Evaluating River Monitoring Network Adequacy Using Remote Sensing, Biotic Indices, and Scenario Modelling

Freshwater monitoring networks are often biased toward major rivers and chemical indicators, creating blind spots in ecologically sensitive areas. This study tests whether a legacy monitoring network is positioned in the right parts of its catchment, using three lines of evidence independent of the chemical record, remote sensing to look back, biological indices to assess the present, and catchment modelling to look forward, applied to the Rangitīkei catchment, New Zealand. NDVI, NDBSI, and water quality trends were assessed with a Bayesian change-point method; a catchment-wide macroinvertebrate survey extended coverage beyond routine sites; and the InVEST Nutrient Delivery Ratio model, calibrated against measured nitrogen loads, simulated land use change scenarios. Neither remote sensing nor water quality indicated widespread deterioration, with no directional trends detected in either. Macroinvertebrate surveys identified severely degraded sites outside the existing network, including two below the national bottom line in a single survey. Converting suitable land to dairy farming predicted nitrogen export increases of up to 110% at the most exposed site, and cropping conversion up to 69%. Riparian buffers combined with a 20% reduction in nitrogen loss from high-producing grassland predicted reductions of up to 32%. Sites with the greatest predicted risk, primarily Tutaenui Stream and Kawhatau River, are not well covered by the existing network. Network design methods are largely built around statistical representativeness or accessibility rather than catchment process. This framework identified priority gaps such designs would likely miss, offering a transferable basis for auditing networks or informing new ones.