This paper presents new estimates of solute fluxes from five high Arctic glacier basins in Svalbard. These estimates are combined with data from two other glacier basins to assess the effectiveness of chemical denudation on Svalbard and to estimate rates of temporary (or transient) CO2 drawdown. We use a solute provenance model to partition solutes into marine, aerosol, atmospheric and crustal components and to estimate their annual fluxes. Crustally derived solute fluxes are equivalent to a mean chemical denudation rate of 350 Σmeq+ m−2 a−1 for Svalbard (range: 160–560 Σmeq+ m−2 a−1), which lies within the global range of 94–4200 Σmeq+ m−2 a−1 for 21 glacier basins in the northern hemisphere, and is close to the continental average of 390 Σmeq+ m−2 a−1. Specific annual discharge is the most significant control upon chemical denudation in the glacierized basins, and basin lithology is an important secondary control, with carbonate-rich and basaltic lithologies currently showing the greatest chemical denudation rates. Estimates of transient CO2 drawdown are also directly associated with specific annual discharge and rock type. On Svalbard transient CO2 drawdown lies in the range 110–3000 kg C km−2 a−1, whilst the range is 110–13000 kg C km−2 a−1 for the northern hemisphere glacial data set. Transient CO2 drawdown is therefore usually low in the Svalbard basins unless carbonate or basalt rocks are abundant. The analysis shows that a large area of uncertainty in the transient CO2 drawdown estimates exists due to the non-stoichiometric release of solute during silicate hydrolysis. Silicate hydrolysis is particularly non-stoichiometric in basins where the extent of glacierization is high, which is most probably an artefact of high flushing rates through ice-marginal and subglacial environments where K-feldspars are undergoing mechanical comminution. Copyright © 2000 John Wiley & Sons, Ltd.