Mapping the vulnerability of irrigation sand traps in a tropical volcanic basin, Indonesia

Sand traps in irrigation networks are typically used in mitigating canal sedimentation. In irrigation networks located in basins of high sediment yield due to the presence of volcanoes, it is essential to assess the vulnerability of sand traps. Using sediment yield at irrigation scheme inlets, sand trap vulnerability can be evaluated. This study aims to understand the vulnerability of irrigation sand traps throughout the Progo–Opak–Serang (POS) Volcanic River Basin, Indonesia, via mapping the sediment yield distributions in the basin. We employed the Revised Universal Soil Loss Equation to estimate soil loss, where the results show that the average soil loss in the POS River Basin is 179.69 tons/ha/year that falls under the category of moderate erosion potential, while the average sediment yield for the whole basin is 51.04 tons/ha/year. Parts of the basin with high yields of more than 180 tons/ha/year were mostly found along the volcanic mountains such as Sindoro, Sumbing, Merapi, Merbabu, and Telomoyo, and the Menoreh Hills. The model demonstrated relatively high performance with R2, NSE, RMSE, and MAE of 0.89, 0.82, 0.14, and 0.11, respectively. Within the POS Basin, Badran, Kalibawang, and Blawong are the three most vulnerable irrigation sand traps, with sediment yield values of 252.83, 178.92, and 63.49 tons/ha/year, respectively; they are all located in sub-watershed outlets. The vulnerability assessment conducted in this study can be used for the decision support system to prioritize irrigation sand traps towards a more effective irrigation system development.

In irrigation networks located in the basins of high sediment yield due to the presence of volcanoes, it is essential to assess the vulnerability of sand traps 18 .The volcanic river basin refers to this particular river basin 19,20 .There are 133 river basins in Indonesia, each corresponding to one of the 34 River Basin Organizations.Among these, 127 are volcanic river basins 14 .Mount Merapi, one of the globe's most active volcanoes, is situated in the Progo-Opak-Serang (POS) River Basin in Indonesia, hence the designation POS Volcanic River Basin 21 .In the past 200 years, Mount Merapi has erupted 41 times, with 15 of those eruptions being large.It typically erupts every three years, with major eruptions occurring every nine years 18 .The eruption material from volcanoes carried by debris flow affects river sedimentation 22 .Sedimentation causes siltation and damages sand traps and irrigation structures which draw water from the related river 23 .
Sediment yield at irrigation scheme inlets can be used to assess the vulnerability of irrigation sand traps.In general, sediment yield estimation is needed for soil and water conservation studies, including reservoir or irrigation channel sedimentation 24,25 .In Indonesia, there is limited study on sediment yield estimation in the volcanic river basin.A previous similar study had only examined a small-scale young volcanic catchment using a longduration field measurement method and had been done 30 years ago 26 .It was discovered that the rainfed agricultural land contributed nearly half of the soil erosion, contributing to the sediment yield 26 .Updated research with a broader scope is needed that discusses the sediment yield pattern in the volcanic basin and its integration including the irrigation system.This can contribute to integrated water resource management.Another study recently conducted in 2019 modeled precipitation-runoff and sediment production at the Upper Opak Basin, that is portion of the POS River Basin.It was found that the highest sediment yield of the area from 2004 to 2013 was 147.8 tons/year.There is no further discussion regarding the comparison with other volcanic catchments or how this sediment yield relates to the irrigation system.In addition, there is no study on the vulnerability of irrigation networks based on sediment yield.Several existing studies discuss susceptible watersheds to sediment yield and sediment yield estimation to assess the area's vulnerability to soil erosion [27][28][29][30][31][32] .
There are several models used to predict soil erosion on a watershed scale, including empirical models, namely the universal soil loss equation (USLE) as well as its revised and modified versions (RUSLE and MUSLE) 33 ; and a physical-based model, namely soil and water assessment tools (SWAT) 34 .USLE is a simple model with accessible and well-established parameters, so it is widely used in various parts of the world, while its disadvantage is that it cannot measure sedimentation [35][36][37] .RUSLE is a refinement of the USLE model, and this model will be better if combined with Geographic Information Systems (GIS) 25 .RUSLE is usually used for estimating annual erosion.The disadvantage of this model is that it cannot measure the amount of sediment produced 38 .MUSLE is a modified version of RUSLE that includes additional parameters that are not always accessible.The advantage of this model is that it can show the amount of sedimentation 39 .While SWAT is a complete model, it contains parameters of information about climate, soil properties, topography, plants, and land management contained in the watershed.It has been integrated with computers so that it is efficient to use, but one disadvantage is that the parameters used are quite large and must be filled when the model is operated, even though the desired results do not actually require these parameters 40 .
In this study, we map the sediment yield pattern that is able to estimate the vulnerability of irrigation sand traps throughout the POS Volcanic River Basin in Indonesia.We employed the revised universal soil loss equation (RUSLE) to forecast soil loss within the basin because of its ease of application and highly accurate projections predicting the quantity of erosion produced, making it one of the most extensively used models in research 38,41 .
In conjunction with the geographic information system and remotely sensed data, the RUSLE model will generate more precise and reliable estimations 38,42 .The RUSLE model's erosion prediction findings are utilized to estimate sediment yield patterns in the river basin 41,[43][44][45] .The results can be used to determine the most vulnerable sand traps and prioritize them for sedimentation studies.It intends to investigate the current performance of sand traps as the principal sedimentation blocker in irrigation networks.Adequate sand trap performance can enhance the overall efficiency of the irrigation system.This study can serve as a resource for decision-makers in managing sedimentation in river basin-integrated irrigation networks.

