Rainfall trend analysis to the change of flood discharge estimation in Bengawan Solo River

Bengawan Solo Watershed is the largest watershed on Java Island, which also has fluctuating rainfall. The fluctuating rainfall may cause an alteration to the river discharge. The alteration can be analyzed through the SCS-CN model and Synder Unit Hydrograph. SCS-CN Java Island, which also has fluctuating rainfalis a well-known model in hydrological modeling that is commonly used because of its simplicity, stability, and predictability. Synder Unit Hydrograph is suitable for ungauged basins and may be used in watersheds about 25 km2 to 25.000 km2 (Sudhakar BS et al, 2015). The analysis and calibration were performed from the baseline data (2001-2022). The trend analysis was carried out by comparing the analysis results from data sequences with a ten-year range with the baseline data. According to the trend analysis, the rainfall design of the Bengawan Solo watershed has increased by approximately 4–6% in 2013–2022 compared to the baseline (2001-2022), which may lead to a 10% increase in flood design discharge in Bengawan Solo.


Introduction
Bengawan Solo watershed is the largest watershed on Java Island, the area is about 16353.74km 2 and runs for 559 kilometers across East Java Province, Central Java Province, and the Special Region Province of Yogyakarta.As the longest river on Java Island, The Bengawan Solo River assumes a vital role in driving environmental development [1].Numerous impactful endeavors spanning three provinces unfold along its banks.Nevertheless, the river's attributes are susceptible to modification due to intensified human activity and the capricious nature of rainfall.This combination of factors has the potential to alter the inherent characteristics of the Bengawan Solo River and give rise to heightened flood discharges.The hydrology of the Bengawan Solo River is a complex and dynamic system that plays an important role in the ecological and socio-economic fabric of Java, Indonesia.The river's hydrological cycle is deeply affected by wet and dry seasons in the region's climate (tropical climate).During the rainy season, which typically spans from November to March, heavy monsoonal rains contribute to high water inflows, leading to increased water levels and potential flooding.One of the causes of flooding is the fluctuating rainfall [2].The increasing rainfall is one of the impacts of changes in climatology and land use [3].Short duration with high rainfall frequency and the land cover change can cause higher runoff discharge in an impermeable area [4].The inaccurate prediction of flood discharge can cause problems for the flood control infrastructure [5] not only that, the trash clogging problems can affect the decrease of drainage capacity and cause flood problems [6].
The increasing rainfall has the potential to induce river discharge, which might result in surface runoff and flooding [7].As a result, it is critical to identify changes in rainfall and river discharge using the trend model and analysis.Determining a hydrograph that corresponds to actual flood events is one of the most difficult processes in hydrological analysis [8], [9], [10].Nowadays, not only land use changes influence flood hydrographs but climate change is also a factor that influences changes, where flash floods are recorded more frequently than before [2], [9], [10], [11], [12].To comprehend the hydrology prediction about the influence of rainfall fluctuations on river discharge, simplification needs to be fulfilled.Modeling is one type of simplification that may be applied.The model is considered sufficient to analyze rain in the watershed with the flood hydrograph.Flood hydrographs are a graphical representation that depicts the reaction of a river or stream to a specific rainfall time.It shows how the discharge flow of water in a river changes over time because of the rainfall and subsequent runoff [13].The flood hydrograph will provide a correlation between rainfall and river discharge within a specific time frame.The SCS-CN model stands as a widely utilized tool for estimating direct surface runoff in unmeasured basins equipped with known rainfall data.This model is highly established and recognized because of its simplicity, stability, and predictability [14].Synder Unit Hydrograph is the formulation of an empirical equation for a synthesis unit hydrograph in a large catchment area.Synder Unit Hydrograph is suitable for ungauged basins and watersheds ranging in area from 25 km 2 to 25.000 km 2 [15].Therefore, the Snyder hydrograph is very suitable for use in the Bengawan Solo watershed which is included in the large watershed category.
Based on several identified problems above, the main object of this research is to analyze trends in changes in rainfall towards changes in flood discharge in the Bengawan Solo River with the right and accurate model.

