Spatial planning concept for flood prevention in the Kedurus River watershed

1.1 Study area

Surabaya is the second-largest city in Indonesia [14], and a map of potential flood disasters in the RTRW document for Surabaya divided the city into three locations, namely, the East Coast of Surabaya, the North Coast of Surabaya, and the Kedurus Watershed [15]. The flood that occurred in the Kedurus watershed was caused by the overflow of the Kedurus River and changes in land use. Therefore, it is necessary to interfere to reduce the potential flood hazard. The worst floods in the city of Surabaya occurred in the Wiyung area, which was up to 1 m [16], more than 100 ha was inundated by the Kedurus River flood. The 2018 Surabaya Drainage Management Plan data show that several areas of Wiyung often experience flooding, with a flood duration of 1 h and an average height of 0.5 m.

Floods during the rainy season in the Wiyung area occur due to the impact of land conversion. This is backed up by the fact that the Wiyung area has a trend of changing the use of agricultural land into a settlement based on the RDTRK UP Wiyung data. As for the data on the RTRW of Surabaya for 2014–2034, the Wiyung Development Unit has the main function for settlements, especially the development of medium-density housing and settlements, education, industry, and protection of nature [17,18]. Apart from these reasons, the Kedurus watershed was also chosen due to the uniqueness of the urban flood area. The aim is to increase the regional economic target in the Kedurus watershed while maintaining its ability to anticipate floods.

In terms of Wiyung’s RDTRK strategic plan, there are several development plans that will have an impact on increasing the chance of flood itself. In terms of infrastructure in the Kedurus watershed in UP Wiyung, the existence of the Middle West Ring Road network will increase settlement activities and business activities around it which will reduce the function of water infiltration into the ground. In addition, when viewed from the spatial planning objective of UP Wiyung’s RDTRK, which is investment friendly, it will make it easy for housing developers to establish housing clusters that will reduce the function of water absorption in the Kedurus watershed.

This is supported by the condition of the existing spatial pattern, namely, formal housing in the Kedurus watershed in UP Wiyung. This also applies to the RDTRK UP spatial plan. Wiyung will have a medium-density settlement space allocation plan (R3) around the Kedurus watershed. Due to the potential for flooding in the Kedurus watershed, which has an impact on the disruption of urban activities, it is necessary to take steps to control the flooding in the Kedurus watershed.

In spatial planning, there are various instruments that can be used to control flooding. There are three spatial planning instruments, namely, land use regulation, building regulation, and drainage systems. From previous studies [19,20,21,22], these three instruments were declared effective in reducing flooding. Where related to the study of flood reduction in the Kedurus watershed through land use regulation, it can reduce the flood discharge up to 12.92% of the flood discharge in existing land uses. In addition, the previous study [23] discussed the related building regulation, specifically setting the basic building coefficient (maximum percentage within a land parcel that can be built [KDB]) and green basic coefficient (minimum percentage of open space within a land parcel [KDH]). The results obtained by adding 10% KDH in residential areas will reduce the flood discharge by 1.4%. Arrangement of the drainage system can be done by providing a drainage network and retention pond. From previous studies [24,25,26], it is found that a good drainage system can reduce floods by 58.79%.

Empirical facts in the field of flood control in Surabaya are still inaccurate considering that, currently, the area in the Kedurus watershed has exceeded 50% of the watershed area and so efforts to add drainage channels and increase river capacity are difficult to carry out [27]. Similar research was conducted [16] related to reducing floods in the Kedurus watershed by using land use regulation instruments and building codes, which were analyzed by using the soil and water assessment tool model. The results of the research were that all areas of the Kedurus watershed were divided into 27 sub-watersheds that experienced a flood, namely, sub-watershed 1, 2, 3, 5, 6, 8, 12, 13, 15, 22, 24, and 27 with a total overflow volume of 546797.53 m³. The combination of land use regulations and building codes can reduce flood discharge by 47.51% [28].

