Robustness analysis of storm water quality modelling with LID infrastructures from natural event-based field monitoring

Modeling the effect of land use change to design a suitable low impact development (LID) system to control surface water pollutants

Growth in urbanization has led to increased impervious surfaces, exacerbating flood risks and water quality degradation. This study investigated the impact of land use change and Low-Impact Development (LID) systems on urban runoff quality and quantity in the second region of Tehran. Pioneering an innovative approach, the integration of the Land Change Modeler (LCM) with the Stormwater Management Model (SWMM) signifies a paradigm shift in urban water management. Combined with other hydrological models, this new approach provides a comprehensive method for assessing the future effectiveness of LID practices. The Event Mean Concentration Method (EMC) was used in this study to measure Total Suspended Solids (TSS), Chemical Oxygen Demand (COD), Total Phosphorus (TP), and Zinc (Zn) in urban runoff from five land uses. Results pinpointed transportation land uses as the primary source of pollutants. Using LCM, the study forecasted a surge in urban runoff pollutants by 2030, particularly in the Northwest area of the region due to anticipated land use shifts towards commercial and residential land uses. Model results showed an 11 % increase in TSS over a decade, highlighting the importance of land use change in runoff quality. The study used three types of LIDs to reduce contaminants in dense urban areas. Assessing the impact of LID scenarios on runoff pollutants using SWMM revealed that the bio-retention cell had the best performance, reducing TSS by 20.92 %, and the vegetative swale had the worst performance, reducing TSS by 8.43 %. The study also concluded that combining LIDs would be more effective than using them separately. The results of this study suggest that LID systems can be an effective way to reduce urban runoff pollutants and improve water quality in the second region of Tehran. However, more research is needed to optimize the design and placement of LID systems in different urban areas.

Catchment-scale life cycle impacts of green infrastructures and sensitivity to runoff coefficient with stormwater modelling

Urban green infrastructure (GI) has been widely used in sponge city construction to manage hydrological processes. While studies on environmental benefits of GI from the perspective of whole life cycle assessment (LCA) have been reported in recent years, few have explored and compared the environmental performance of different GIs within a single catchment, which is directly linked to catchment-scale hydrological control. This study focuses on a Sponge City pilot project in Shenzhen, China, including three typical types of GI: permeable pavement, green roof, and sunken green space. By collecting hydrological data, land use, and life cycle inventory of GI and employing SWMM (Storm Water Management Model)-based stormwater modelling, we have revealed the environmental impacts at different stages of the life cycle of the GI scenario and three GIs through comparative and sensitivity analyses. Notably, we have disclosed, for the first time, the effect of the runoff coefficient in LCA. Our findings indicate that over the 30-year life cycle, the total environmental impact of the GI scenario is 24 % smaller than that of the hypothetical grey scenario. Permeable pavement exhibits the largest environmental impact per unit area, being 1.8 times and 7.6 times greater than that of the green roof and sunken green space, respectively. The operation stage of the three GIs significantly mitigates eutrophication and climate change. Furthermore, sensitivity analysis demonstrates that an increase in surface runoff undermines the environmental benefits of GIs. These results highlight the importance of embedding stormwater modelling into LCA, enabling catchment-scale integrated evaluation and equivalent assessment of different GIs within a single catchment whereby the influence of external factors such as climate change can be described, which aids in understanding the dynamic environmental performance of GIs. The proposed research framework and results are anticipated to provide valuable guidance for future GI construction and carbon-neutral policies.

Investigating non-point pollution mitigation strategies in response to changing environments: A cross-regional study in China and Germany

Climate change and urbanization have altered regional hydro-environments. Yet, the impact of future changes on the pollution risk and associated mitigation strategies requires further exploration. This study proposed a hydraulic and water-quality modeling framework, to investigate the spatiotemporal characteristics of pollution risk mitigation by low impact development (LID) strategies under future Representative Concentration Pathways (RCP) and Shared Socioeconomic Pathways (SSP) scenarios. Results demonstrated that the LID strategies exhibited an effective performance of pollutant removal in the current hydro-environment, with the removal rates ranging from 33% to 56%. In future climate and urbanization scenarios, the LID performance declined and turned to be uncertain as the greenhouse gas (GHG) emissions increased, with the removal rates ranging from 12% to 59%. Scenario analysis suggested that the LID performance was enhanced by a maximum of 73% through the diversified implementation of LID practices, and the performance uncertainty was reduced by a maximum of 67% through the increased LID deployment. In addition, comparative analysis revealed that the LID strategies in a well-developed region (Dresden, Germany) were more resilient in response to changing environments, while the LID strategy in a high-growth region (Chaohu, China) exhibited a better pollutant removal performance under low-GHG scenarios. The methods and findings in this study could provide additional insights into sustainable water quality management in response to climate change and urbanization.

