How does flow alteration propagate across a large, highly-regulated basin? Dam attributes, network context, and implications for biodiversity

Large dams are a global driver of river ecosystem degradation. Although dams progressively alter natural flow regimes as water moves downstream, flow alteration research has largely ignored the dendritic structure of river networks. As a consequence, it remains largely unknown how altered flow regimes propagate over space. Here we used the highly-regulated Colorado River Basin as a model system to understand how flow alteration is spatially controlled by dams and the incorporation of free-flowing tributaries. Streamflow on a convergent dendritic network can be described by a spatial Markov process; accordingly, we used a spatial network in which flows downstream of dams were informed by flows immediately upstream. The network was informed by 117 upstream-downstream sets of monthly flow series from USGS gaging stations (2003-2017), which allowed estimating the impacts of 84 intermediate-to-large dams representing > 83% of the total storage in the basin. Using LASSO regression, we then explored how such contributions to flow alteration were explained by local dam properties (e.g., primary and secondary purpose, storage capacity) vs. network-level attributes (e.g., position, upstream cumulative storage). We found that flow alteration was highly variable across the network, and dams with significant impacts often overlapped with areas of high fish biodiversity, or areas with native fishes requiring seasonal flows that are no longer present. Importantly, dam impacts were explained to a higher extent by network-level attributes (63%) than by local dam properties (37%). The overriding effect of context shows that contrary to common practice, the impacts of dams on flow regimes may not be captured by studying single dams in isolation. Our methods are transferrable, and could guide screening for reoperation by identifying dams that disproportionally contribute flow alteration while being located in hydro-ecological contexts with high restoration potential.