Material and Method
In this research, the data used was obtained from the Central River Region of Bengawan Solo.The study utilized rainfall data from thirteen rainfall gauges between 2001 and 2022 and they are spread across the Bengawan Solo Watershed.The observation discharge data is from Cepu Water Level Station between 2001 and 2022.These observational discharge data were obtained from the results that have been analyzed by the Central River Region Bengawan Solo which is a part of The Ministry of Public Works and Housing, Indonesia.The calibration method in this research is by comparing the flood discharge modeled for a certain return period with the observed discharge that has been frequency analyzed in the same return period as the flood discharge model.After calibration, the flood discharge model with several return periods will become the basic data used to analyze trends in changes in flood discharge that occur.
Using the Polygon Thiessen technique, the design rainfall was analyzed based on rainfall data between 2011 and 2022 from 13 rainfall stations located across the Bengawan Solo Watershed.Following that, the design discharge was assessed using the hydrograph model of the Soil Conservation Service-Curve Number (SCS-CN) method and Synder Unit Hydrograph.The SCS-CN model offers significant benefits in macro-scale quantitative studies and planning analysis based on the hydrological simulations model.The SCS-CN model is a rainfall-runoff model created in 1954 by the Soil Conservation Service of the United States Department of Agriculture.It has a simple structure and multiple parameters, and it specifically examines the influence of soil type, land use, soil moisture, and other factors on rainfall-runoff [14].Synder Unit Hydrograph was developed by F. Synder by utilizing the empirical coefficients to connect the unit hydrograph elements with the watershed characteristics.
To obtain good accuracy in simulation and represent the real condition of a flood event, an adequate dataset for input into a flood model and a good calibration is needed.The Calibration was performed using a trial-and-error approach based on observation discharge data from the Cepu Water Level Station from 2001 to 2022.Meanwhile, to determine the relationship between observation data and modeling results, the correlation test used in this research was adapted from statistical tests.Essential for flood runoff generation, rainfall datasets input must be considered.In a distributed flood model, the spatial distribution of rainfall should also be taken into account, as it impacts parameter estimation and may result in calibration disagreements [16].Rainfall representation for runoff estimation was researched by Segond, et.al.[17] and Pechlivanidis, et.al.[18] and the results confirmed the importance of rainfall spatial variability.In this research, rainfall and flood discharge trends are analyzed using a comparison method between groups of data within the baseline data with the same return period.The following table is a dividing data sequence with the period for each data group being ten years.
The rainfall variability data group plays a pivotal role in understanding and analyzing the intricate patterns of precipitation that shape our environment [19].This data offers insights into the temporal and spatial distribution of rainfall, enabling scientists, policymakers, and researchers to comprehend the fluctuations in weather patterns, identify potential trends, and predict extreme weather events.As referenced by Smith, et all [20], studies utilizing long-term rainfall variability data have demonstrated the changing nature of precipitation regimes and the implications for ecosystems and societies.By delving into these datasets, we can enhance our grasp of the complex interactions between climate, geography, and human activities, ultimately fostering more informed and resilient decisionmaking processes.
The following figure is a work step in analyzing changes in trends that occur.

.Rainfall Design Trend Analysis
This study utilized the Thiessen Polygon approach to examine the design of rainfall depending on the rainfall region.This technique divides a given geographical region into multiple polygons, with each polygon encompassing a unique point called a rain gauge station.The rainfall data collected at each station is then assigned to the entire area within its corresponding polygon, assuming that the station's recorded rainfall is representative of that entire polygon's rainfall.By doing so, the method enables the creation of a detailed map that showcases the spatial distribution of rainfall intensity.
The rainfall data was obtained from several rainfall stations across Bengawan Solo Watershed from 2001 to 2022.The calibration watershed is located inside the Bengawan Solo Watershed.

Figure 3. The Impact of Rainfall Station Area in Calibration Watershed and Bengawan Solo Watershed
After analyzing the rainfall pattern, a frequency of occurrence analysis was conducted utilizing the Gumbel Distribution, focusing on the return period.this study will be using periods of 5,10,25,50 and 100 years.To analyze the accuracy of frequency analysis, Chi-Square, and Kolmogorov-Smirnov statistical tests are conducted.The Chi-Square statistic test is used to compare observed frequencies with expected frequencies under a given hypothesis, while the Smirnov Kolmogorov test evaluates the similarity between a sample distribution and a theoretical distribution.The following is the result of the analysis.

