Understanding which hydrological data types provide the most valuable information for models is crucial, given the limitations of data availability. This study applies data worth analysis to evaluate the impact of various observation types on predictive uncertainty in a coupled SWAT-MODFLOW-RT3D model simulating water flows and nitrate transport in a small headwater catchment in New Zealand. We assessed the worth of continuous nitrate concentrations, in-catchment flow measurements, and SkyTEM-derived groundwater levels for predicting stream flow and in-stream nitrate concentrations. Using PEST software for model calibration and linear uncertainty analysis, we determined the relative worth of different observation types. Results indicate that SkyTEM estimates of groundwater levels and continuously measured nitrate concentrations were particularly effective in reducing predictive uncertainty. This study highlights the value of integrating high-resolution SkyTEM data into models to enhance prediction accuracy for groundwater levels, stream flow, and nitrate pollution. It also demonstrates nitrate's utility as an environmental tracer, refining our understanding of surface water–groundwater interactions and solute transport in the Piako Headwaters Catchment.