RAINFALL CHARACTERISTIC ON THE SLOPES OF MOUNT MERAPI REGION

Debris flow phenomenon on the slopes of Mount Merapi area has potentially became catastrophic natural due to its great destructive force and velocity. One of many triggers is the rainfall with a certain intensity and duration. The temporal and spatial characteristic of rainfall varies and influenced by various factors, such as topography and climate. Dharma (2012) suggested defining the characteristics of the intensity of rainfall using rainfall data with a shorter duration according to statistical tests to establish the best empirical IDF formula. This study was using 30 minutes rainfall data to determine the representative duration and distribution pattern of rainfall. The empirical formula of Sherman, Kimijima, Haspers and Mononobe method is applied to determine the relationship of rainfall intensity and duration. Debris flow occurrence and rainfall data were analyzed using method A of MLIT guidelines to establish standard rainfall for the warning and evacuation criteria of debris flow prediction. Sherman formula has shown the best fit to the IDF characteristics of the rainfall on Mount Merapi. Rainfall distribution pattern has shown the high intensity rainfall in the first hour and then decreased in the next hour which means the distribution for the duration of 6 half-hours are, 12%, 28%, 25%, 16%, 12%, and 7%, respectively. Based on the critical line, 5 mm of standard rainfall was gained for warning (R1) and 28 mm for evacuation (R2).


INTRODUCTION
Mount Merapi eruption in 2010 has damaged and caused huge losses for Yogyakarta and Central Java Province.Debris flow has caused fatalities and major damage to public and private infrastructures.Moreover, the volume of ejected material is estimated nearly 140 million that is led to debris flow disaster.Debris flow on the slopes of Mount Merapi is triggered by the occurrence of rainfall with a certain duration and intensity.In order to determine the characteristics of short duration rainfall (<3 hours), the analysis of extreme rainfall distribution as the form of rain magnitude, intensity, duration and frequency was carried out.That rainfall characteristic information at Mount Merapi can be used for analyzing the debris flow.This research aims to determine the rainfall intensity-duration formula analytically and empirically which is in accordance with the short duration rainfall characteristics, and rainfall distribution, and determine the Critical Line (CL), Warning Line (WL), and Evacuation Line (EL) in order to develop warning system regarding to debris flow in Woro River.
Rainfall characteristics is expected to be able to support the development of early warning systems in Mount Merapi corresponding to debris flow by applying rainfall intensity (mm/h) and working rainfall (mm).

Debris Flow Prediction using Critical Line
Rainfall intensity is the depth of rain falling into the ground surface per unit time, usually in mm/h and mm/day (Triatmodjo, 2008).According to Sosrodarsono and Takeda (1985) rain can be classified based on its intensity.
Debris flow is a flow of a mixture of water with a high concentration of sediments slide down through the high slopes.Debris flow is not directly related to the eruption, but it can occur in volcanic and nonvolcanic regions (Kusumobroto, 2006).Such conditions, which may lead the debris flow, are steep slope of the valley, sediment material availability around the mountain as the ejected material, and rainfall in considerable amounts.One sequence of rain within more than 24 hours of non-rainfall duration before and after is called by "a series of rain".Total amount of rainfall during that period is called by "continuous rainfall (RC)" This relationship is shown in Figure 1.
The working rainfall (R W ) is a cumulative rainfall which affected by antecedent rainfall.In general, debris flows occur under the influence of under not only rainfall triggering debris flow, but also antecedent rainfall.The degree of influence of an antecedent rainfall is deemed declining as duration time of causing rainfall increase.CL curve separates rainfall data of debris flow occurs termed as causing rainfall and rainfall data which not directly caused the debris flow termed as non-causing rainfall (see Figure 2).

