Integrating fluvial processes into the global Si cycle

Sammanfattning: Silicon (Si) is ubiqutious in natural waters as dissolved silica (DSi) and is used as a nutrient by many plant and animal species. DSi ultimately derives from the weathering of silicate minerals, a process which also consumes atmospheric CO2 and couples the silicon and carbon cycles. These processes, among others, impart observable fractionations to the stable isotopes of silicon, which makes the Si isotopic composition (expressed in per mil as δ30Si) a useful tracer. The most important term in the global Si balance is the flux of Si from land to oceans, which has interest both as the integrative product of the continental cycle and as the input to the ocean cycle. The vast majority of this flux is delivered via the fluvial system as both DSi and particulate silica phases. The fluvial system can be conceptualised as a series of filters, including alluvial deposits, floodplains and wetlands, lakes and reservoirs, and the hyporheic and ripararian zones, and each of these has the potential to modify the river Si flux in ways that are not well understood. Overall, this thesis aims to improve the understanding of processes that can alter the magnitude, form and isotopic composition of the river Si flux and is based on a synthesis of literature data and the results from field campaigns in three large watersheds – the Okavango, the Ganges and the Paraná. Based on >500 published river DSi δ30Si values, including those developed in this thesis, the first order control on δ30Si is inferred to be the degree of incorporation of Si into secondary phases, including biogenic silica structures. River water DSi δ30Si values have an unweighted mean = +1.38‰ and are consistently above the δ30Si of their parent material. This thesis suggests that these values can be exploited to develop a metric of silicate weathering intensity, and therefore to correct observed DSi fluxes and derive initial silicate weathering rates. Within the fluvial system there is a substantial reduction of the dissolved silica flux mediated by the development of a pool of silica sequestered in lake and reservoir sediments. This is a semi-permanent sink. There are similar pools in soils, alluvial deposits and wetlands that can act as sources or sinks. Together, the papers underscore the importance of incorporating biological processes into a mechanistic understanding of the global Si cycle. Finally, while particulate silicates carried by rivers are typically assumed to be relatively inert, we were able to show that the fraction of the sediment liable to release DSi when it enters seawater may be larger than previously thought. Overall, the fluvial filter is not just a simple conduit but instead actively processes Si in a way that needs to be better understood in order to understand and predict the magnitude, form and isotopic composition of Si fluxes from land to ocean.