Multivariate Approach of End-Member Contributions to Streamflow in the Critical Zone. Case Study: Valles Caldera, New Mexico

Abstract

Multivariable End-member Mixing Analysis (EMMA) that incorporates principal component analysis (PCA) is a widely utilized tool to identify the sources of water that generate streamflow in catchment hydrology. In this study, we investigated how different combinations of principal components (PC) allow assessing the importance that potential end-members have to surface waters. We evaluated mixing spaces of different dimensions in order to justify the number of end-members needed to generate streamflow. Furthermore, this multidimensional approach provided further evidence of the hydrologic processes that dominate in the headwaters at the Jemez River Basin Critical Zone Observatory (JRB-CZO). Our results showed that the U-mixing spaces of three dimensions of the La Jara and Upper Jaramillo catchments highlight the contributions of deep groundwater that the two-dimensional mixing space neglected. Conversely, in the History Grove catchment, a two-dimensional U-mixing space was enough to explain streamflow generation. Groundwater, snowmelt, rainfall and soil water are the end-members identified in each catchment. The geomorphology (e.g. aspect, topography, and geology) of each watershed and climate variability, however, influence the contribution of these source waters in each system. Groundwater contributions dominate streamflow generation in the JRB-CZO. Moreover, increments of snowmelt, rainfall and soil water contributions are observed specifically during base-flow conditions. We argue that the contributions of these end-members do not correspond specifically to overland flow, but rather contributions of shallow groundwater and subsurface lateral flow that possess the chemical signature of these source waters.