Preconditioning an ensemble Kalman filter for groundwater flow using environmental-tracer observations

摘要: Abstract Groundwater resources management requires operational, regional-scale groundwater models accounting for dominant spatial variability of aquifer properties and spatiotemporal recharge. We test the Ensemble Kalman filter (EnKF) to estimate transient hydraulic heads recharge, as well conductivity specific-yield distributions a virtual phreatic aquifer. To speed up computation time, we use coarsened grid in simulations, reconstruct head measurements at observation points by local model vicinity piezometer part operator. show that EnKF can adequately both mean patterns when assimilating daily values from highly variable initial sample. The also temporally recharge satisfactory level, long ensemble size is large enough. Constraining parameters on concentrations groundwater-age tracers (here: tritium) hydraulic-head observations cannot reasonably be done because depend history over longer times while have much shorter temporal support. thus different method, Generator (KEG), precondition tracer data time-averaged values. preconditioned exhibits smaller spread improved means patterns. preconditioning improves particularly sizes, allowing operational assimilation with reduced computational effort. In validation scenario delineating protection zones, performs clearly better than using original ensemble.