The impact of standard and hard‐coded parameters on the hydrologic fluxes in the Noah‐MP land surface model

Summary

This technical study applied global sensitivity analysis to the Noah-MP land surface model to characterise how 117 model parameters (both standard and hard-coded) influence hydrologic predictions across diverse US catchments. The work revealed that two-thirds of standard parameters significantly affect hydrologic output, but notably, the most oversensitive parameter is a hard-coded value governing soil surface resistance for direct evaporation. The findings highlight the importance of including both plant and soil parameters in model calibration, and suggest that hard-coded parameters—currently fixed in model code—should be exposed as adjustable variables to improve model robustness and calibration capability.

UK applicability

The methodological approach of exposing and sensitivity-testing hard-coded parameters could be applied to land surface models used in UK hydrological and weather prediction systems. However, the specific parameter sensitivities identified are calibrated to US catchment characteristics and may differ under UK climate and soil conditions.

Key measures

Sobol' global sensitivity indices for standard parameters (42 of 71) and hard-coded parameters (75 of 139); latent heat flux; total runoff; surface runoff; component flux sensitivities

Outcomes reported

The study identified 139 hard-coded parameters in the Noah-MP land surface model and performed global sensitivity analysis on hydrologic output fluxes (latent heat and total runoff) across 12 US catchments with contrasting hydrometeorological regimes. The analysis quantified which model parameters most influence hydrologic predictions and calibration outcomes.

Topic tags