Methodological framework
This study uses the RUSLE model to predict average soil loss in the POS River Basin.It is used to estimate the distribution pattern of sediment yield in the river basin with the help of the sediment delivery ratio (SDR).The estimate of the SDR is derived from the delineation of the sub-watershed area within the basin derived employing the digital elevation model (DEM).The predicted findings of the sediment yield distribution obtained based on the model were then validated by using data pairs of total suspended solids and discharge at several measurement locations throughout the basin.Figure 3 illustrates the framework of this study.

Data collection
The datasets used in the analysis were obtained from several institutions with the format, resolution, and time frame as shown in Table 1.Most data were collected through the Large River Basin Organization of Serayu-Opak (BBWS SO).This information includes tabular data on rainfall and daily discharge (gathered from automatic water level recorders, or AWLR) and vector maps of the POS River Basin's soil type and catchment area.We also collected data on total suspended solids at 10 locations within the basin from the Department of Environment and Forestry of the Special Region of Yogyakarta (DLHK DIY).Other raster maps are in the form of national DEM from the Geospatial Information Agency of Indonesia (BIG) and land cover from the European Space Agency (ESA).

Prediction of soil erosion
Due to data availability issues, RUSLE presented various ways of determining parameter values other than the original concept of USLE 47 .The RUSLE model is only capable of anticipating the rate of soil erosion due to sheet erosion and does not include trench erosion.The model cannot directly forecast sediment yield 34 .The RUSLE incorporates rainfall-runoff erosivity, soil erodibility, length and steepness of the slope, crop management, and supports practice factors.According to 48 , it is defined by Eq. 1.
where A is average potential soil loss (tons/ha/year), R is rainfall-runoff erosivity factor (MJ•mm/ha/hour/year), K is soil erodibility factor (tons•hour/MJ/mm), LS is length and steepness of slope factor (dimensionless), C is crop management factor (dimensionless), and P is support practice factor (dimensionless). (1) Overall methodological framework.The RUSLE is the preferred approach for predicting sediment yield throughout the POS River Basin.The rainfall-runoff erosivity (R) component represents the soil's capacity being eroded and carried downstream as a result of precipitation intensity 48 .R is the total rainfall energy at a 30-min intensity, as computed by the equation R = EI 30 , in which E is the rainfall kinetic energy and I 30 is the 30-min rainfall intensity 49 .Rainfall intensity is a significant element influencing R value; hence persistent precipitation data are necessary during R calculation.The R value is frequently derived from daily or monthly precipitation data 50 .An R value is computed in this study using the Lenvain equation using monthly rainfall in centimeters as an input (Eq.2).It is derived on studies using precipitation data collected at various places on the Indonesian island of Java 51 .
where R is rainfall-runoff erosivity factor (MJ•mm/ha/hour/year), P m is monthly rainfall (cm) For the Lenvain formula to estimate the value of R, the average monthly precipitation in centimeters is necessary.There are 83 rainfall stations within the POS River Basin, with 54 stations in the Progo watershed, 25 in the Opak watershed, and 4 in the Serang watershed.Each rainfall station has a different available data length.The Godean station, located in the Progo watershed, has the longest accessible data, spanning 46 years from 1978 to 2020.Before they can be used for R factor analysis, the entire rainfall data from all rain stations must be checked for consistency using a homogeneity test.Assessing the homogeneity of a time-series dataset of rainfall is critical for discovering data inconsistencies caused by non-climate-related variables, such as equipment failure, operator mistakes, and unexpected changes in the nearby region of the instrument 52 .In this study, the rescaled adjusted partial sums (RAPS) method was employed to deal with the homogeneity test.
The soil erodibility (K) factor expresses the soil's vulnerability to separation and transfer as a result of erosional energy 43 .K value can be measured directly using a Nomograph or a reference value.The Wischmeier and Smith method 48 , that incorporates soil nutrient content, structure, and permeability, is often used for direct measurements.For several soil types on Java Island, K values in this study were established using reference values, as stated in Table 2 53 .
The length and steepness of slope (LS) factor indicates the correlation between slope length and steepness and soil erosion.The slope ratio was symbolized as L, which was identified as the slope length component approximated from the place where water flow first occurred above the soil surface to the site where precipitation first occurred 54 .S represents the slope steepness factor, which is given in slope angle degrees or percentage (%) 54 .In actuality, L and S variables are concurrently calculated as an LS component.Consequently, LS is referred to as the ratio of the amount of erosion on a surface area with a specific slope length and gradient compared to the identical land with a slope length of 22 m and a gradient of 9 percent 42 .Equation 3 can be employed to determine the LS's value.
where g is slope steepness (%), x is slope length (meters) Using Eq. 3, the slope steepness is used to get the LS component.Table 3 contains the LS categorization depends on the class of slope steepness 54 .The quantity of the LS component is established by examining the slope class map obtained from the DEM data analysis and then matched with the LS value table.
The crop management (C) refers to the impact of vegetation cover on soil erosion 55,56 .As shown in Table 4, the C factor values range from 0 to 1, with 0 indicating that the area is completely protected from erosion 47 .
The support practice (P) component represents the influence of soil erosion management techniques.P values vary from 0 to 1, with 1 signifying a region with no conservation practices, often known as undisturbed.We utilized the Wener Equation (Eq.4) to estimate the P factor due to the inadequate observed data for river basin conservation efforts 57,58 .The DEM is used to generate the slope.
where P is the support practice factor, S is the slope (%)

