COS 177-7 - Regional and global controls and potential significance of dissolved silica retention in lakes and reservoirs

Friday, August 10, 2012: 10:10 AM
F150, Oregon Convention Center
John A. Harrison1, Patrick Frings2, Arthur Beusen3, Daniel Conley2 and Michelle McCrackin4, (1)School of Earth and Environmental Sciences, Washington State University Vancouver, Vancouver, WA, (2)Department of Geology, Lund University, (3)Netherlands Environmental Assessment Agency, (4)School of Earth and Environmental Sciences, Washington State University
Background/Question/Methods

Dissolved silica (DSi) availability can exert a strong influence on phytoplankton community structure and composition in freshwaters and the coastal zone.  Recent work suggests that through reservoir construction, humans have dramatically altered the efficiency with which dissolved silica is transported downstream through watersheds to the coastal zone.  Lentic water bodies (lakes and reservoirs) offer favorable conditions for DSi burial in sediments.  However, the patterns and controls of lentic DSi removal at regional to global scales remain largely unexplored and un-quantified.  In this paper we describe, evaluate, and apply a new, spatially explicit, annual-scale, global model of lentic DSi removal called SiRReLa (Silica Retention in Reservoirs and Lakes) and combine results of SiRReLa with results of a previously published model of TN retention by lentic systems (NiRReLa) to develop a global view of regions where lentic nutrient trapping could significantly distort N and Si availability.

Results/Conclusions

We found that lentic DSi removal (kg Si yr-1) is significantly and positively related to DSi loading (P < 0.0001; r2 = 0.96) and that DSi removal efficiency is significantly and positively related to system water residence time (P <0.0001; r2=0.65).  In addition, greater DSi uptake rates were associated with eutrophic systems than with non-eutrophic systems (median Vf: 11.1 and 1.7 m y-1 for eutrophic and non-eutrophic systems, respectively; P < 0.01).  The SiRReLa model, which is based on these insights, performed well in predicting both total DSi removal (kg Si yr-1) by individual lakes ( Nash-Sutcliffe Efficiency (NSE): 0.96) and DSi removal efficiency (% Si removed; NSE: 0.80), with no detectable bias in the model.  Global model runs indicate: 1) lentic systems are important sinks for DSi and are conservatively estimated to remove 19-38% of all DSi inputs to surface waters (227-466 Tg DSi yr-1); 2) small lakes and reservoirs (< 50 km2) are critical traps for DSi, retaining 81% (72 Tg DSi yr-1) of the global total; and 3) although reservoirs occupy just 6% of the global lentic surface area they retained approximately 35% of the total DSi removed by lentic systems.