First flush stormwater pollution in urban catchments: A review of its characterization and quantification towards optimization of control measures

Identification, quantification, and control of First-Flush (FF) are considered extremely crucial in urban stormwater management. This paper reviews the methods for FF phenomenon identification, characteristics of pollutants flushes, technologies for FF pollution control, and the relationships among these factors. It further discusses FF quantification methods and optimization of control measures, aiming to reveal directions for future studies on FF management. Results showed that statistical analyses and Runoff Pollutographs Applying Curve (RPAC) fitting modelling of wash-off processes are the most applicable FF identification methods currently available. Furthermore, deep insights into the pollutant mass flushing of roof runoff may be a critical approach to characterizing FF stormwater. Finally, a novel strategy for FF control is established comprising multi-stage objectives, coupling LID/BMPs optimization schemes and Information Feedback (IF) mechanisms, aiming towards its application for the management of urban stormwater at the watershed scale.

Linking downstream river water quality to urbanization signatures in subtropical climate

Urbanization has several hydro-ecological effects on receiving waters. Hence, understanding how urbanization influences river water quality is essential for proper river management. However, an inappropriate approach that correlates urbanization signatures with water quality may result in spurious correlations. This study aimed to investigate the relationship of urbanization signatures with two key pollutants of stream flows: nutrients and pathogens. In contrast to the commonly used approaches that are based on economic or demographic metrics, our approach represents urbanization signatures using related anthropogenic activities and evaluates the effect of such activities on water quality parameters. The approach was also applied to evaluate the impacts of urbanization on nutrient and pathogen trends in the river waters of Hong Kong. The data were collected for the period of 1986-2020 from the Environmental Protection Department and monthly measurements were performed. Total nitrogen (TN), total phosphorus (TP), Escherichia. coli (E. coli), and fecal coliforms (FC) showed consistently decreasing trends. However, the long-term seasonality of nutrients differed from that of pathogens. TP and TN exhibited homogenous seasonality with an approximately sinusoidal relationship from January to December, whereas the seasonality of pathogens was more complex and not dependent on river flow dilution effects. Additionally, urbanization impacts on station nutrients and pathogen characteristics were found to be unevenly distributed; under high water temperatures, nutrient concentrations were found to be decreased because of the rainfall dilution effect on river flows. Both urban point and diffuse sources of pollution significantly contributed to nutrient pollution in rivers. Furthermore, the concentrations of FC were not highly influenced by suspended solids, and dissolved oxygen was negatively correlated with all pathogens. Furthermore, the river flow rate was found to improve the water quality in terms of both nutrients and pathogens; urban point source pollution and river temperature alteration were shown to mainly contribute to seasonal variations in both nutrients and pathogens.

The control efficiency and mechanism of heavy metals by permeable pavement system in runoff based on enhanced infiltration materials

As one of the commonly used stormwater management measures, permeable pavement system (PPS) played a prominent role in controlling runoff pollution and alleviating urban waterlogging. In this study, new enhanced infiltration materials (construction waste brick, coal gangue, activated carbon, multi-walled carbon nanotube, multi-layer graphene) were applied in PPS and the control efficiency and mechanism of typical heavy metals (HMs, Mn²⁺, Pb²⁺, Zn²⁺, Cu²⁺, Cd²⁺, Ni²⁺) was investigated in runoff. Furthermore, the influences of different rainfall intensities and antecedent dry periods on HMs removal by PPS were evaluated. The results showed that all PPS with enhanced infiltration materials have little leaching effect on HMs (＜3 μg/L). All the selected enhanced infiltration materials meet the requirements of PPS. The concentration of HMs in the effluent of PPS dropped sharply first, followed rebounded and then maintained at a stable range. Activated carbon PPS (AC), Multi-walled carbon nanotube PPS (MCN), and Multi-layer graphene PPS (MG) could significantly improve the control effect of PPS on nearly all selected HMs. The average removal rates of AC, MCN and MG for six HMs were 75.48%-99.35%, 81.30%-97.59%, and 73.03%-99.33%, respectively. Compared with Traditional PPS (TR), the effluent concentrations of HMs in construction waste brick PPS (CW) and coal gangue PPS (CG) were relatively higher and unstable. AC, CN and MG could adapt to different rainfall conditions and the maximum removal rates of most HMs exceed to 99%. With antecedent dry periods increased, the control effect of HMs was significantly improved. The influences of the antecedent drying period on HMs removal followed as: CW＞CG＞TR＞MG＞CN＞AC. This study provided novel methods to eliminating HMs in runoff and provides implications for the design of PPS.

