosion in gullies in the sediment deposits in Mt. Bawakaraeng from 2007 to 2017. The category of debris flow disaster was analyzed based on the volume of sediments and the impact of the debris flow. This impact includes damage to roads, destruction of agricultural land and sediment load to the catchment dam in the area (Table 4). Until now, there has not been a casual event. From the information provided by the communities living around the Caldera of Mt. Bawakaraeng no affectations have occurred. However, the incidence of debris flow promoted by gully erosion becomes a threat to people who carry out activities in the Caldera or its surroundings, which brings together 3 172 inhabitants in the vicinity.

From the point of view of disaster prevention, the construction of sabo dams (dams built in the high parts of the mountains that prevent sediment runoff) would control erosion and sedimentation in the Mt Bawakaraeng Caldera; however, the procedure is expensive and time consuming, so it is necessary to have an early warning system in the area, especially in the rainy season when events of great intensity usually occur (November-April).

^{Hasnawir and Kubota (2010)} reported that average rainfall above 3 000 mm per year occurs in the Mt. Bawakaraeng Caldera. Rainfall thresholds as an application of an early warning system for the prevention of sediment-related disasters according to the methodology proposed by ^{Sánchez-Castillo et al. (2017b)} could be developed in case there is not enough specific historical data in terms of time (minutes) and place. Figure 4 shows erosion in gullies in the sediment deposits at Mt. Bawakaraeng Caldera.

Figure 4 Erosion in gullies and sediments deposited in Mt. Bawakareang Caldera.