Prediction of sediment yield
We estimated the sediment delivery ratio (SDR) that is necessary to forecast sediment yield at the catchment outlets 53,59 .This study calculated SDR based on the sub-watershed division within the POS River Basin.The procedure for delineating the sub-watershed using ArcGIS is using a DEM as an input and then creating fill sinks, flow direction, flow accumulation, pour points, and watersheds.After obtaining the sub-basin division, the SDR can be calculated using the formula adapted from the USDA SCS, as demonstrated in Eq. 5. We calculated the sediment yield (SY) by multiplying the RUSLE model results (spatial distribution of potential soil loss/A) by the SDR as demonstrated in Eq. 6 32,60 .
where SDR is sediment delivery ratio (dimensionless), A c is catchment area (km 2 ), SY is sediment yield (tons/ ha/year), A is average potential soil loss (tons/ha/year).

Model validation
The prediction of soil loss based on RUSLE needs to be validated using field measurement data.The data available is in the form of TSS at 10 locations in the POS River Basin.Most of these data were measured in tributaries of the Opak Watershed.In general, data is accessible for eight years, namely from 2013 to 2020.In each year, there are three data collection periods, namely in February, May, and September.Since the available TSS data is not continuous, developing a continuous data approach is necessary using the available daily discharge data.A rating curve depicting the connection between flow discharge (Q) and sediment discharge represents this method (Qs).For this reason, it is necessary to find a TSS measurement location where there is a discharge recording station using an AWLR.The POS River Basin has 32 AWLR stations: 15 in the Progo watershed, 12 in Opak, and 5 in Serang.Based on the investigation, eight TSS measurement sites are adjacent to the AWLR.The results of the TSS measurement (mg/L) are paired with the Q (m 3 /s) on the same day so that the Q-TSS data pair can be obtained.Then calculate Q s with input Q and TSS as shown in Eq. 7.
After obtaining the Q−Q s data pair, a rating curve can be made by the power regression.The rating curve is used to generate daily Q s data along with the accessible daily Q data.Therefore, cumulative sediment volume (m 3 ) can be obtained, converted into sediment load in tons/year, and finally converted into an average sediment yield in tons/ha/year.The results of the sediment yield measured at the location concerned are then compared with the value of the sediment yield at the same location derived from RUSLE modeling findings.The comparison is evaluated by performance indicators, namely coefficient of determination (R 2) , Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Nash-Sutcliffe efficiency index (NSE).
where Qs is sediment discharge (m 3 /s), Q is river flow discharge (m 3 /s), TSS is total suspended solids (mg/L), ρ is the average density of sediment (kg/m 3 ).