Impact of rainfall characteristics on urban stormwater quality using data mining framework

Understanding the impact of rainfall characteristics on urban stormwater quality is important for stormwater management. Even though significant attempts have been undertaken to study the relationship between rainfall and urban stormwater quality, the knowledge developed may be difficult to apply in commercial stormwater management models. A data mining framework was proposed to study the impacts of rainfall characteristics on stormwater quality. A rainfall type-based calibration approach was developed to improve water quality model performance. Specifically, the relationship between rainfall characteristics and stormwater quality was studied using principal component analysis and correlation analysis. Rainfall events were classified using a K-means clustering method based on the selected rainfall characteristics. A rainfall type-based (RTB) model was independently calibrated for each rainfall type to obtain optimal parameter sets of stormwater quality models. The results revealed that antecedent dry days, average rainfall intensity, and rainfall duration were the most critical rainfall characteristics affecting the event mean concentrations (EMCs) of total suspended solids, total nitrogen, and total phosphorus, while total rainfall was found to be of negligible importance. The K-means method effectively clustered the rainfall events into four types that could represent the rainfall characteristics in the study areas. The rainfall type-based calibration approach can considerably improve water quality model accuracy. Compared to the traditional continuous simulation model, the relative error of the RTB model was reduced by 11.4 % to 16.4 % over the calibration period. The calibrated stormwater quality parameters can be transferred to adjacent catchments with similar characteristics.

Sponge City Construction and Urban Economic Sustainable Development: An Ecological Philosophical Perspective

The Party's 14th Five-Year-Plan and the 2035 Visionary Goals point out that green and sustainable development is the direction of China's road in the present age, and provide a theoretical basis for further improvement of ecological civilization. "Sponge city" is a new type of urban construction idea in China; moving from pilot to vigorous implementation at present, it is an important element of China's promotion of green development and development of ecological civilization. At present, a number of sponge city pilot projects have been built in China, and evaluation of their effects is already a matter of urgency. The overall planning of China's current policies in sponge city construction and the specific analyses conducted by experts from both subjective and objective aspects have not been able to completely fill the gap in this regard, thus making it particularly urgent to conduct in-depth studies. Based on this, this paper discusses the performance assessment of sponge cities in China on the basis of the prediction and analysis of the development trend of sponge cities in China. In the performance assessment system, the correctness and timeliness of the index system should be considered in terms of practical effects; in the city performance assessment, the ideas of new city development such as low-carbon cities and smart cities should be combined to build a comprehensive and multi-perspective intelligent assessment system, so as to provide a strong boost to promote the development of city construction and its evaluative research. Firstly, a system-dynamic model is applied to sort out and combine its internal operation mechanism, and a set of evaluation systems based on the ecological philosophical perspective of the sponge city and urban sustainable development performance is established. Second, to improve the accuracy of the research results, parallelism tests and robustness analysis were conducted on this performance index evaluation system. The study's results show that sponge city construction has achieved good results in sustainable urban development and has contributed to future development.

Quantifying the effects of land use change and aggregate stormwater management practices on fecal coliform dynamics in a temperate catchment

