Supplemental Materials for "Smart-Green Hybrid Digital Twin Framework for Vetiver NbS Scour Countermeasures"
Description
This dataset contains supplemental materials supporting a peer-reviewed journal manuscript. Tables S1-S8 provide: S1) HEC-RAS 2D mesh sensitivity analysis at the abutment; S2) Manning’s n calibration for vetiver channels from literature and field tests; S3) Digital twin alert thresholds for velocity, shear stress, and scour depth; S4) Sources of 140 laboratory cases used for Random Forest model training; S5) RF hyperparameter ranges and optimal values from 5-fold cross-validation; S6) UAV image dataset summary for vetiver health classification; S7) Bill of quantities comparing vetiver NbS vs. riprap, using NHA CSR 2023 rates; S8) Embodied carbon calculations using Ecoinvent 3.9 and ICE v3.0 factors. Figures S1-S4 include: S1) HEC-RAS 2D mesh layouts showing 4m, 2m, and 1m refinements; S2) Sample UAV training images from DJI Mavic 3E showing classes for healthy vetiver, stressed vetiver, bare soil, and water; S3) Field photographs of vetiver growth stages at Kala Khatai site, Mar-Jun 2026; S4) Lifecycle cost comparison charts for vetiver NbS vs. conventional riprap. All data were generated using HEC-RAS 6.4.1, Python 3.11.5 with scikit-learn 1.3.2, Agisoft Metashape 2.0, and publicly available sources. Raw code and processing scripts are included where applicable. The main article is complete without this supplemental data.
Files
Steps to reproduce
Steps to reproduce Tables S1-S8 and Figures S1-S4: Software required: HEC-RAS 6.4.1, Python 3.11.5, pandas 2.1.1, scikit-learn 1.3.2, Agisoft Metashape 2.0, matplotlib 3.8.2, Microsoft Excel. Table S1 - Mesh Sensitivity: 1. Build HEC-RAS 2D model using geometry in manuscript Fig. 2 2. Run unsteady flow with 100-yr hydrograph for mesh sizes 4m, 2m, 1m at abutment 3. Export max shear stress at Section A-A. Values recorded in Table_S1_Mesh Table S2 - Manning's n: Literature review of n values for vetiver channels. Sources: Truong et al. 2008, Yoon et al. 2015, NHA Field Test 2024. Calibrated value n=0.045 used in HEC-RAS 2D Table S3 - Thresholds: 1. Run 100-yr design flood in HEC-RAS → record max velocity=2.08 m/s, shear=82.7 Pa, scour=1.47 m 2. Apply Factor of Safety: Velocity 1.2x, Shear 1.03x, Scour 1.7x → thresholds in table Table S4 - RF Sources: 140 laboratory cases compiled from Melville 1992, Froehlich 1989, Sturm 2004, Kothyari 2007, Bento et al. 2023. Raw data available from original publications Table S5 - RF Hyperparams: Python script 'RF_tune.py' using GridSearchCV, 5-fold CV. Optimal params: n_estimators=500, max_depth=10, min_samples_split=4, min_samples_leaf=2, max_features=sqrt. CV R²=0.934 Table S6 - UAV Dataset: 1240 images collected with DJI Mavic 3E, 20MP RGB+multispectral, Mar-Jun 2026. Classes: Healthy vetiver 660, Stressed 330, Bare soil 165, Water 85. 80/20 train/val split Table S7 - BOQ: Quantities from AutoCAD Civil 3D. Unit rates from NHA CSR 2023 Pakistan. Riprap D50=300mm. Vetiver slips @ PKR 60/slip. Exchange rate 280 PKR/USD Table S8 - Carbon: Emission factors: Ecoinvent 3.9, ICE v3.0, Pakistan NDC 2021. Diesel 2.68 kgCO2e/L. Calculations in 'carbon_calc.xlsx' sheet Figure S1 - Mesh Layouts: Exported directly from HEC-RAS 6.4.1 RAS Mapper → TIFF 300dpi. Shows 4m, 2m, 1m refinements at abutment used in Table S1. Figure S2 - UAV Samples: Images acquired with DJI Mavic 3E, 20MP RGB+multispectral, Mar-Jun 2026. Processed in Agisoft Metashape 2.0. Class labels verified by field survey at Kala Khatai site. Figure S3 - Vetiver Growth: Field photographs taken with Canon EOS R5 at Kala Khatai site. Documents growth stages at 0, 30, 60, 90 days post-planting. Figure S4 - Cost Charts: Generated in Python 3.11.5 using matplotlib 3.8.2. Script 'plot_S4.py' included in zip. Input data from Table_S7_BOQ. Unzip Supplemental_Figures_S1-S4.zip to view figures. All source files and scripts included where available. Runtime: All tables and figures reproducible in
Institutions
- Shandong Jianzhu UniversityShandong, Jinan