Contributing Factors to soil erosion
The findings of the homogeneity test of rainfall data using the RAPS method indicate that 40 out of 83 rainfall stations have consistent data, including 25 stations in the Progo Watershed, 12 stations in the Opak Watershed, and 3 stations in the Serang Watershed (see Fig. 4a).There is no noticeable trend in the distribution of stations recording consistent precipitation.The inhomogeneity of the data is the result of missing or imperfectly recorded data in the field, which is the consequence of a variety of factors, including the negligence of recording officers, the damage to rainfall recording equipment caused by a lack of maintenance, and sudden changes in the condition of the equipment.In 2010, most rainfall data obtained at stations spread on the slopes of Mount Merapi or in the middle portion of the Opak Watershed and upstream of the Progo Watershed was not recorded.This is presumably related to the October 2010 eruption of Mount Merapi, but we lack sufficient information to be specific.In contrast to the Progo and Opak watersheds, where only half of the rainfall stations have consistent data, three out of four stations in the Serang have consistent data.This is since the rainfall station in the watershed were installed in the 2000s as opposed to the 1980s (such as Progo and Opak).
The consistent rainfall stations optimize the monthly precipitation and R factor.Figure 4b depicts R values interpolated with inverse distance weighted (IDW).It demonstrates that the average R factor value for the entire POS River Basin is 1656.66MJ•mm/ha/hour/year.In the upstream and central regions of the Progo Basin, as Indonesia is positioned between two continents (Asia and Australia) as well as two oceans (Pacific and Indian).The monsoon circulation is caused by changes in physical attributes between the ocean and the continent 61 .As a monsoon region, Indonesia experiences considerable rainfall in the summer and autumn seasons.In monsoon regions, an orographic effect can enhance the amount of precipitation on the windward slope 61 .The windward slope is the side of a mountain, island, or other tall things that the wind hits.This causes some regions of the POS River Basin to experience intense rainfall, as it covers hilly and mountainous terrain.Moreover, maximum rainfall intensity data indicate that the west and southwest of Mount Merapi experience more precipitation than the southeast 62 .The south-eastern area of Mount Merapi is shown by the green space on the lower right in Fig. 4b.The K value is strongly connected to soil physical parameters, including texture, structure, permeability, and organic content 50,63,64 .The erosion resistance of soils with a higher permeable level and a significant amount of organic matter is enhanced.Figure 4c depicts a map of the soil types within the POS River Basin with six distinct soil types, including alluvial, grumusol, latosol, litosol, regosol, and yellow-red latosol, with proportions of 7%, 8%, 36%, 25%, 24%, and 0.0001%, respectively.Latosols, litosols, and regosols dominate this river basin, with most latosol types covering the upstream area of the Progo Watershed, litosol covering the downstream portion  To account for the LS component, we utilized the DEM-derived percentage slope.The slope of the terrain influences surface water runoff, recharge, and flow 68 .In general, it is classified into categories such as moderate (8-15%), moderately steep (16-25%), and steep (26-40%) 67 .Figure 4e depicts the slope class.The map shows that the Progo, Opak, and Serang Watersheds have average slopes of 16.95%, 18.74%, and 15.48%, respectively, while the whole river basin has a moderately steep slope of 16.38%.The Slope class was matched with Table 3's LS reference values.Figure 4f shows LS factor spatial distribution; it shows that the average LS factor for the river basin is 2.72, while Progo, Opak, and Serang Watersheds have 2.79, 3.26, and 2.58, respectively.Land slope distribution has a considerable impact on LS factor values and is one of the most important factors controlling the erosion process.LS increases the velocity of surface runoff, resulting in the separation of soil particles and erosion 69 .Understanding the slope and LS value distribution facilitates soil erosion control approaches that use engineering techniques such as terracing to lessen slope length and gradient impacts 47 .
C factor attempts to measure the contribution of management techniques (tillage, cover crops, and plant residues) to agricultural land soil loss 70 .When establishing the C component in RUSLE, land use and land cover classification are considered.Figure 4g indicates that 73.61% of the POS River Basin is dense forest, followed by 13.10% agricultural land and 7.09% built-up land.We assigned C values based on these categories, and Fig. 4h depicts the spatial distribution of C component.It shows that the average C component throughout the POS River Basin is 0.022, whereas the average C factor for every watershed is 0.024, 0.022, and 0.019 for the Progo, Opak, and Serang Watersheds, respectively.The C factor values range from 0 to 1, with a value of 0 indicating total protection against erosion.According to the average value of component C, it may be inferred that the POS river basin has adequate erosion protection.This is in accordance with Fig. 4g that most of the land is densely forested.Numerous ecosystem services are offered by forest, notably protection against soil erosion.Forests mitigate soil erosion by providing a surface cover using tree crowns that can reduce the kinetic energy of rainfall by intercepting precipitation 71 .Using the predicted value of the C factor, we can determine an area's susceptibility to erosion due to inadequate land management.Consequently, the C component is perhaps the most important aspect in policymaking and land use considerations, since it reflects the circumstances which are most easily managed to reduce erosion 70 .
P factor is the last component contributing to soil erosion in RUSLE.It describes the rate of soil loss produced by a particular soil conservation or protection method to slope loss 57 .It is a type of conservation used to evaluate the soil's space or rate of soil erosion with preservation treatments such as terracing, crop pattern, and total soil erosion controlled in accordance with the slope under the same conditions 42 .Limited in situ data prompted us to use a modified version of the Wener equation to estimate the P factor with the input parameter being slope in percent (Eq.4) 57,58 .The spatial pattern of P component is illustrated by Fig. 4i; the value spans from 0.2 to 131.8, with an average of 5.1.Typically, P scores range from 0 to 1, with 1 being a territory without conservation efforts.The high P values result from the slope value, which is utilized as an input in the computation, is more significant than 100% at many locations in the POS river basin.A slope greater than 100% is feasible because some POS locations are mountainous and hilly.The location has a high slope degree, with a massive rise over a relatively short run, resulting in an extraordinarily high slope percentage.
All soil erosion factors are generated in raster format for use with a raster calculator to predict the average soil loss, and Fig. 4j displays the spatial pattern of the outcomes.The average soil loss in the POS River Basin is 179.69 tons/hectare/year, with 255.31, 159.24, and 96.73 tons/ha/year for the Progo, Opak, and Serang watersheds, respectively.In accordance with the Indonesian Department of Forestry's guidelines, the soil loss values can be classified according to Table 5 47 .The POS River Basin is under the category of moderate erosion potential, whilst the Progo watershed falls under the category of heavy erosion potential.According to 72 , 38% of the POS river basin was categorized as critical land in 2010, with an erosion rate of 235 tons/hectare/year.Compared to these findings, the erosion rate obtained in this study is realistic, and the POS River Basin has exhibited a declining trend in erosion risk over the previous decade.It should be emphasized, however, that each model will produce values that are not identical since they employ various inputs, methodologies, and assumptions.We also examined additional data in the Progo and Serang watersheds, although there were no comparative investigations in the Opak watershed.The average erosion rate in the Progo Basin increased from 165 tons/hectare/year just in 1999 reached 184 tons/ha/year in 2011 73 , whereas the erosion hazard level in the Upper Progo Sub-watershed in 2016 was 178.432 tons/ha/year 74 .According to 75 and P, is a method that has been proven and well-established.The five factors determine the average potential soil loss in tons/ha/year.However, the magnitude of the contribution of each factor is different for various study areas in the world.In this regard, we conducted a correlation analysis to check how significant the five RUSLE factors are in calculating average potential soil loss in the POS River Basin.The correlation coefficient of RUSLE factors with average potential soil loss is shown in Table 6.The five factors (R, K, LS, C, and P) positively correlate with average potential soil loss (A).The highest coefficient value is shown by the P factor, followed by LS (0.054 and 0.0521).This results from the study area being dominated by hilly and mountainous areas, with high slope values.Furthermore, the LS factor is also closely related to the P factor, which is the main component in P factor mapping.The strong influence of the LS factor is in line with findings from other research, which show that the LS factor is one of the most significant agents contributing to soil erosion, representing the combined effect of slope steepness and length [76][77][78] .

