Screening of representative rainfall event series for long-term hydrological performance evaluation of grassed swales

Feikai Yang; Dafang Fu; Chris Zevenbergen; Floris C Boogaard; Rajendra Prasad Singh

doi:10.1007/s11356-024-32355-5

Screening of representative rainfall event series for long-term hydrological performance evaluation of grassed swales

Environ Sci Pollut Res Int. 2024 Feb 23. doi: 10.1007/s11356-024-32355-5. Online ahead of print.

Authors

Feikai Yang^{1

2

3}, Dafang Fu^{1

2}, Chris Zevenbergen³, Floris C Boogaard^{4

5}, Rajendra Prasad Singh^{6

7}

Affiliations

¹ School of Civil Engineering, Southeast University, Nanjing, 210096, China.
² Southeast University-Monash University Joint Research Centre for Future Cities, Nanjing, 210096, China.
³ Department of Civil Engineering, Delft University of Technology (TU Delft), Gebouw 23, Stevinweg 1, 2628CN, Delft, the Netherlands.
⁴ Research Centre for Built Environment NoorderRuimte, Hanze University of Applied Sciences, 9747 AS, Groningen, the Netherlands.
⁵ Deltares, Daltonlaan 600, 3584 BK, Utrecht, the Netherlands.
⁶ School of Civil Engineering, Southeast University, Nanjing, 210096, China. rajupsc@seu.edu.cn.
⁷ Southeast University-Monash University Joint Research Centre for Future Cities, Nanjing, 210096, China. rajupsc@seu.edu.cn.

PMID: 38393564
DOI: 10.1007/s11356-024-32355-5

Abstract

Evaluation of the hydrological performance of grassed swales usually needs long-term monitoring data. At present, suitable techniques for simulating the hydrological performance using limited monitoring data are not available. Therefore, current study aims to investigate the relationship between saturated hydraulic conductivity (K_s) fitting results and rainfall characteristics of various events series length. Data from a full-scale grassed swale (Enschede, the Netherlands) were utilized as long-term rainfall event series length (95 rainfall events) on the fitting outcomes. Short-term rainfall event series were extracted from these long-term series and used as input in fitting into a multivariate nonlinear model between K_s and its influencing rainfall indicators (antecedent dry days, temperature, rainfall, rainfall duration, total rainfall, and seasonal factor (spring, summer, autumn, and winter, herein refer as 1, 2, 3, and 4). Comparison of short-term and long-term rainfall event series fitting results allowed to obtain a representative short-term series that leads to similar results with those using long-term series. A cluster analysis was conducted based on the fitting results of the representative rainfall event series with their rainfall event characteristics using average values of influencing rainfall indicators. The seasonal index (average value of seasonal factors) was found to be the most representative short rainfall event series indicator. Furthermore, a Bayesian network was proposed in the current study to predict if a given short-term rainfall event series is representative. It was validated by a data series (58 rainfall events) from another full-scale grassed swale located in Utrecht, the Netherlands. Results revealed that it is quite promising and useful to evaluate the representativeness of short-term rainfall event series used for long-term hydrological performance evaluation of grassed swales.

Keywords: Bayesian network; Grassed swale; Long-term hydrological performance; Rainfall event series; Representativeness; Saturated hydraulic conductivity.

Grants and funding

201906090081/China Scholarship Council