Changes in land use and land cover (LULC) due to land development can lead to an increase in diffuse microbial pollutions and, consequently, degradation of the receiving aquatic ecosystem. However, the mechanisms underlying these phenomena are rarely considered in hydrological models. Therefore, in this study, fecal indicator bacteria (FIB) and total suspended solids (TSS) in a temperate catchment were simulated using a well-established water quality model (Personal Computer Storm Water Management Model) to systematically quantify the factors influencing their dynamics and the effects of stormwater management. Additionally, high-resolution data (e.g., water quality variables and LULC changes) were used to calibrate the model, which accurately reproduced the physical and biological features of the catchment. The results showed that increases in bare land areas and impervious cover in the catchment exceeded the Korean (as well as the USEPA-based) standard recreational water quality criteria for fecal contamination and TSS. Dissolved organic compounds (only during storm events), TSS, and total nitrogen (except during the pre-development phase) were the strongest predictors in shaping FIB dynamics. The multiple control of stormwater management reduced the FIB and TSS concentrations by approximately 65% in the catchment. The results of this study not only provide conclusions on the drivers of FIB and TSS dynamics and their quantitative contribution but also help in designing a methodology for empirical and ecological predictions of diffuse microbial and TSS pollution in a catchment with ongoing land development.

Evaluation of comprehensive benefits of sponge cities using meta-analysis in different geographical environments in China

With the rapid progress in urbanization, frequent urban waterlogging and non-point source pollution are threatening the living and health of human beings. Sponge city construction has become an effective means to curb urban waterlogging. Although related studies have explored the comprehensive benefits of sponge cities, few studies have been conducted on the effects of different geographical environments on runoff control and suspended solid (SS) removal. Based on 76 cities with sponge cities in China, this study used the meta-analysis method to evaluate the relationships of climate, terrain, underlying surface conditions, and construction area with the increase in the total annual runoff control rate and SS removal rate. The results reveal that the runoff control benefit can be significantly improved by sponge cities under the combined conditions of average annual precipitation of approximately 1000 mm, high fractional vegetation cover, sufficient soil fertility, a terrain slope i of ≤2%, and a permeability coefficient of strata of 100-200 m/d, especially in northern China, where the weight representing the quantity of comprehensive benefits was calculated to be 25.5%. In addition, the study results assist in reforming unfavorable geographical environments in the construction of sponge city, thus providing more effective solutions for tackling SS pollution. The most significant benefits of SS removal were obtained in north central China, where the weight was 21.4%. This study comprehensively investigated the effects of geographical environmental factors on the comprehensive benefits of sponge city reflected by the improvement in the total annual runoff control rate and the SS removal rate. The results will provide guidance for the planning and design of global sponge cities and effectively optimize the practice, scale, and location of existing construction based on specific geographical environments.

The Spatial Interaction Effect of Green Spaces on Urban Economic Growth: Empirical Evidence from China

This paper measures the impact of urban green space construction rate on urban economic growth from the perspective of spatial interaction. To this end, we collect panel data of 31 provincial capital cities in China from 2001 to 2020 and use spatial economics models for empirical testing. The research results are summarized as follows: the level of green space construction can attract talents and investment by improving the environmental level of the city, and these financial expenditures, foreign investment, and talents are conducive to urbanization, thus having a significant positive impact on urban economic development. In addition, it also has a significant positive spatial spillover effect. In addition, the construction of urban green space will also stimulate the environmental protection of neighboring cities, which has a significant positive spatial dependence. At this time, talents and investment are affected by the environmental construction of neighboring cities, and the economic development of the city has also been significantly improved. The spatial spillover effect of green space construction on the economic level of surrounding cities is also positive. The empirical conclusions provide references for implementing green development strategies and promoting high-quality economic development.

A Discussion on the Application of Terminology for Urban Soil Sealing Mitigation Practices

Soil sealing is one of the most serious environmental problems today regarding its impact on cities. This article presents an analysis of the different urban practices currently used to mitigate the effects of soil sealing in urban areas. The main typologies, characteristics, differences, similarities and objectives have been considered. The practices analyzed were SuDS (Sustainable Drainage Systems), LIDs (Low Impact Developments), BMPs (Best Management Practices), WSUD (Water Sensitive Urban Design), GI (Green Infrastructure), and NbS (Nature-based Solutions). To understand the impact of these terms, an analysis of their presence in the scientific literature over the last 10 years is carried out. The results indicate that the trend in the use of these terms is increasing, with the number of articles having doubled in the last 10 years. This indicates the importance that the problem of soil sealing has acquired in the world, and the relevant environmental benefits of addressing it.

