Development of Decision Support System (DSS) for surface runoff and erosion control structures planning: a case study in the Upper Citarum Watershed & towards sustainability inland waters ecosystem

The watershed ecosystem is a complex system in which there are numerous transportation and transfer mechanisms of mass and energy. To make management more effective in all national critical watersheds, there needs to be more scientific, evidence-based policymaking that is based on an understanding of the system and mechanisms of the socio-hydrological processes of the watershed. The availability of Decision Support System (DSS) technology can be the appropriate approach to this need because DSS can be an interface between scientific and practical needs (easy-to-use, easy-to-access, user-friendly). However, the availability and implementation of DSS as an important tool in the optimal design of sustainable watershed management in Indonesia are still very limited. This study aims to create a prototype DSS that practitioners and policymakers can use to identify priority areas and optimize technical solutions for controlling surface runoff and soil erosion at various scales in the internal watershed. Herein, the spatial-based numerical modeling system and process mechanism; the database and knowledge; and the Graphical User Interface (GUI) are the three main components that have been used as a framework for model-based DSS development. A distributed rainfall-runoff-erosion model (EcoHydro) is the main engine of DSS for spatially quantifying dimensions of hydrological responses, erosion rate, and sediment production according to the user’s specifications and providing design options for control measures of them. The upper Citarum river basin in West Java, which is a critical and first national priority watershed, has been selected as the case study for DSS development and to demonstrate its application. Furthermore, the resulting DSS can later be developed for use in watersheds and other nationally critical lakes.


Introduction
There are currently at least 2,145 watersheds in Indonesia that need to be restored due to the degradation of their ecosystems, out of a total of 17,000 watersheds.According to the findings of the study [1], there were 22 watersheds in critical condition in 1984, 39 watersheds in 1994, and 62 watersheds in 1999.The condition is deteriorating as a result of increasing human intervention, which is reducing the environment's support and carrying capacity of the watershed ecosystem.Currently, the government estimates that as many as 108 watersheds in Indonesia are in critical condition, with 15 of them needing to be addressed on a priority scale as stated in the National Medium Term Development Plan (RPJMN) 2014-2019.If not immediately addressed, this critical condition will result in a greater disruption of the four functions that make up the watershed ecosystem, including (1) water availability (quantity and quality); (2) biodiversity function (diversity of habitats and species); (3) ecosystem service functions (including fisheries, agriculture, and hydropower); and (4) resilience to disturbances and disasters.
Since a long time ago, Indonesia has implemented a number of programs and technical measures to restore the above four dimensions of watershed function; however, these management efforts have thus far run into a number of challenges [2].One of the main causes is the inadequacy of a comprehensive planning system [3,4,5].A watershed is an integrated ecosystem that contains numerous cycles and transfers of material (water, soil, contaminants), as well as energy.Therefore, more scientifically evidence-based policymaking that is based on an understanding of the systems and mechanisms of watershed socio-hydrological processes needs to be implemented in order to improve the effectiveness of management in all national critical watersheds [6,7].There has been a lot of technical research done on managing watersheds, conserving soil and water, and mitigating disasters.However, those who practice watershed management have not used science and knowledge as the basis of policymaking [8].One of the causes is the technical knowledge barrier that prevents practitioners from using modeling tools for the analysis and design of watershed management.
Because Decision Support System (DSS) technology can serve as a bridge between the demands of practitioners and scientists, it may be the ideal solution to the needs outlined above.Using information from data, documents, knowledge, and simulation models, DSS is an interactive information system that helps decision-makers (planners) find issues, resolve issues, and make decisions.To meet practical needs, such as being simple to use, simple to access, and user-friendly, a DSS will integrate a complex modeling system with information technology [9].Operational activities for watershed management have made extensive use of DSS on a variety of platforms [10,9,11,12,13].However, there are still very few facts in Indonesia where DSS is available and implemented as a useful tool in the best management strategy design in order to restore the watershed function [14].
On the other hand, land demands and changes have been resulting in a decrease in the areas where rainwater is recharged into soil, which happens in almost all Indonesian watersheds.Thus, the high rates of surface runoff, soil erosion, and the transport of sediment material to water bodies are the main issues currently plaguing Indonesia's 108 national priority watersheds.As an example, the upper Citarum River Basin in West Java has erosion and sedimentation issues.The Upper Citarum's outlet, the Saguling Reservoir, had a sedimentation rate of 4.2 million m 3 /year, according to the Ministry of Public Works & Houses' (2012) analysis of bathymetric data from 1985 to 2012.Given that the Saguling Reservoir's expected sedimentation rate was only 1.5 million m 3 /year, this situation threatens to reduce the reservoir's operational life time from approximately 183 years to 34 years.In addition to endangering the reservoir's function, soil erosion can also make it more likely to flood by making water bodies shallower and decrease agricultural yields by making the top soil layer thinner, which results in nutrient leaching for plants.Therefore, innovations are urgently required to reduce the rate of soil erosion and sediment transport in the upper area of the Citarum watershed.
The goal of this research was to create a prototype DSS that practitioners and policymakers could use to identify hotspot areas and optimize technical solutions to manage surface runoff, soil erosion, and sediment transportation at different scales within the internal watershed.The three main elements that have been used as a framework in the current modeling system-based DSS development are spatiallybased numerical modeling system (physically-based distributed model type), database and knowledge, and Graphical User Interface (GUI).