Based on the description of the aforementioned problem, the aim of this study is to formulate a spatial planning concept in the form of land use regulation, building codes, and drainage systems for flood reduction in the Kedurus River watershed. The scope of the area in this study is the flood area identified by [29] in the Kedurus watershed. The research area used is the sub-watershed in the administrative area of Surabaya, and the justification is the availability of data on the surface model digitization, with an accuracy of 0.33 m × 0.33 m. The research area is the Lakarsantri District. The administrative boundaries of the research area are Lakarsantri District to the north, Wiyung District to the east, Karangpilang District to the south, and Menganti District to the west (Figure 1).

Figure 1

Research map area.

2.1 Building a spatial flood model based on existing hydrological conditions that cause floods in the Kedurus watershed

This target is divided into several stages in the process of developing a spatial flood model [26] in the Kedurus watershed, namely:

1. Identification of the peak discharge (Q _p) in the Kedurus watershed. At this stage, it is carried out to determine the maximum load volume of rain that flows into the channel by using the unit synthesis hydrograph calculation to convert the rainfall in the study area to peak discharge. In identifying the peak discharge, it is necessary to determine the value of the flow coefficient (runoff), which is carried out using the Cook flow coefficient method. This is done by inputting a map of the level of soil infiltration capacity, flow density, vegetation cover, and slope [30] and furthermore by analyzing the frequency of potential extreme rainfall with the formula of the Gumbel distribution method [31] and converting the daily rainfall into the effective rain per hour using the Ø Index method and the Horton method. Then, the hydrograph of the Snyder synthesis unit is calculated and the superposition hydrograph process is carried out, which produces a superposition hydrograph at the outlet of the Kedurus watershed segment, which is the outlet of all sub-watersheds in the upstream watershed segment. The outlet will be defined as an inlet in hydraulic modeling. This process is carried out in the upstream segment whose outlet is adjacent to the downstream segment inlet, which was affected by the flood, namely, the Kedurus sub-watershed. These discharge data then become the input for hydraulic modeling in the HEC-RAS software. The area of the sub-watershed that was affected by the flooding based on the last flood event in 2017 and which was the delineation of the analysis area in HEC-RAS was 70 km².

2. The geometry correction process of Kedurus Main River on Digital Terrain Model Lidar Surabaya is performed. Digital Terrain Model (DTM) data from Lidar Surabaya with a resolution of 0.33 m × 0.33 m are processed into variable topographic conditions as input for hydraulic modeling in the HEC-RAS and also as a watershed characteristic variable in processing discharge data.

3. A map of the distribution of the Manning roughness coefficient in the Kedurus watershed is created. At this stage, a global reference is used to assess the Manning roughness coefficient in the study area [32,33]: the Manning roughness coefficient is not directly used, but it is combined with the green baseline coefficient.

4. Spatial modeling of floods in the Kedurus watershed. At this stage, the data obtained from steps 1–3 are used as data input. These data through HEC-GeoRAS will be packaged into a file in RAS format, as input for geometric data in HEC-RAS [32]. The hydraulic analysis was performed using the unsteady flow method. This stage produces a flow profile, which is then converted through the HEC-GeoRAS extension into spatial information in the form of the distribution of the affected location, the extent, and the depth of the flood [33].

2.2 Evaluating land use regulations, building codes, and existing drainage systems to reduce flooding in the Kedurus watershed

In this section, planning evaluations related to land use arrangements, building codes, and drainage systems are carried out by adjusting the conditions of the plans and referred documents. The planning document referred to for regulating land use is the Detailed Spatial Plan for the Wiyung Development Unit with the data of the Public Housing Service for Settlement Areas and the Surabaya City Spatial Planning Cipta Karya. Meanwhile, the drainage system uses the Surabaya Drainage Management Plant document with the data of the Public Works Office of Bina Marga. This target is divided into several stages:

1. Identification of the peak discharge due to the implementation of the land use plan and the basic building coefficient plan: This stage is carried out to determine the discharge load, which is caused by the implementation of the land use plan and the planning of the basic building coefficient. First, it is necessary to determine the runoff value due to the application of the land use plan and the basic building coefficient (KDB) plan using the Cook flow coefficient approach, in which the vegetation cover parameters are adjusted to the land use plan and the basic building plan. The reference value of the runoff coefficient for each class of land use plans is derived from ref. [36]. The runoff value of each land use plan classification is then multiplied by the planned basic building coefficient per land use plan block. Furthermore, the conversion of daily rainfall to effective rain per hour is carried out by running off the plan and carrying out the superposition hydrograph process with a runoff plan.

2. Geometry correction of the main Kedurus River based on the River Cross Section Plan in the Surabaya Drainage Management Plan: At this stage, the River Cross Section Plan for SDMP 2018 will be applied to the Model Lidar Surabaya Digital Terrain data 0.33 m × 0.33 m. The SDMP application in DTM uses the Cross-Section Editor in the HEC-RAS 5.0 Software. The value in the section that will be changed is in accordance with the value in the 2018 SDMP.

3. Developing a manning roughness coefficient distribution map based on the Land Use Plan and the Basic Building Coefficient Plan: At this stage, the manning roughness coefficient will be created based on the land use plan data and the green baseline coefficient. The value will then be multiplied by the green base’s coefficient because the larger the non-built land, the greater the green base’s coefficient value.

4. Spatial Modeling of Kedurus Watershed Flood for Evaluation of Land Use Plans, Basic Building Coefficient Plans, and Drainage Channel Plans: At this stage, the HEC-RAS 5.0 software is still using the same as target 1. The difference in settings is only in the value of the hydrograph graph according to the Run-Off Plan, the Terrain model uses the results of the DTM correction with SDMP 2018, and the manning coefficient map uses the land use plan data and the basic building coefficient.

2.3 Formulation of the concept of spatial planning to reduce flooding in the Kedurus watershed

In this section, the conceptual formulation is carried out using sensitivity analysis tools to assess the significance of variable changes in the area of flood inundation. From the sensitivity analysis, it is known that the priority variables are significant in reducing the area of flood inundation. Then, the formulation is carried out in the form of scenarios 1 and 2. Sensitivity analysis has several stages:

1. Sensitivity 1: Changes in Land Use Variables. Sensitivity testing of land use intensity parameters was carried out by increasing the agricultural area and open area by 15% from current conditions. This is because agriculture and open areas have the least coefficient of runoff and are optimal for absorbing water into the soil

2. Sensitivity 2: Basic Building Coefficient. Testing of the basic building coefficient parameters is carried out by lowering the KDB for Developed Land by 15% from the current condition. The basic building coefficient is decreased from KDB 50–100%. This is because KDB below 40 is a nonbuilt area with dominance in open areas.

3. Sensitivity 3: Drainage Channel Capacity. Testing of the drainage channel capacity parameters is carried out by deepening the main Kedurus drainage channel by 15% from its existing depth. The technique used is to change the Lidar Surabaya DTM data from 0.33 × 0.33 m with XS Interpolation with the help of RAS Mapper in HEC-RAS 5. The main Kedurus drainage channel will be divided into 31 cross-sections to identify the depth of the existing channel. Then it will be set by 15% of the existing condition.

4. Sensitivity 4: Combination of Land Use Intensity, Building Base Coefficient, and Drainage Capacity. In this sensitivity test, a combination of parameters of land use intensity, basic building efficiency, and channel capacity will be carried out. The technique is used to combine data on sensitivity 1, 2, and 3.

5. After knowing the impact of changing each parameter on reducing the area of flood inundation. Then, it is prioritized based on how sensitive parameter changes are to flood reduction.