Numerical Experiments on Low Impact Development for Urban Resilience Index

Low impact development (LID) has become one of the strategies that effectively mitigate the impacts of climate change. In addition to the ability to reduce nonpoint source (NPS) pollution caused by flash floods from the surface runoff, LID has also been applied to control water quantity under extreme rainfall events. Due to the fact that studies about LID configuration optimization tended to control water quantity and gradually ignored the main functions of water quality treatment, this study aims to consider water quantity and quality to estimate the benefits and optimal configuration of LID by Non-Dominated Genetic Algorithm (NSGA-II). In addition, regarding to the outlet peak flow, hydrologic footprint residence (HFR) was considered to be the water quantity indicator due to the ability to represent the dynamics of flow changes, and the modified quality indicator (Mass Emission First Flush ratio, MEFF30) was corrected to represent the pollutant transport process in a large catchment area. The results show that the flood and MEFF30 reduction rate of LID are inversely proportional to rainfall duration and intensity. The benefit of pollutant reduction, which can still be maintained by 20% and 15% under a big return period and the long duration was about three times than the quantity control. Taking the cost into account, although the rain barrel had the best effect of reduction per unit area, green roofs and permeable pavements had a higher unit cost reduction rate due to the lower costs. The upper and middle reaches of the open channel and the confluence of rainwater sewers should be the optimal LID configuration to achieve the benefits of both flood and pollution reduction.

Runoff control simulation and comprehensive benefit evaluation of low-impact development strategies in a typical cold climate area

With the acceleration of urbanization, the proportion of surface imperviousness is increasing continuously in cities, resulting in frequent waterlogging disasters. In this context, storm water management, based on the low-impact development (LID) concept, offers an effective measure for the management of urban storm waters. First, the storm water management model (SWMM) was built for a typical cold climate city (Changchun) in China. Next, the two-stage calibrated model was employed to explore the surface runoff and storm sewer control effects of four LID combination plans. Finally, these plans were put through a "cost-benefit" evaluation through an analytic hierarchy process. According to the results, after using four LID plans, the reduction rates of peak runoff exceeded 40% and the problem of overflow load of the storm sewage was significantly mitigated. The infiltration-oriented Plan I proved to be the optimal plan, with the lowest proportions of the overflow nodes and full-load pipe sections in each return period, as well as with maximum overall performance. This study offers technical and conformed methodological support to cold cities for the prevention and control of waterlogging disasters and recycling of rainwater resources.

Sustainable Urban Drainage System (SUDS) modeling supporting decision-making: A systematic quantitative review

Decision Support Systems (DSS) for Sustainable Urban Drainage Systems (SUDS) are a valuable aid for SUDS widespread adoption. These tools systematize the decision-making criteria and eliminate the bias inherent to expert judgment, abridging the technical aspect of SUDS for non-technical users and decision-makers. Through the collection and careful assessment of 120 papers on SUDS models and SUDS-DSS, this review shows how these tools are built, selected, and used to assist decision-makers questions. The manuscript classifies the DSS based on the question they assist in answering, the spatial scale used, the software selected, among other aspects. SUDS-DSS aspects that require more attention are identified, including environmental and social considerations, SUDS trains performance and criteria for selection, stochasticity of rainfall, and future scenarios impact. Suggestions for SUDS-DSS are finally offered to better equip decision-makers in facing emerging stormwater challenges in urban centers.

Effects of low impact development on the stormwater runoff and pollution control

The frequent urbanization and extreme rainfall events have posed the threat to the urban environment. The implementation of low impact development (LID) practices with great potential for control urban flood and overflow pollution is not comprehensively understood yet due to the influence of complex factors (i.e., hydrological pattern, installation location, and vertical parameter setting). In this study, the hydraulic and water quality model were used to analyze the hydrological and pollution reduction of outfall and storage under different hydrological patterns, vertical parameter setting, and green infrastructure installation locations, which can determine the best implementation of the scheme for overflow pollution control. The results showed that nine parameters of the vertical layer regarding the four parameters impacted the peak value and load of suspended solids (SS). The combination scheme of the LID practices was further proposed based on the selection and analysis of the single LID practice. Besides, considering the installation location, the downstream installed location was a better choice. The horizontal connection of overflow runoff and pollution could be reduced by up to 9.75% and 36.46%, respectively. In addition, the horizontal connection can effectively reduce the peak value of inflow and pollutants at the time of assessing storage tank impact, which reach the maximum of 14.08% and 29.25%, respectively. The pollutants distribution became uniform and showed better resilience against rainfall intensity, which is beneficial to the management of stormwater. Our findings can provide guidance for Sponge City construction and effectively alleviate the combined sewer overflow.