Sediment yield distribution and validation
The sediment delivery ratio (SDR) relates the quantity of sediments degraded and carried from a watershed's gradient to the quantity that reaches the watershed's outflow via streams and rivers 79 .SDR shows the relationship between the amount of erosion across the entire basin and the sediment yield at the watershed outlet 80 .The SDR value varies from 0 to 1, and a number close to one indicates that all soil moved by erosion enters the river.The assessment of annual sediment yield and sediment delivery has been a primary concern for watershed management authorities due to the impact of sedimentation on reservoir storage capacity and the yearly financial impact of disasters caused by sedimentation 79 .As depicted in Fig. 5a, the SDR was determined according to the division of the sub-watershed region in the POS River Basin.The average SDR within the POS River Basin is 0.28, and sub-watersheds originating in Mount Merapi, including the Bedog, Winongo, Gadjahwong, Code, Gawe, and Kali Kuning sub-watersheds, have the highest SDR values (0.33-0.39).Additionally, the downstream subwatersheds of Progo, Opak, and Serang also have a high SDR.After obtaining the spatial variability of average soil loss (A) and SDR, the distribution of sediment yield in the river area can be estimated.The spatial pattern of sediment yield can be seen in Fig. 5a.The average sediment yield in the POS River Basin is 51.04 tons/hectare/ year, with values of 71.75, 50.23, and 27.87 tons/ha/year for the Progo, Opak, and Serang watersheds, respectively, as shown in Fig. 5b.High sediment yields more than 180 tons/ha/year are mostly found in the Sindoro and Sumbing Mountains area, which are located upstream of the Progo watershed on the left; the Merapi, Merbabu, and Telomoyo Mountains, which are located in the middle of the Progo watershed and upstream of the Opak watershed on the right; and the Menoreh Hills, which are located in the middle of the Progo watershed and upstream of the Serang Basin on the left.
To verify the accuracy of the RUSLE model technique-obtained framework, the sediment yield distribution was confirmed using observed data.Figure 6 represents eight total suspended solids (TSS) measurement locations that are subsequently converted to sediment discharges and linked with AWLR discharge data.Based on the power function, each location has a comparable rating curve pattern with a coefficient of determination greater than 0.80.Using the rating curve, we estimated the annual average sediment yield in tons/hectare/year.The sediment yield that was measured at the location of interest is then compared with the sediment yield that was predicted to be present at the same location based on the results of modeling with RUSLE.These results are presented in Table 7.The comparison is evaluated using R 2 , RMSE, MAE, and NSE and the results of the performance indicators are presented in Table 7.It is possible to conclude that the model has satisfactory accuracy because of the high values of R 2 and NSE (both of which are greater than 0.8) and the low values of RMSE and MAE (0.89, 0.82, and 0.14, respectively).This study validates models using the TSS and Q methods, which are transformed into sediment yields using measurement data from most of the Opak Watersheds.We advise www.nature.com/scientificreports/including measurement locations uniformly dispersed among the three watersheds.This will be highly beneficial for future sedimentation studies in the POS River Basin, which will assist in developing integrated water resource management.This must also be implemented in other watersheds in Indonesia, mainly volcanic basins, because, according to our observations, it is still challenging to gather sediment measurement data at the watershed scale, as opposed to hydro climatological data.The only watersheds where sedimentation measurements have been taken thus far are those with reservoirs.Institutions responsible for river basins, watersheds, water resources, and the environment are required to participate in the planning for sediment measurement at the watershed scale.