Methodology
A Decision Support System (DSS) is a component of a computerized information system and knowledge (management knowledge) used to support decision-making.A DSS is made up of three main parts in general (figure 2 left), namely a database and knowledge, a model management system that serves as a Decision Support Tool (DST), and information technology (graphical user interface, or GUI).The three of them serve as general framework models for DSS package development, allowing the DSS to be simple to use, simple to access, and user-friendly when it is made available to the general public (figure 2 right).
The management system model component quantifies the variables of the hydrological responses that occur based on the chosen scale (slope unit, catchment, basin, watershed, or multiple watersheds) by processing all available data, information, knowledge, and control strategy scenario options.The primary DSS engine for this study is hydrological modeling, which serves as the management system model and can yield information that can be used as a decision-making reference.In addition, the modeling system is a DST component utilized in the DSS.The DST type (platform) that was used to meet the DSS requirement is a physically based distributed model type or a spatially based numerical modeling system with process mechanism formulation.One type of the author's research output [16,17,18,19] is a hydrological modeling system known as EcoHydro, which has been advanced into the primary DST of the obtained DSS.Originally, EcoHydro was a numerical modeling system for assessing fluxes (productions), rates (loads), or changes within one or more components of the hydrological cycle, surface soil erosion, and sediment transportation at various scales (from the grid to the scale of multiple watersheds).
Figure 3.The generic process of system development and decision-making flow in the DSS package that will be built by using the DSS structure created by Giupponi and Sgobbi [11] and adding a new concept of management strategy called the ecohydrological approach.
The EcoHydro DST has been intended to be further developed so that currently it has four constituent sub-models, namely: 1. the hydrological sub-model; 2. the soil and sediment erosion sub-model; 3. the contaminant sub-model; and 4. the sub-model optimization of various control measures and management strategies.This will result in a DSS package for controlling surface runoff flows, soil erosion rate, and sedimentation.Two sources of direction and scientific reference were used in develop the structure and formulation of the four sub-models of EcoHydro DST: (1) bottom-up approach of local participatory planning and (2) ecohydrology concept [20,21,22].Ecohydrology is a new global concept in water resource management that takes a more comprehensive approach by comprehending and quantifying the synergistic interactions that occur between the hydrological cycle, ecological dynamics, and social dynamics of communities within a watershed.The diagram in figure 3 illustrates the existence and role of these two approaches in the development of EcoHydro DST and the DSS components as a whole.
In general, the mechanism of the hydrological cycle, erosion and transport of sediment material is formulated and modeled within DST in a distributional (spatial) manner which includes two main elements of the watershed system, namely the catchment area (land/slope) sub-system and the river section sub-system.Figure 4 shows a general concept of the EcoHydro DST model structure.The current DST version was built with the Fortran programming language, GNU Plot, and GIS Software.Its main data inputs include: topography, soil characteristics, land use, climate/meteorology, river discharge observation data, river sediment concentrations, administrative data, and control structures for runoff and soil erosion.

