top of page

Take a look at all the sensibilities displaced together, presented earlier.

See also that observed fluxes for drains, streams and recharge has being suitably achieved by this model realization, (including a 27% rate between RCH-EVPT), as well as the attained RMSE criteria for the 18 Hydraulic Heads.

This problem end solution comes next, but not yet, indeed. Not before a small secret, in consonance with the applied routines: - Only a very good weighting strategy allows PEST the lowering the objective function (φ).

Calibracao

Betami weighting strategy: From STD analysis of all observations and its confidence intervals Z(%)

10.PNG

Meeting the RMSE unavoidable criteria – Calibration and Mass Balance.

11.PNG

Calibrated outputs and particle tracking: - A model ready to be used for a demanded prediction.

Be that as it may, the fact remains that no formal uncertainty analysis has been made (at least not yet), what does not discredit the present solution - given the full report of all the modelling proceedings.

This paper presents “kind of primary assumptions” of a much more extended subject: The PEST calibration. Here we have not only, but “just” a quite thoroughly analysis of inputs needed, and outputs expected.  

[Yes, undoubtedly more work remains to be done]. The reader is very welcome to contest and expand the subject. All the data needed can be found in archives of the cited GMDSI Tutorials. We also suggest the https://pesthomepage.org/ Although, don't hesitate being back for some prompt available examples at https://www.betami.com.br/01-geostatistics Or for a little more of this example at https://www.betami.com.br/case-c

All things considered, remains self-evident the importance of such judicious and assertive analysis, regardless the size or the model application per si. The extension of which will depend on the inherent modelling objectives.

Disclaimer:

14.PNG

Betami comes from MODFLOW | T-PROGS concepts described in the master's thesis: Markov chains applied to the definition of lithological layers for numerical models of groundwater flow.

We conduct our analyses using our own algorithm for pre-and-post-treatment of numerically processed data.

Our primary focus is on the accurate visualization of each and all aspect of your database.

Do you have a suitable database? Hand me over.

I can recast you model for a fraction of the market price, in one week (!).

Versão completa desta página  apenas em formato Desktop.
Content fully developed only in desktop format.
bottom of page