Vulnerability of irrigation sand traps based on sediment yield
The vulnerability of the sand trap was evaluated based on criteria for sediment yields at the inlet of the irrigation network, i.e., at a location in the river basin before entering the weir with irrigation intakes.The greater the sediment yield at the entrance of the irrigation network, the more susceptible the performance of the irrigation sand trap is to sedimentation within the irrigation network.The susceptibility of this sand trap can be employed as a consideration in the decision support system (DSS) for prioritizing irrigation sand traps that are incorporated into a river basin for further performance-related research.For example 7 , evaluated the performance of the sand trap in the Wonji Shoa Sugar Estate Irrigation Scheme (Ethiopia) 81 , simulated the sand trap in the Sapon irrigation network (Indonesia) 82 , studied how to improve the performance of the sand trap in large scale irrigation scheme (Thailand) 83 , modeled the efficiency of settling basin in small scale irrigation projects (Kenya) 9 , assessed sand traps in the Pendowo and Pijenan Irrigation Scheme (Indonesia), while 8 modeled sand trap behavior in the Pengasih Irrigation Scheme (Indonesia).These studies conclude that a sand trap's performance can be measured by capacity to hydraulically accumulate and flush sediment throughout regular flushing operations and that this will allow the sand trap to keep working correctly due to sedimentation problems in the irrigation network.Information on the current condition and performance of irrigation sand traps can be used as a reference in the management of integrated irrigation networks in a river basin, particularly in river basins that supply numerous irrigation networks.
In the POS River Basin itself, thousands of irrigation schemes draw water from the river, and these irrigation schemes can be classified into three categories of authority.The responsibility division for the region of authority is based on the criteria and areas stipulated in the Minister of Public Works and Housing Regulation of the Republic of Indonesia Number 14/PRT/M/2015, namely the existence of irrigation networks in administrative areas and strata of the irrigation network area.The responsibility division for the region of the authority of the irrigation schemes is based on the existence of the irrigation network in the strata of the irrigation network area, which includes central, provincial, and regency government authority irrigation schemes.Irrigation schemes with more than 3000 hectares, irrigation schemes across provinces or countries, or national strategic irrigation schemes are under the authority of the Central Government.Furthermore, irrigation schemes that have an area of 1000-3000 hectares or irrigation schemes across regencies or cities are under the authority of the Provincial Government.And finally, irrigation schemes with fewer than 1000 hectares and located in one regency/city area are under the authority of the Regency/City Government 8 .Based on the Ministerial regulation mentioned earlier, in the POS River Basin, there are 5, 45, and 1,235, respectively, the number of central, provincial, and regency government irrigation schemes, with a total irrigated area of 16,208, 18,081, and 37,986 hectares, respectively.
Based on the significance of their irrigated area, we only included the locations of irrigation networks under central and provincial government jurisdiction in this analysis.There are 3 irrigation schemes under the authority of the central government that fulfil the criterion of a service area greater than 3000 hectares; the remaining two are cross-provincial irrigation schemes with an insignificant service area of 149 and 115 ha.In addition, there are 9 irrigation schemes under the jurisdiction of the provincial government with a service area of between 1000 and 3000 hectares.At the same time, the remaining 36 are cross-regency irrigation schemes with a range of 1 to 541 hectares, which is negligible.On the POS River Basin sediment yield map, we displayed 12 selected irrigation schemes based on authority and service area, as demonstrated in Fig. 7a.They were represented by the weir's position, the irrigation network's main structure, or the inlet of sediment yield from the river system.Regarding Fig. 7a, the three most vulnerable sand traps of irrigation networks are Badran, Kalibawang, and Blawong, with sediment yield values of 252.82, 178.92, and 63.49 tons/ha/year at these locations, respectively.Badran and Kalibawang gain water from the Progo River, whilst Blawong receives water from the Opak river system.The  7b illustrates the sub-watersheds within the POS River Basin; we can observe that Badran is located at the Upstream Progo outlet, Kalibawang at the Middle Progo outlet, and Blawong at the Winongo outlet, or the junction of the Winongo and Opak Rivers.This implies that the massive sediment yields at all three locations are acceptable.After the prioritization is based on this sediment yield estimate, a field investigation must be conducted to determine the actual condition of sedimentation at the irrigation network's intake.This assessment should include irrigation-related institutions, operators, and water user organizations.
A greater number of stakeholders may be engaged, and irrigation system management can be conducted more efficiently.The broader perspective provided by a river basin can capture dimensions that are normally not included in the regulation of irrigation systems, such as the factors for the scarcity of water, the quality of the water, legal problems involving the usage of water, and unequal water distribution and use.From this study, it can be inferred that river basins, including volcanoes, have a more significant potential for sedimentation than plain areas.Consequently, planning for integrated water resource management (IWRM) in a volcanic river basin necessitates a distinctive strategy.The IWRM components paradigm posits two governance layers: policy-making and management environment.These layers are positioned at the national, regional, and local levels of government.The initial module includes the political context within the scope of water resource management legislation in Indonesia.The second layer illustrates the interrelationships between stakeholders and their responsibilities in the IWRM aspects, which are structured as governance levels and management pillars: Water conservation, utilization, and risk management 19 .An integrated approach to water resources management in a river basin will increase both the productivity and sustainability of natural resource usage, contributing to the accomplishment of sustainable agriculture and food security in compliance with Sustainable Development Goals 2 and 6.
In addition, this study represents an important advance in mapping the vulnerability of irrigation sand traps based on sediment yield patterns, particularly in river basins influenced by volcanism.Before this study, no other research had been conducted that addressed this specific issue, even though there are numerous volcanic river basins around the globe, 127 of which are in Indonesia.This study demonstrates that the volcanic river basin has the characteristics of a high average sediment yield, with specific hilly and mountainous areas averaging more than 180 tons/ha/year.The factor that contributes the most to the high sediment yield is the high value of the slope's length and steepness.This must be considered when formulating an IWRM strategy for volcanic river basins.This addresses the sedimentation problem in downstream water resource infrastructure, including the irrigation network.A network of irrigation systems with minimal sediment has high conveyance efficacy.This will positively affect agricultural productivity, particularly rice, Indonesia's staple diet.It will improve the welfare and social stability of the people in the study area whose primary livelihood is farming.This aligns with the first Sustainable Development Goal: to end all forms of poverty everywhere.

