A Bayesian chemistry-assisted hydrograph separation (BACH) approach was previously demonstrated using 15 years of monthly total phosphorus (TP) and total nitrogen (TN) data from eight mesoscale catchments in New Zealand’s North Island. Calibration was done separately for three 5-year data periods, and in each period, concentrations of the two tracers (TP and TN) discharged from each of the three separated flow paths—fast (event-response near-surface flow), medium (seasonal shallow local groundwater flow), and slow (persistent deeper regional groundwater flow)—were assumed to be constant. This approach has now been extended to reveal non-linear trends in the tracer concentrations in each flow path, each represented using a four-parameter curve (initial and final values of a linear trend plus two harmonics). The extended method (called BACH2) identified clear TP and TN concentration trends in the medium and slow flow paths in most of the eight catchments. TP and TN concentration trends in the fast flow path were generally uncertain, however, due to the infrequency and inherent variability of concentrations sampled during high flow conditions. Concentrations closely matched previously published results from the constant-concentration BACH model calibrated to shorter data series. The BACH2 approach is a powerful tool for revealing concentration trends in the different pathways that sustain stream flow using commonly available water quality and flow data. This type of analysis has not previously been available outside of complex distributed simulation models.

以前使用15年来自新西兰北岛八个中尺度集水区的每月总磷（TP）和总氮（TN）数据证明了贝叶斯化学辅助水文学分离（BACH）方法。分别对三个为期5年的数据周期进行校准，并且在每个周期中，分别从三个分离的流路中的每个（快速（事件响应近地表流），介质（季节性的局部浅层地下水流量）和缓慢的（持久的深层区域地下水流量）被认为是恒定的。现在已经扩展了该方法，以揭示每个流动路径中示踪剂浓度的非线性趋势，每个趋势都使用四参数曲线表示（线性趋势的初始值和最终值加上两个谐波）。扩展方法（称为BACH2）可在八个流域中的大多数中，缓慢流径中确定清晰的TP和TN浓度趋势。但是，由于在高流量条件下采样的浓度不高且固有变化，通常无法确定快速流动路径中的TP和TN浓度趋势。浓度与校准至较​​短数据系列的恒定浓度BACH模型的先前发布结果非常匹配。BACH2方法是使用常用的水质和流量数据揭示维持水流的不同途径中浓度趋势的强大工具。在复杂的分布式仿真模型之外，以前没有这种类型的分析。