@article{fdi:010073195,
  title = {{S}ensitivity and identifiability of hydraulic and geophysical parameters from streaming potential signals in unsaturated porous media},
  author = {{Y}ounes, {A}nis and {Z}aouali, {J}. and {L}ehmann, {F}. and {F}ahs, {M}.},
  editor = {},
  language = {{ENG}},
  abstract = {{F}luid flow in a charged porous medium generates electric potentials called streaming potential ({SP}). {T}he {SP} signal is related to both hydraulic and electrical properties of the soil. {I}n this work, global sensitivity analysis ({GSA}) and parameter estimation procedures are performed to assess the influence of hydraulic and geophysical parameters on the {SP} signals and to investigate the identifiability of these parameters from {SP} measurements. {B}oth procedures are applied to a synthetic column experiment involving a falling head infiltration phase followed by a drainage phase. {GSA} is used through variance-based sensitivity indices, calculated using sparse polynomial chaos expansion ({PCE}). {T}o allow high {PCE} orders, we use an efficient sparse {PCE} algorithm which selects the best sparse {PCE} from a given data set using the {K}ashyap information criterion ({KIC}). {P}arameter identifiability is performed using two approaches: the {B}ayesian approach based on the {M}arkov chain {M}onte {C}arlo ({MCMC}) method and the first-order approximation ({FOA}) approach based on the {L}evenberg-{M}arquardt algorithm. {T}he comparison between both approaches allows us to check whether {FOA} can provide a reliable estimation of parameters and associated uncertainties for the highly non-linear hydrogeophysical problem investigated. {GSA} results show that in short time periods, the saturated hydraulic conductivity ({K}-s) and the voltage coupling coefficient at saturation ({C}-sat) are the most influential parameters, whereas in long time periods, the residual water content (theta(s)), the {M}ualem-van {G}enuchten parameter (n) and the {A}rchie saturation exponent (n(a)) become influential, with strong interactions between them. {T}he {M}ualem-van {G}enuchten pa-rameter (alpha) has a very weak influence on the {SP} signals during the whole experiment. {R}esults of parameter estimation show that although the studied problem is highly nonlinear, when several {SP} data collected at different altitudes inside the column are used to calibrate the model, all hydraulic ({K}-s, theta(s), alpha, n) and geophysical parameters (n(a),{C}-sat) can be reasonably estimated from the {SP} measurements. {F}urther, in this case, the {FOA} approach provides accurate estimations of both mean parameter values and uncertainty regions. {C}onversely, when the number of {SP} measurements used for the calibration is strongly reduced, the {FOA} approach yields accurate mean parameter values (in agreement with {MCMC} results) but inaccurate and even unphysical confidence intervals for parameters with large uncertainty regions.},
  keywords = {},
  booktitle = {},
  journal = {{H}ydrology and {E}arth {S}ystem {S}ciences},
  volume = {22},
  numero = {7},
  pages = {3561--3574},
  ISSN = {1027-5606},
  year = {2018},
  DOI = {10.5194/hess-22-3561-2018},
  URL = {https://www.documentation.ird.fr/hor/fdi:010073195},
}