Results & Discussion
Surface runoff, soil erosion rate, and sediment yield information are produced spatially and temporally by the DSS output through the EcoHydro DST, which has sub-models of surface runoff generation processes, soil detachments, and sediment transports (rainfall-runoff-erosion). The model output is a text file that can be converted into tables, graphs, or maps in GNU Plot, Excel or ArcGIS after being processed there.The DSS package presents temporal data in hourly to monthly time intervals and spatial data based on a specific grid size.Surface flow, erosion, and sediment rates, however, are computed on a hourly basis and then added up or averaged over the course of a month or year.Evaluating EcoHydro's DST output to reflect the observed value of the hydrological response in the Upper Citarum watershed is the first step in the DSS performance testing process.The Upper Citarum   Figures 5 and 6 show the periods of calibration and validation results from the application of modeling system-based DSS in representing the observed cumulative daily water and sediment yields at the study area.The NSE values for the cumulative daily discharge in 2018 and 2019 were 0.71 and 0.95, respectively, while the NSE values for the cumulative daily sediment in 2018 and 2019 were 0.96 and 0.79.This value is excellent for a model's ability to accurately represent the actual condition.Since the distributed rainfall-runoff-sediment model has been verified and calibrated, and it displays good NSE values, it is possible to predict erosion and sediment production in addition to hydrological conditions using the EcoHydro DST.
Because no soil particles are released as a result of surface flows and rain splashes, there is little sediment production during the dry season compared to high production during the rainy season.It is estimated that sediment accumulation through Nanjung Station will be 6,980,402.5974tons/year in 2018 and 6,158,238.2771tons/year in 2019.Six million tons of sediment are produced annually, and most of that material may end up in the Saguling Reservoir.Additionally, figure 7 depicts the spatial information on the potential sources of erosion and deposition over the target area obtained from the developed DSS.Quantitative information on the distribution of annual erosion and deposition rates can be used as a reference for priority plans for handling erosion based on administrative areas.The process of developing the decision support system DSS for watershed management, especially for land degradation mitigation and controlling river water pollution, is still ongoing, and 2024 is the target year to finalize the system.The formulation of DSS has been intended to help decision-makers in strategic planning and operation control through four ultimate functions, as below: x determining priority locations for handling watershed degradation and pollution control spatially, equipped with the level of hazard, damage, or risk of losses; x To figure out what the problem is by looking at the quantitative (spatial and temporal) data that comes from simulating how the watershed system's key components interact with each other; x tool for designing and evaluating the effectiveness of various strategic scenarios for controlling both issues-watershed land degradation and river pollution; x optimizing control strategies from the aspects of type, dimension, costs, location, and benefit

Conclusion
The DSS package's main component, EcoHydro DST (rainfall-runoff-erosion modeling system), performs well in simulating observed river flow and sediment concentration, according to the results of its development and preliminary application.It also has a great deal of potential for use in determining the spatiotemporal source areas of soil erosion and deposition within the watershed as well as in determining how much sediment is produced at a specific location.The resulting prototype DSS can be used as a tool for scenario-based management plans, such as designing control strategies for improving water availability and quality and reducing surface runoff, erosion rates, and sediment production at certain identified priority locations or sub-watershed.Quantitative assessment (reconstruction, prediction, and projection), understanding the hydrological process mechanisms, and scenario-based management plans can all be done with the current developed prototype DSS.

1 .
Upper Citarum WatershedThe Indonesian government has determined 15 national priority critical watersheds that must be restored, namely: Citarum, Ciliwung, Cisadane, Serayu, Bengawan Solo, Brantas, Asahan Toba, Siak, Musi, Way Sekampung, Way Seputih, Moyo, Kapuas, Jeneberang, and Saddang.Located in West Java Province, the Citarum Watershed has the highest priority for restoration targets.This has been accomplished through the implementation of Presidential Regulation No. 15 of 2018 relating to the Program Percepatan Pengendalian Pencemaran & Kerusakan DAS (P3K) for the 2018-2025 period, which controls river pollution and watershed degradation.With the goal of making a DSS model that can contribute to P3K, DSS development and practical use have taken place in the Upper Citarum watershed, which includes the area up to the Saguling Reservoir as the upper Citarum River outlet.

Figure 1
depicts the entire Citarum watershed (682,227 ha) on the left and the upstream sub-watershed division areas on the right.The study conducted by Cahyaningsih & Harsoyo [15] is a reference for site selection in the upper Citarum watershed (231,124 ha).It stated that the level of river pollution and the area of critical land are the highest in this region.

Figure 1 .
Figure 1.The separation of the sub-watersheds in the study area (right) and the boundaries of the Citarum watershed (left).

Figure 2 .
Figure 2. The DSS technology package's framework algorithm (right) and the three main components that make up the DSS package's structure (left).

Figure 4 .
Figure 4. General description of the DST structure of the distributed rainfall-runoff-erosion model (EcoHydro).

6
River outlet at Nanjung Station collected daily observation data on sediment concentration and river water flow discharges.These data in 2018 and 2019 were used to calibrate and validate the model output.In order for the model output to accurately reflect the observed values, the calibration stage involves tuning up model parameter values either manually or automatically.Next, what is referred to as the "model output validation stage" involves applying well-calibrated model parameter values again at different time periods.Sixteen (16) parameter models pertaining to topography, land use, and soil components are included in the calibration process of the EcoHydro DST for the Upper Citarum watershed model.Utilizing the Nash-Sutcliffe model Efficiency coefficient (NSE), the output performance of DST EcoHydro was evaluated during the calibration and validation process.

Figure 5 .
Figure 5.Comparison cumulative daily water yields at the river outlet of upper Citarum Watershed in 2018-2019 between model-based DSS and observed.

Figure 6 .
Figure 6.Comparison cumulative daily sediment yields at the river outlet of upper Citarum Watershed in 2018-2019 between model-based DSS and observed.

Figure 7 .
Figure 7. Spatial information of DSS for the potential sources of erosion and deposition in the upper Citarum Watershed.