The seasonality of gross primary productivity (GPP) in rivers reflects complex interactions between multiple environmental drivers. Annual light regimes are impacted by light attenuation from riparian canopies, periods of high flow can reduce autotrophic biomass, and thermal regimes reflect changes in channel characteristics or land use. The degree to which these drivers influence productivity may also change with network position through changes in channel characteristics or flow regimes associated with changing river size. This means that network scale productivity should reflect not only environmental drivers, but also the structure of the river network. Here we ask, are there characteristic river productivity regimes? Further, does variation in GPP within individual river reaches culminate in an emergent river network productivity regime? To address these questions, we first performed a classification on continuous time series of GPP for 47 U.S. rivers to generate a typology of river productivity. This typology provided a framework for mapping observed patterns in seasonality to relevant network positional information such as width or watershed area. We applied these regimes to individual river reaches within simulated Optimal Channel Networks (OCNs) based on the distribution of river widths to analyze the emergent annual productivity regime for a river network.
Results/Conclusions
From our dataset we identified four productivity regimes of rivers that differed with respect to their magnitude and seasonality of GPP. Across our dataset annual rates of GPP were positively correlated to watershed area (r = 0.69) and width (r = 0.49). The most productive regime, which we termed Summer Peak Rivers, were typically larger rivers with peaks in productivity coincident with peak incoming solar radiation during the summer. By contrast, Spring Peak Rivers were smaller, generally unproductive rivers that reached maximum rates early in the year before declining to near zero during the summer. These two regimes represent the most productive, and the most numerous segments within a network respectively. When we applied these results to the network scale we found that daily GPP for a river network exhibits a bimodal annual regime with seasonal peaks during the spring and summer. Although GPP in smaller rivers is relatively low, the cumulative influence of spring peaks in primary productivity across a large number of headwater rivers is substantial at the network scale. Despite peak GPP in the most productive reaches occurring during the summer, network scale GPP is often maximized during the spring vernal window in small-medium sized river networks.