Conclusion
The average soil loss in the Progo-Opak-Serang (POS) Volcanic River Basin is 179.69 tons/ha/year; it falls under the category of moderate erosion potential.The average sediment yield in the POS River Basin is 51.04 tons/ha/ year.High sediment yields of more than 180 tons/hectare/year are mainly concentrated in the Sindoro, Sumbing, Merapi, Merbabu, and Telomoyo Mountains areas and the Menoreh Hills.The model has good accuracy, identified by R 2 , NSE, RMSE, and MAE values of 0.89, 0.82, 0.14, and 0.11, respectively.The significance of the sediment yield identified at the inlet of the irrigation network can be one of the criteria for assessing the vulnerability of irrigation sand traps within the POS River Basin.The greater the sediment yield at the entrance of the irrigation network, the more susceptible the performance of the irrigation sand trap is to sedimentation within the irrigation network.The susceptibility of this sand trap can be employed as a consideration in the decision support system (DSS) for prioritizing irrigation sand traps that are incorporated into a river basin for further performance-related research.Badran, Kalibawang, and Blawong are the three most vulnerable irrigation sand traps, with sediment yield values of 252.83, 178.92, and 63.49 tons/ha/year at these locations; they are all located in sub-watershed outlets.Furthermore, a field investigation must be conducted to determine the actual condition of sedimentation at the irrigation network's inlet.This assessment should include irrigation-related institutions, operators, and water user organizations.A more significant number of stakeholders may be engaged, and irrigation system management can be conducted more efficiently.Furthermore, planning for integrated water resource management (IWRM) in a volcanic basin necessitates a specific strategy because it is more susceptible to sedimentation than plain regions.