3.1 Building a spatial flood model based on existing hydrological conditions that cause floods in the Kedurus watershed

In the analysis of target 1, a flood map with five time series is generated: the flood on November 25, 2017, 2-year return period, 5-year return period, 10-year return period, and 20-year return period (Figure 2). The results of the spatial model of the flood are in the form of a.TIF raster file that describes the results of the model that has not been filtered/cut in areas that are waters (rivers and other water areas).

Figure 2

Comparison of spatial flood modeling results in Kedurus watershed.

The level of accuracy of the model to the actual flood conditions is measured by the value of root mean square error (RMSE) and (LE90 Linear Error 90%), which have become the standard in the vertical accuracy test of spatial data [34,35,36,37]. The results of the flood model were validated by testing the RMSE flood model on November 25, which had a value of 0.15, which means that on average the model had an error rate in describing the actual flood conditions of 0.046 m. Meanwhile, the calculation of LE09 shows that the value of the model is 0.24, which means that the maximum difference between the flood model and the actual condition of the flood is 0.24 m. From the RMSE value of 0.15 below the value of errors 1 [35,36], the model is declared valid. When compared with a similar study from ref. [37], which modeled flooding with the same method, the RMSE value was 0.076, and then the flood model in this study is better. Because the LE09 value is less than the value of 1, the model is categorized as quality model 1 [38]. From the spatial flood model, it is known that the flood area covers 8.27% of the research area or an area of 228.88 ha in the most extreme conditions during the 20-year return period and from the worst-case scenario, and the peak discharge reaches 20.90 m³/s, which is assumed to be the maximum load that must be accommodated by the Kedurus watershed.

3.2 Evaluating land use regulations, building codes, and existing drainage systems to reduce flood in the Kedurus watershed

In the analysis of target 2, the Kedurus watershed runoff coefficient is generated based on the land use plan and the basic building coefficient plan. As a result of the implementation of the land use plan and the KDB plan in the Kedurus watershed, the Run-Off value of 0.38 is obtained, meaning that of 100 mm of rain in a day, 38 mm could not be infiltrated into the ground and had to be accommodated in drainage channels. The runoff parameter of the plan increased by 0.2 points from the existing runoff parameter. Thus, RDTRK in the Kedurus watershed does not decrease the Run-Off value and even increases the Run-Off value, which has the potential to increase flooding. The peak discharge in the planned condition also increases compared to the peak discharge in the existing condition.

The results of the spatial flood modeling in the Kedurus watershed using the parameters of the land use plan, KDB plan, and Kedurus Drainage Plan (Figure 2) show that the existing land use plans, KDB plans, and drainage plans that currently exist do not reduce the potential for flooding, and in fact, they tend to increase the potential for flooding, in slightly.

3.3 Formulating the concept of spatial planning to reduce flooding in the Kedurus watershed

The resulting spatial layout concept for flood reduction in the Kedurus watershed is a combination of two scenarios (Table 1). Scenario 1 is a concept consisting of changes in land use variables, basic building coefficients, and channel capacity based on the significant value and the possibility to do so. Scenario 2 is a concept that consists of optimizing all variables and using green infrastructure components to reduce flooding by 100%. With the sensitivity analysis and the concept of reducing inundation area, the application of the spatial planning concept for reducing flooding in the Kedurus watershed still leaves a flood inundation area of 6.26 ha from 80.27 or has decreased by 97.26% from the existing conditions. In addition, the results obtained by changing the basic coefficient of buildings have the greatest impact in reducing the flood inundation area by 42.29%. Then in second place is changing the intensity of land use, which has an impact on reducing the inundation area by 36.77%. Finally, changing the capacity of the drainage channel reduces the inundation area by 16.00%. Combining these three parameters will reduce the inundation area by 69.38%. From the results of the process of formulating the spatial planning concept for flood reduction in the Kedurus watershed, the overall conclusion is presented in Table 1.