. Flood Hydrograph Analysis
Flood hydrograph analysis constitutes a fundamental aspect of hydrology, enabling comprehensive insights into the dynamic behavior of rivers and streams during extreme rainfall events.This analysis involves the examination of hydrographs, which graphically depict the variation of discharge over time.By dissecting the rising and recession limbs of hydrographs, hydrologists can discern crucial information about the basin's response to precipitation, including peak flow rates, lag times, and overall hydrological characteristics.The model calibration is one of the stages in hydrological modeling to obtain a good model of flood hydrograph.Calibration is a pivotal step that involves the adjustment of model parameters to align the simulated outcomes with observed data.This process fine-tunes the model and enhances its predictive accuracy and reliability.The initial step in the model calibration process is conducting a frequency analysis with Gumbel Distribution centered on the return period from the discharge observation data.The Gumbel Distribution is effective in depicting the upper or lower tails of datasets making it a preferred choice for modeling extreme events [21].The following are the results of the observation discharge frequency with Gumbel Distribution from 2001 to 2022 that has been analyzed.After conducting a frequency of occurrence analysis (Gumbel Distribution) with several periods, the next step of the calibration process is the trial-and-error method.The trial and error method for several hydrology parameters was analyzed based on comparing the modeled flood discharge for a certain return period with the observed discharge that has been analyzed for frequency in the same return period as the flood discharge.
The research showed the time lag and peaking coefficient are 19.20 hours and 0.5 in the calibration watershed.The correlation test between the flood model discharge and observation discharge between 2001-2022 at all return period times achieved a value of 0.9974.It shows that the calibration results were in a very strong relation category.[6], exemplifies the significance of flood discharge trend analysis in assessing changing hydrological dynamics.This research showcases the vital relationship between the analysis of the flood discharge model and the prediction of future flood risks.By leveraging insights from flood discharge trend analysis, stakeholders and policymakers can make informed decisions about flood mitigation strategies and sustainable water resource management, ensuring the resilience of communities and ecosystems in the face of changing hydrological patterns.In this study, the flood discharge trend changes were examined by comparing groups of data with the baseline data.The analysis of the flood discharge trend generated the following results.Based on the analysis, the extreme rainfall in 2001-2010 had a decreasing trend of around 7-10% compared to the baseline data.From 2007-2016, it displayed a decreasing trend ranging from 3-5%.However, in the last year, 2013-2022, the data showed a trend of increasing design rainfall by 10%.By this analysis, when we use the latest data for analysis, we obtain higher design rainfall, which is crucial for designing hydraulic infrastructure.Therefore, using the newest rainfall data is essential to ensure that the infrastructure has the appropriate capacity to overcome future flood discharge.

Conclusions
The analysis of evolving trends in rainfall and discharge design centered around the baseline data spanning from 2001 to 2022 within the Bengawan Solo Watershed unveiled significant insights.Specifically, the exploration of rainfall and discharge trends exposed a remarkable 4-6% rise in precipitation levels from 2013 to 2022.Consequently, there was a substantial 10% surge in flood flow within the basin.These findings give rise to a few considerations.First, the variations in rainfall intensity correspondingly mirror fluctuations in flood discharge volume.Second, the use of more recent data can yield increased results in the analysis.Lastly, apart from changes in rainfall, increased flood discharge can occur due to other factors such as changes in land use.

Figure 2 . 3 .
Figure 2. Research Flow Chart Extreme rainfall data from 2001 to 2010 indicates a 3-6% decrease compared to the baseline data.It also shows a declining trend ranging from 1-3% between 2007 and 2016.From 2013 to 2022, the pattern indicates an increasing design rainfall ranging from 4-6%.The figure below shows the correlation between rainfall fluctuations across different groups of years.

Figure 4 .
Figure 4. Percentage of Several Return Periods of Rainfall Design in Bengawan Solo Watershed

Figure 6 .
Figure 6.Percentage of Several Return Periods of Design Discharge in Bengawan Solo Watershed

Table 1 .
Value Criteria of the Correlation Test

Table 2 .
Data Sequences

Table 3 .
Various Periods of Design Rainfall of Bengawan Solo Watershed

Table 4 .
The Observation Discharge based on Several Return Periods

Table 5 .
Calibration Result of Flood Discharge in 2001-2022 in Cepu Calibration Watershed

Table 6 .
Flood Discharge 2001-2022 (Baseline Data) in Bengawan Solo Watershed Figure 5. Flood Discharge Model in Bengawan Solo Watershed (2001-2022) 3.3 .Flood Discharge Trend Analysis Flood discharge trend analysis aims to unravel patterns and changes in river flow for several periods.By scrutinizing flood discharge model data, we can discern trends that might reflect alterations in precipitation patterns, land use practices, or other factors influencing watershed behavior.Such analysis plays a crucial role in understanding the potential effect of climate change on flood occurrences.A study by 8