Figure 2 .
Figure2.The study area: POS Volcanic River Basin, Indonesia.It is a basin comprised of three major watersheds: Progo, Opak, and Serang.This river basin contains Mount Merapi, one of the world's most active volcanoes, as well as other volcanoes, and is therefore known as the volcanic river basin46 .

Figure 4 .
Figure 4.The spatial distribution of components and factors used in RUSLE within the POS River Basin.(a): The distribution of overall and consistent rainfall stations; no specific trend; (b): Rainfall-runoff erosivity factor (R) with an average value of 1656.66MJ•mm/ha/hour/year; ©: Six distinct soil types, including alluvial, grumusol, latosol, litosol, regosol, and yellow-red latosol, with proportions of 7%, 8%, 36%, 25%, 24%, and 0.0001%; (d): K factor with high range of 0.40-0.56tons•hour/MJ/mm is distributed in the middle and downstream parts of the Progo Watershed and upstream Opak; (e): five slope classes, including such as moderate (8-15%), moderately steep (16-25%), and steep (26-40%); (f): LS factor with an average value for the entire river basin is 2.72; (g): LULC map indicates that 73.61% of the POS River Basin is dense forest; (h): C factor with an average value is 0.022; (i): P factor spans from 0.2 to 131.8, with an average of 5.1; (j): The spatial pattern of the estimated average soil loss with an average of 179.69 tons/ha/year in the POS River Basin 46 .

Figure 5 .Figure 6 .
Figure 5.The spatial variability of SDR and sediment yield throughout the POS River Basin.(a): The distribution of SDR, with an average value of 0.28 within the river basin; (b): The distribution of sediment yield.The average sediment yield in the POS River Basin is 51.04 tons/ha/year, with values of 71.75, 50.23, and 27.87 tons/ha/year for the Progo, Opak, and Serang watersheds, respectively 46 .

Figure 7 .
Figure 7.The spatial distribution of selected weirs on sediment yield map and sub-watersheds delineation.(a): On the sediment yield map, 12 weirs (the primary structure of irrigation networks) represent the location of irrigation networks.Based on the amount of sediment at the inlet, Badran, Kalibawang, and Blawong are the three most vulnerable sand traps in irrigation networks; (b): The location of weirs or irrigation networks inside sub-watersheds.Badran, Kalibawang, and Blawong are in the outlets of sub-watersheds 46 .

Table 2 .
K rate (tons•hour/MJ/mm) for several soil types on Java Island, Indonesia.

Table 3 .
Length and steepness of slope factor.

Table 4 .
C factor value for land use and land cover classes.

use and land cover C factor value Land use and land cover C factor value
well as in the upstream regions of the Opak watershed, high R values ranging from 1700 to 2600 MJ•mm/ha/ hour/year can be found.This is related to the high monthly rainfall input in the region, which includes Mount Merapi, Mount Merbabu, Mount Sindoro, Mount Sumbing, Mount Telomoyo, and Menoreh Hills.

Table 2 ,
67 assigned the K quantities to every type of soil, and the spatial distribution of K values is demonstrated in Fig.4d.High K concentrations of 0.40-0.56tons•hour/MJ/mmareobserved in the central and downstream portions of the Progo Basin, as well as upstream Opak.This refers to the regosol, alluvial, and yellow-red latosol soils with reference K values of 0.4, 0.47, and 0.56 tons•hour/MJ/mm, respectively.Regosols have weak structural cohesion and are common in eroding soils, especially in arid climate zones and mountainous regions65.Alluvial deposit, also known as alluvium, is loose clay, silt, sand, or gravel accumulated by moving water in a river bottom, in a floodplain, on an alluvial fan or seashore, or in other comparable environments66.This kind of soil is highly susceptible to erosion.Furthermore, there are seven classes of soil erodibility in Indonesia, namely very low (0.10 tons•hour/MJ/ mm), low (0.10-0.15 tons•hour/MJ/mm), moderately low (0.15-0.20 tons•hour/MJ/mm), moderate (0.20-0.25 tons•hour/MJ/mm), moderately high (0.25-0.30tons•hour/MJ/mm), high (0.30-0.35 tons•hour/MJ/mm), and very high (> 0.35 tons•hour/MJ/mm)67.The average values for soil erodibility in the Progo, Opak, and Serang watersheds are 0.34, 0.30, and 0.28 tons•hour/MJ/mm, respectively, while the average value in one river basin is 0.31 tons•hour/MJ/mm.Consequently, the POS river region falls into the category of highly erodible soils that have the potential to erode at a pace that is significantly larger than the ratio of soil loss considered tolerable.

Table 5 .
Soil Erosion Dispersion in the POS River Basin.

Table 6 .
Correlation coefficient of RUSLE factors used to calculate the soil erosion in the POS River Basin.

Table 7 .
Comparing observed sediment yield to expected sediment yield.