Flood Control and Aquifer Recharge Effects of Sponge City: A Case Study in North China

Sponge City is an integrated urban stormwater management approach and practice to tackle waterlogging, flooding, water scarcity, and their related problems. Despite many positive effects of Sponge City on flood control that have been investigated and revealed, the effect on aquifer recharge is still less known. Considering maximizing the function of natural elements such as surface water bodies and subsurface storage space, to minimize the use of a gray drainage system, a Sponge City design was proposed to substitute the planning development scheme in the study area. The stormwater management model of SWMM (storm water management model) and the groundwater flow model of MODFlow (Modular Three-dimensional Finite-difference Groundwater Flow Model) were adopted to evaluate the flood-control effect and aquifer-recharge effect, respectively. Compared with the traditional planning scenario, the peak runoff is approximately 92% less than that under the traditional planning scenario under the condition of a 5-year return period. Due to the increase in impervious areas of urban construction, the total aquifer recharge from precipitation and surface water bodies was decreased both in the present planning scenario and the Sponge City design scenario. However, the Sponge City design has a positive impact on maintaining groundwater level stabilization and even raises the groundwater level in some specific areas where stormwater seepage infrastructure is located.

Impact of green infrastructure on the mitigation of road-deposited sediment induced stormwater pollution

As a vital stormwater pollution source, the pollutants associated with road-deposited sediment (RDS) have become a growing concern in urban water management. Green infrastructure has exhibited great potential in stormwater pollution mitigation, but is not comprehensively understood yet due to the influences of complex RDS-associated pollutant migration processes (i.e., build-up, wash-off, and discharge). In this study, a city-scale hydraulic and water quality model was used to analyze the migration and removal processes of four RDS-associated pollutants (total suspended solids (TSS), chemical oxygen demand (COD), total nitrogen (TN), and total phosphorus (TP)) under different hydrological patterns, land-cover types, and green infrastructure installation locations. Results show that the antecedent dry-weather period was the main factor influencing RDS build-up, while the precipitation pattern strongly impacted RDS wash-off, discharge, and removal. The downstream-installed green infrastructures reduced the RDS-induced stormwater pollution by up to 68% and relieved the pollution-mitigation pressure of the studied drainage networks by almost 50%. The TSS and COD removal rates were higher (62.22-68.09%) near green space, while those of TN and TP were higher around buildings and roads (40.00-62.50%). Sensitivity analysis indicated that seven parameters regarding the surface layer characteristics and soil texture class strongly impacted the pollution-mitigation performance among the 31 technical parameters of green infrastructure. The results of this study would assist urban water management by optimizing green infrastructure for stormwater pollution mitigation.

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

A gated recurrent unit (GRU) network, which is a kind of artificial neural network (ANN), has been increasingly applied to runoff forecasting. However, knowledge about the impact of different input data filtering strategies and the implications of different architectures on the GRU runoff forecasting model’s performance is still insufficient. This study has selected the daily rainfall and runoff data from 2007 to 2014 in the Wei River basin in Shaanxi, China, and assessed six different scenarios to explore the patterns of that impact. In the scenarios, four manually-selected rainfall or runoff data combinations and principal component analysis (PCA) denoised input have been considered along with single directional and bi-directional GRU network architectures. The performance has been evaluated from the aspect of robustness to 48 various hypermeter combinations, also, optimized accuracy in one-day-ahead (T + 1) and two-day-ahead (T + 2) forecasting for the overall forecasting process and the flood peak forecasts. The results suggest that the rainfall data can enhance the robustness of the model, especially in T + 2 forecasting. Additionally, it slightly introduces noise and affects the optimized prediction accuracy in T + 1 forecasting, but significantly improves the accuracy in T + 2 forecasting. Though with relevance (R = 0.409~0.763, Grey correlation grade >0.99), the runoff data at the adjacent tributary has an adverse effect on the robustness, but can enhance the accuracy of the flood peak forecasts with a short lead time. The models with PCA denoised input has an equivalent, even better performance on the robustness and accuracy compared with the models with the well manually filtered data; though slightly reduces the time-step robustness, the bi-directional architecture can enhance the prediction accuracy. All the scenarios provide acceptable forecasting results (NSE of 0.927~0.951 for T + 1 forecasting and 0.745~0.836 for T + 2 forecasting) when the hyperparameters have already been optimized. Based on the results, recommendations have been provided for the construction of the GRU runoff forecasting model.