Estimating peak flow of Farahzad River in Tehran under climate change and debris flow scenarios: A novel approach and its implications for urban flood hazard mapping

Peak runoff is one of the potential hazards in urban rivers, which is exacerbated by climate change and debris flows (Zhang et al. 2016; Li et al. 2019; Banihabib et al. 2020b; Gao et al. 2020; Gong et al. 2023), in areas where debris flow occurred, the trees became large woody debris, and the trees were supplied to the river from banks affected by debris flows and eroded riverbanks, which can increase the potential risk related to local wood accumulation, bridge clogging, or damage to hydraulic structures (Ruiz-Villanueva et al. 2022). If the current climate changes and debris flow continue in the coming years, any floodplain of the world threatened by the increased flooding due to the river systems may fail to accommodate the increased peak runoff (Bibi et al. 2023).

Flood hazard maps show the extent and expected water depths/levels of an area flooded in different scenarios, a low probability scenario or extreme events, with different return periods (Gigović et al. 2017). Within flood hazard mapping, extreme events, such as those influenced by climate change on peak flow, are prominently featured. These events are typically simulated using general circulation models to forecast precipitation patterns in future timeframes (Booth et al. 2013; Ault et al. 2014; Cullather et al. 2014; Ostad-Ali-Askari et al. 2020). When it comes to runoff or floods, hydrological models and empirical methods are usually employed to investigate the impact of climate change on the flows (Zhang et al. 2015).

Another scenario is the effect of debris flow on the peak flow and corresponding flood map. Debris flow assessment schemes are commonly based on empirical, physical, geomorphological, and numerical methods and techniques (Hungr et al. 1984; Van Asch & Van Steijn 1991; Bovis & Jakob 1999; Bianco & Franzi 2000; Gatwood et al. 2000; Marchi & D'Agostino 2004; Glade 2005; Jakob 2005; Gartner et al. 2008; Pak et al. 2009; Milne et al. 2012; Guo et al. 2016; Banihabib & Nazarieh 2019; Villacorta et al. 2020; Vagnon et al. 2022; Dias et al. 2022). These methods are either very expensive to implement or require historical data on debris flows to be collected (Banihabib et al. 2017; Banihabib & Forghani (2017)). A procedure to estimate debris flow volume is proposed that uses peak discharge as ‘fluid flow’ and then, by applying the relations described by Takahashi (2007), estimates debris flow and sediment volumes from the fluid flow.

Finally, for the woody debris scenario, knowing the vulnerability of riparian vegetation to uprooting is a key aspect (Francalanci et al. 2020). Several studies have focused on the mechanics of uprooting trees under the effect of intense flow (Preti et al. 2010; Edmaier et al. 2011, 2015; Tron et al. 2015; Giambastiani et al. 2017; Perona & Crouzy 2018; Bau et al. 2019; Calvani et al. 2019; Jin et al. 2022).

A procedure by Francalanci et al. (2020) to estimate tree uprooting was used in this study. The method uses a detailed vegetation survey and the hydraulic modeling of the flood event to investigate the hydraulic forcing on the tree. They concluded that the Froude number represents the key parameter in the evaluation of the instability parameter so that we can predict the total amount of woody debris produced during very intense floods (Francalanci et al. 2020).

Creating a flood hazard map under different scenarios can be achieved through three major approaches: physically based, empirical, and physical modeling methods (Bellos 2012; Teng et al. 2017). Among physically based hydrodynamic modeling tools is the Hydrologic Engineering Center's River Analysis System (HEC-RAS), which can be used for sediment transport modeling (Mudashiru et al. 2021). However, it is not capable for modeling trees uprooting, which can cause a divergence between observed and simulated floodplains (Gibson et al. 2022). In this article, the prediction of uprooted trees is discussed using the Froude number output from the software and then incorporated into the model.

According to the literature, possible scenarios are considered separately. It seems necessary to have a comprehensive approach that includes all possible scenarios and multiple factors in estimating peak flow and flood hazard map. The present study introduces a versatile and practical framework for estimating peak flow and floodplain mapping, considering the impacts of climate change, sediment, and woody debris by integrating hydrological, sedimentological, and hydraulic methods and models. This framework is designed to be adaptable for implementation in various watersheds. By incorporating these factors, our approach improves the accuracy of peak flow and floodplain extent estimations, contrasting with conventional methods that tend to underestimate the peak flow and overlook woody debris effects. The application of this framework supports urban river flood mitigation planning by policymakers and aids stakeholders in minimizing economic losses associated with such floods (Mudashiru et al. 2021).

