Adapting a regulatory water quality index under data scarcity: a reduced-parameter framework for tropical agricultural basins

Reliable application of regulatory Water Quality Indices (WQIs) is often constrained in data-scarce basins by limited monitoring and incomplete class representation. This study develops a parsimonious adaptation of the Colombian IDEAM Water Quality Index (ICA) for the Tocaría River basin while preserving consistency with the official regulatory formulation. The framework combines exhaustive subset selection, constrained weight calibration, and controlled synthetic-data support during calibration, with validation performed exclusively on observed data. Based on 70 monitoring samples, a four-subindex formulation (DO, TSS, pH, and NT/PT) reproduced the six-variable ICA with strong continuous agreement (mean R2 = 0.847; RMSE ≈ 0.023) and high categorical consistency within the observed IDEAM class support (accuracy ≈ 0.93; Cohen’s κ ≈ 0.85). Results suggest that reduced-parameter regulatory approximations can decrease monitoring requirements while maintaining fidelity to the reference index. However, conclusions remain conditional on limited sample size and incomplete representation of official ICA classes.