Nagavali River Water-Quality Dataset and Reproducibility Code for Cost-Efficient Reduced EWQI Monitoring

This repository provides the cleaned analytical dataset and complete reproducibility code accompanying the study “Evaluating Cost-Efficient Reduced Monitoring for River Water-Quality Indices: Spatial, Forward-Temporal and Forward-Period Model Transfer Evidence from the Nagavali River Basin.” The dataset contains 420 monthly station–month observations from seven monitoring stations in the Nagavali River Basin, eastern India, covering January 2020 through December 2024. The analytical workbook contains station and month identifiers together with 24 reported water-quality variables. The source monitoring observations originate from the National Water Quality Monitoring Programme (NWMP) of the Central Pollution Control Board, Government of India; the distributed workbook is the cleaned analytical version used in the study. The accompanying Python workflow reproduces the computational analyses reported in the manuscript and Supplementary Information. These include construction of a 14-parameter data-driven relative Environmental Water Quality Index (EWQI); entropy, CRITIC and equal-weight robustness analyses; principal component analysis and empirical orthogonal function analysis; evaluation of fixed Core-4 and Consensus-6 reduced-monitoring configurations; OLS, ridge-regression and XGBoost surrogate modelling; spatial leave-one-station-out, forward-temporal and 2020–2022 → 2023–2024 forward-period validation; analytical cost–performance analysis; EOF structural-fidelity analysis; and Morlet wavelet-coherence analysis. The Consensus-6 configuration comprises dissolved oxygen, chemical oxygen demand, sulphate, ammonia-N, pH and total coliform. The EWQI is a data-driven relative index rather than a regulatory compliance index, with higher values representing relatively better conditions within the analysed record. The repository contains four analytical Python scripts and a master runner (05_reproduce_all.py). Running the master script with the full profile reproduces the computational workflow, including manuscript Figures 2–10 and Supplementary Figures S1–S10. Manuscript Figure 1 is a GIS/cartographic study-area figure and is not regenerated because the original spatial layers are outside this analytical package. Dataset period: January 2020–December 2024Stations: 7Station–month observations: 420Primary analysis: Core-14 relative EWQIReduced configuration: Consensus-6Software: PythonRandom seed: 42 The exact SHA-256 checksum of the analytical workbook used in the reported full reproduction run is: c06ec438dc004f2dfafb26faa1c68d38a97b0bcd655ac386af3c10023f46cb87 Users reusing the dataset should acknowledge the original NWMP/CPCB monitoring-data source in addition to citing this Zenodo record and the associated article.