Premium
Kayak drifter surface velocity observation for 2D hydraulic model validation
Author(s) -
Barker J.R.,
Pasternack G.B.,
Bratovich P.M.,
Massa D.A.,
Wyrick J.R.,
Johnson T.R.
Publication year - 2018
Publication title -
river research and applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.679
H-Index - 94
eISSN - 1535-1467
pISSN - 1535-1459
DOI - 10.1002/rra.3238
Subject(s) - drifter , kinematics , global positioning system , software , tracking (education) , environmental science , computer science , marine engineering , simulation , geodesy , remote sensing , geology , mathematics , engineering , psychology , telecommunications , pedagogy , physics , classical mechanics , lagrangian , mathematical physics , programming language
Advances in remote sensing, informatics, software, and microprocessors enable meter‐resolution two‐dimensional (2D) hydrodynamic models that produce nearly a census of ecohydraulic conditions over long river segments with 10 5 to 10 8 computational elements. It is difficult to test statistical and spatial model performance at such scope using fixed‐point velocity measurements, because field methods are so expensive, laborious, slow, and restricted by safety factors. This study evaluated low‐cost water surface particle tracking by kayak with real‐time kinematic GPS for 2D model validation using 7.2 km of the lower Yuba River in California. Observed flows were between 15 to 140 m 3 /s, which were in‐channel up to and including bankfull conditions. The coefficients of determination between 5,780 observations and 2D model predictions were 0.79 and 0.80 for velocity magnitude and direction, respectively. When surface speed was downscaled and compared to modelled depth‐averaged velocity, median unsigned difference was 15.5%. Standard hydrological model performance metrics affirmed satisfactory validation. Surface tracking provided the novel benefit of enabling validation of velocity direction, and that testing found satisfactory performance using all metrics. Having 10 to 1,000 times more data enables robust statistical testing and spatial analysis of both speed and direction, which outweighs the loss of depth‐averaged data. Both fixed‐point and kayak particle tracking methods are useful tools to help evaluate 2D model performance.