Spatial and Temporal Characteristic of Rainfall
Spatial characteristics of rainfall in Mount Merapi area can be analyzed by considering several matters, i.e. correlation between elevation of rainfall station and frequency of rainfall intensity, and distribution or rainfall intensity, total day rainfall, and annual rainfall which can be calculated by comparing elevation and location of rainfall station towards wind direction.In order to get the correlation between elevation of rainfall station and annual rainfall, the following formula (see Equation 1) is used: where R is annual rainfall depth (mm), a and b are constant values for curve slope, while H is elevation (m).
Temporal characteristics of rainfall in Mount Merapi area can be analyzed using rainfall intensity and frequency of rainfall occurrence, which is shown at Equation ( 2): where I T is mean rainfall intensity at T-time (mm/hour), R T is rainfall depth during T-time (mm), and T is duration of rainfall (hour).
The accuracy of IDF analysis can be observed using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE).Formula of RMSE and MAE is shown at Equation ( 3) and ( 4): where m is number of variation of rainfall duration, n is number of variation of return period, I jk is analytical rainfall intensity for j duration and k return period, and

RESULTS AND DISCUSSIONS
The results of frequency analysis of rainfall intensity for a return period of 2, 5, 10, and 25 years at Ngepos Station can be seen at Table 2.The methods for determining the performance criteria or model calibration against field observations can be approached using Root Mean Square Error (RMSE) Method (Drogue, et al., 2002).Table 5 and Table 6 show that Sherman formula has the least error among other formulas.Average of rainfall distribution for whole stations has a similar pattern.The first hour of rainfall is the highest, then declined for the next hours (see Figure 6).

Figure 1 .
Figure 1.Concept of a series of rain and the antecedent rain

Figure 5 .
Figure 5. IDF Curve at Ngepos Station with 2 years of return period.

Figure 6 .
Figure 6.IDF Curve at Ngepos Station with 5 years of return period.

Figure 7 .
Figure 7. IDF Curve at Ngepos Station with 10 years of return period.

Figure 8 .
Figure 8. IDF Curve at Ngepos Station with 25 years of return period.
The comparison between rainfall intensity and working rainfall at Deles and Batur station are shown in Figure10(a) and 10 (b), respectively.Critical line divides the area into two zones, namely safe zone and unsafe zone.Critical line for Woro River case is depicted at Figure811.It is created from two points of causing rainfall at the most left side and extends to cross X and Y axis.Safe zone consists of non-line crossing Y axis has value of 17mm/h.Since RH1M at Station Deles is 18 mm/h (>17 mm/h), the evacuation line from intersection of RH1M and critical line apparently could not be drawn.RH1M value from rainfall data at Batur Station is 44 mm/hour, while RH2M value is 67 mm.RH1M that is 44 mm/hour horizontally crossed critical line, then created vertical line crossing x axis, so called evacuation line.The evacuation line crossing x axis value is called R2. R2 value is 28 mm.Next step is to create warning line.RH2M minus RH1M is equal to 23 mm (RH2M -RH1M).The warning line is created from parallelize evacuation line to the left 23 mm (RH2M -RH1M).The warning line crossing x axis value is called R1. R1 value is 5 mm.

Figure
Figure 9. Average of rainfall distribution * jk is rainfall intensity based on empirical formula., while Deles and Batur rainfall stations and PTMA Karangbutan are used for CL, WL and EL analysis.The locations of each rainfall station are listed in Table1.The selection of the five rainfall stations are based on data availability and distribution of 30 minutes rainfall, which representing variability of elevation.Deles station is chosen since it is located in the most upstream area.Figure3shows location of research.Research flow chart is depicted in Figure4.
o 35' 48.7" 904 Sidorejo, Kemalang Sub-district, Klaten Regency Figure 3. Location of research area RAINFALL AND WATER LEVEL STATION MAP IN MERAPI MOUNTAIN AREA ARR Batur

Table 2 .
Rainfall intensity at Ngepos Station

Table 3 .
The constant empirical formula at Ngepos Station

Table 5 .
RMSE test results of the rainfall intensity formula

Table 6 .
MAE test results of the rainfall intensity formula