3.4 Discussion

The results of the process of formulating the spatial planning concept for flood reduction in the Kedurus watershed (Table 1) were obtained from the planned changes in each parameter. These changes use 15% incremental increase/decrease as follows:

1. In the parameter of land use intensity, the changes that must be made are increasing the agricultural area and open area by 15% from current conditions.

2. In the basic building coefficient parameter, the changes that must be made are reducing the KDB of built land by 15% from the current condition.

3. In the drainage channel capacity parameter, changes that must be made are increasing the capacity of the drainage channel by 15% from the current condition.

4. In the parameters of land use and the basic coefficient of buildings and drainage channels, the changes that must be made increase the agricultural area and open areas by 15% from current conditions and reduce the KDB of built land by 15% from current conditions and increase the capacity of the drainage channels by 15% of the current state.

5. In the concept of reducing the area of flood inundation in scenario 1, the changes that must be made are adjusting land use and basic building coefficients as well as channel capacity based on possibilities.

6. In the concept of reducing the area of flood inundation in scenario 2, the changes that must be made are adding retention ponds.

Overall, the spatial planning concept for reducing flooding in the Kedurus watershed is divided into four variables: land use (Table 2 and Figure 3), basic building coefficient (Table 3 and Figure 4), channel capacity, and use of the retention pond.

Figure 3

Land use recommendation for flood reduction.

Figure 4

Building coefficient recommendation for flood reduction.

To illustrate that the recommendations for engineering actions on land use variables, basic building coefficients, and channel capacity can be implemented in the field. The implementation level of each recommendation was determined using a series of calculation to determine the suggested value, which are the minimum or maximum values to produce a significant flood reduction. The recommendations for flood reduction in the Kedurus Watershed using those values are as follows:

1. The land of Surabaya City Government assets is turned into a green open space (green open spaces which are managed by the government [RTH]).

2. Space allocation for protected area zones follows the spatial plan for the Surabaya City RDTRK spatial plan.

3. Spatial allocation for cultivation in the sub-zone of settlement, trade, public service, and industry, are allocated on existing non-built land, with maximum 60% percentage of the land, and must have infiltration wells.

4. Decreasing the basic building coefficient by 5% in various industries, utility installations, government offices, public facilities, defense, and security.

5. A 10% decrease in the basic building coefficient for trade and services.

6. Decrease the base building coefficient by 15% for high-density housing, medium-density housing, and high-density housing.

7. Allocation of the basic building coefficient is a maximum of 60% on the allocation of cultivation that converts the existing undeveloped land.

8. Increase the capacity of the Kedurus drainage channel by deepening it by 25% from the existing condition to a maximum of 4.3 m or according to the 2018 SDMP.

9. The construction of two new retention ponds covering an area of 22.98 and 8.53 ha.

10. Expansion of the existing Kedurus retention pond covering an area of 4.9 and 3.8 ha.

The main result obtained from this research is a spatial modeling of the flood with the existing conditions, resulting in a flood inundation area of 228.88 ha. The model was validated by RMSE with a value of 0.046 and declared excellent. Evaluation of the planned conditions of the 2018 Surabaya City RDTRK and 2018 Surabaya Management Drainage Plan with a model that has been built resulting in a flood area of 226.41 ha. The implementation of RDTRK and SDMP in the City of Surabaya tends not to accommodate the reduction in the area of flood inundation. The stages of the formulation of the spatial planning concept for flood reduction resulted in several proposals, namely, the allocation of green open space for all the land assets of the Surabaya City government and the provision of a maximum requirement of 60% development for space allocation on the existing undeveloped land. In addition, it is proposed to reduce the maximum KDB from 5 to 15% according to land use. It is also proposed to increase the channel capacity by deepening 25% of the existing condition. Then, a flood reduction instrument in the form of a detention pond is involved. The implementation of these recommendations has the potential to reduce the area of flood inundation by 222.62 ha or 97.26%.