The specific objectives are to (1) investigate the impact of climate change on the peak flow, (2) investigate the debris flow occurrence under climate peak discharge and debris flow effect on peak flow, (3) investigate the woody debris occurrence from hydraulic modeling and culvert blockage, and (4) generate and validate the flood hazard map in the baseline period and develop future flood hazard maps by considering multiple factors.

Farahzad basin has geographical coordinates of 22′- 51° to 20′- 51° east longitude and 45′-35° to 53′-35 ° north latitude (Figure 1(a)). In addition to the main tributary, the Farahzad River has two tributaries called Behroud and Moradabad. There are three culverts, Khoshmaram, Abshar, and Niyayesh, in the study area (Figure 1(b)). The watershed boundary and sub-basins of the region are as follows: Farahzad mountain and urban catchments. The physiographic parameters of the basin are described in Table 1.

The hydrometric station with an acceptable range of precipitation data was presented in the study by Binesh et al. (2019).

In assessing peak flood discharge within the study area, encompassing both mountainous and inner-city catchments, we applied two distinct rainfall–runoff methodologies.

This method is one of the most popular methods for computing the volume of surface runoff for a given rainfall event from small agricultural, forest, and urban watersheds (Boughton 1989; Mishra & Singh 2003; Lian et al. 2020), in which the CN (curve number) must be calculated first. CN is the number of curves related to the amount of water infiltration in the soil of the basin, which is calculated based on the area's hydrologic soil group and land use; then according to the calculation of the flood peak flow using the recorded flood data of hydrometric stations, the CN value of each sub-basin is calibrated and determined according to the calculated peak.

In this method, the rational runoff coefficients (the rational C) must be calculated first (Young et al. 2009). The calibration values are presented in Table 2; with the assumption of considering CN, the same for the future period, the peak flood flow under climate change is calculated from Equations (1)–(3) of the soil conservation service (SCS) method and Equation (4) of the rational method. The results are presented in Table 2.

In the proposed procedure to estimate sediment volume, is used to convert fluid flow discharge into debris flow discharge. Banihabib & Masumi (2008) verified this equation for the Masoleh watershed on the northern slopes of the Alborz Mountains, and Banihabib & Forghani (2017) verified this equation for the Tajrish watershed on the southern slopes of the Alborz Mountains.

Generally, the angle of repose is related to the static friction coefficient and the angle of internal friction. As previously mentioned (Santamarina & Cho 2004; Das 2010), the angle of repose is often assumed to be equal to the internal friction angle. Moreover, the river is separated into two segments according to changes in and along the river, and so the calculated results for are presented in Table 3.

Table 3

coefficient according to changes of and along the river

.	.	.	.	.
From the beginning of the basin to the Abshar bridge	0.07	0.4	0.13	1.27
From Abshar bridge to Niyayesh bridge	0.04	0.22	0.13	1.3

.	.	.	.	.
From the beginning of the basin to the Abshar bridge	0.07	0.4	0.13	1.27
From Abshar bridge to Niyayesh bridge	0.04	0.22	0.13	1.3

View Large

Floodplain mapping in HEC-RAS integrated various parameters, including a digital elevation model (DEM), hydraulic parameters (n value), boundary conditions, river structures, and peak flow data. Following the simulation, the results were converted into 2D using HEC-GeoRAS, and a Google Earth projection was created to better visualize the inundation in the study area.

• i.

  Five stations were strategically chosen to examine the cumulative frequency of flora, considering the number, average distance, and species type. Figure 6 shows the cumulative frequency of flour per station. The number (Figure 7), the average distance (Figure 8), and the type of species (Table 4) in the study area were calculated by site investigations, geographic information systems (GIS) techniques, and remote sensing within Google Earth Pro software and satellite images.

• ii.

  The approach introduced by Francalanci et al. (2020) was adopted to assess uprooted trees based on channel Froude numbers. In river segments where the channel Froude number exceeds 0.5, the study anticipates that force-driving trees will exceed the resistance forces, resulting in uprooting. Froude numbers obtained from hydraulic modeling outcomes, particularly using HEC-RAS, play a crucial role in this analysis (Figure 9).

• iii.
  The volume of extracted trees was computed using Equations (10)–(12) by Tabacchi et al. (2011), accounting for species-specific coefficients. Table 4 provides the prevailing tree species in the region along with their associated coefficients.

  

  (10)

  

  (11)

  

  (12)

  where D (cm) is the diameter, h (m) is the height of the plant, and V (dm³) is the volume of the trunk and the larger branches.

Due to the presence of three culverts in the study area (Figure 1(b)), the timber volume determined in Section 2.5.1 is computed for each river segment preceding every culvert. Subsequently, this calculated volume is applied through the block obstruction tool to the HEC-RAS model (Figure 13).

For the 50-year return period, the climate change peak flow is 2.05 times the base peak flow, and for the 100-year return period, the climate change peak flow is 2.5 times the base peak flow. The total debris peak flow for the 100-year return period is 3.25 times the base peak flow. Also an increasing trend is observed in flood parameters as expected due to the increase in peak flow. This trend leads to inundation that can have immense consequences, especially in the Khoshmaram and Abshar bridge sections due to non-engineered construction in these areas. However, the Niayesh bridge is not heavily affected due to its engineering design. (Figure 14).

This framework also emphasizes the determination of hotspots concerning inundation and flood risk by Google Earth projection, especially for urban parts containing buildings, roads, parks, and farmlands (Figure 13). Between hotspots are culverts and bridges, which are the major bottleneck in the river channels (Hadidi et al. 2020) and experience the exacerbating effects of flood due to blockage by woody debris. By increasing the flood velocity at the entrance of bridges, there is the possibility of increasing scour in these parts.

Another critical hotspot is the inundated riparian zone with a high Froude number – the primary variable controlling the uprooting process of trees – which creates woody debris in the rivers. During flooding, large wood in debris flows can cause significant damage to check dams, bridges, and other major infrastructure in its flow path (Chen et al. 2020).

These results are based on the assumption that the CN value of the future period is the same as the current time. The Tehran government has implemented the Farahzad District limitation of construction and development plan, and so key impact factors of CN include soil properties, land use type, slope, antecedent, vegetation coverage, land management practices (Hawkins 1975; Boughton 1989; Chin 2017; Lal et al. 2019; Zouré et al. 2019; Lian et al. 2020) are stable. It is recommended for future applications of this framework in different watersheds to take into account these assumptions.

Another assumption is the neglect of obstacles posed by woody debris along the river. Field observation and hydraulic model results during the peak flow indicated that wood size is smaller than the river bankfull width that can be floated downstream during high flows (Marston 1982; Maser et al. 1988; Mao et al. 2013; Ruiz-Villanueva et al. 2016).

The limitation lies in the absence of observational data for debris under climate flow conditions, necessitating the assumption of a constant Manning's coefficient for both the base peak flow and debris under climate conditions. It is recommended that the future research endeavors focus on gathering more relevant data, conducting comprehensive sensitivity analyses, and exploring alternative modeling approaches capable of accommodating variability and changing conditions.

It is believed that this comprehensive approach will assist researchers in different fields, including urban design, river restoration, bridge design, and flood risk management.

This study aimed to estimate the peak flow of urban rivers by considering the effects of climate change and debris flow on flood hazard. A novel approach was proposed that integrates hydrological, sedimentological, and hydraulic models and methods to account for the influence of sediment volume, woody debris, and culvert blockage on peak flow and floodplain estimation. The approach was applied to the Farahzad River basin in Tehran, Iran, using future data downscaled from a global climate model under the RCP8.5 scenario for 50- and 100-year return periods. The results indicated a noticeable increase in climate change and debris flow scenarios, leading to higher peak flood discharge and associated factors. Specifically, there was a 10–20% expansion in the flood zone, a 15–30% rise in flood velocity, and a 10–40% elevation in inundation depth.

The occurrence of debris flows in the basin due to its location in the Alborz mountain range and the occurrence of debris flows in this area for many years were aligned with historical instances mentioned in Section 1 (Banihabib et al. 2020a).

The study establishes that tree uprooting caused woody debris accumulation, backwater rise, and overflow at upstream bridges, aligning with Martín-Vide et al. (2023), Harada et al. (2023), Gibson et al. (2022), and Okamoto et al. (2020), who also reported blocking upstream. Notably, downstream wide-span highway bridges remained unblocked.

By incorporating these factors, our approach improves the accuracy of peak flow and floodplain extent estimations, contrasting with conventional methods that tend to underestimate the peak flow and overlook woody debris effects. The application of this framework supports effective urban river flood risk mitigation planning by policymakers and aids stakeholders in minimizing economic losses associated with such floods.

These results indicate a significant increase in urban flood risk and suggest the need for more effective mitigation measures. The proposed approach provides a useful tool for estimating the peak flow of urban rivers by considering multiple factors in an integrated manner. It also offers a more accurate and reliable basis for determining urban rivers' buffer zone and designing flood protection structures. The study contributes to the advancement of urban flood risk management and enhances the resilience of cities to climate change and debris flow impacts.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.