HydroMix v1.0: a new Bayesian mixing framework for attributing uncertain hydrological sources

Beria, Harsh; Larsen, Joshua R.; Michelon, Anthony; Ceperley, Natalie C.; Schaefli, Bettina

doi:https://doi.org/10.5194/gmd-13-2433-2020

Articles | Volume 13, issue 5

https://doi.org/10.5194/gmd-13-2433-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-13-2433-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 13, issue 5

Model description paper

|

27 May 2020

Model description paper |

| 27 May 2020

HydroMix v1.0: a new Bayesian mixing framework for attributing uncertain hydrological sources

Harsh Beria, Joshua R. Larsen, Anthony Michelon, Natalie C. Ceperley, and Bettina Schaefli

Download

Final revised paper (published on 27 May 2020)
Preprint (discussion started on 28 Mar 2019)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

SC1: 'Executive editor comment, code and data availability.', David Ham, 15 May 2019
- AC1: 'Response to the executive editor', Harsh Beria, 21 May 2019
RC1: 'Review of "HydroMix v1.0: a new Bayesian mixing framework for attributing uncertain hydrological sources"', Anonymous Referee #1, 26 Jun 2019
- AC4: 'Response to Anonymous reviewer #1', Harsh Beria, 19 Jul 2019
RC2: 'Reviewer comment', Anonymous Referee #2, 02 Jul 2019
- AC2: 'Response to Anonymous reviewer #2', Harsh Beria, 17 Jul 2019
RC3: 'Revision', Anonymous Referee #3, 10 Jul 2019
- AC3: 'Response to Anonymous reviewer #3', Harsh Beria, 17 Jul 2019

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Harsh Beria on behalf of the Authors (07 Oct 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (25 Oct 2019) by Bethanna Jackson

RR by Anonymous Referee #1 (06 Nov 2019)

RR by Anonymous Referee #4 (27 Nov 2019)

Suggestions for revision or reasons for rejection

Review of “HydroMix v1.0: a new Bayesian mixing framework for attributing uncertain hydrological 1 sources” by Harsh Beria et al. (manuscript ID gmd-2019-69, revised submission)

The manuscript presents an end-member mixing-model analysis using Bayesian methods. The presented approach directly addresses uncertainty introduced by small sample sizes in combination with heterogeneity of the source composition in a Markov-chain Monte Carlo framework.
This is a revised submission, in which the authors addressed the main concerns of the reviewers from the first round. The manuscript is well structured and explains the problem and approach well. However, at times the wording is imprecise (i.e., the authors say “sources” when they actually mean “source concentrations”, and see further comments below). I recommend improving this for better clarity of the text. I recommend publication subject to technical corrections.

Minor comments:
p.1, l.17: you fit a pdf to the source concentrations, not to the sources themselves.
p.2, l.32: variability refers to spatial and temporal variability I assume. Please specify.
p.3, l.13: concentrations are “distributed heterogeneously in space”, rather than “spatially distributed”
p.3, l.14-17: grammar of the sentence is incorrect. It should be: “…if we were to characterize…., we would need to capture…”
p.4, l.16/17: “tracer concentration” rather than “concentration of tracer” (two instances)
p.7, l.20: I assume “sources” in this case actually refers to the concentrations of these sources, rather than the sources themselves. Please be precise.
p.7, l.44: “stable isotopes of water in deuterium”: deuterium IS one of the stable water isotopes, it is not IN the stable water
p.8, Eq. 15: Is P_s (snow precip) defined previously?
p.8, l. 31: not explicitly considered, or excluded?
p.12, l.21: is this the correct reference? I believe this has been known much longer, e.g. since Dansgaard (1961).
Fig 8: do I understand this figure correctly, that d2H and d18O result in different predictions of the fraction of snow in groundwater? Is this because the isotopic composition of the two sources (GW, Snow) is less distinct for one tracer than for the other?
p.21, l19: what is a “normal model error”? A normally-distributed one? Please use precise terminology.

References
Dansgaard, W.: The isotopic composition of natural waters withspecial reference to the Greenland ice cap, Medd. Groenland,165, 120, 1961.

Hide

ED: Reconsider after major revisions (12 Dec 2019) by Bethanna Jackson

AR by Harsh Beria on behalf of the Authors (17 Mar 2020) Author's response Manuscript

ED: Publish as is (27 Apr 2020) by Bethanna Jackson

Short summary

We develop a Bayesian mixing model to address the issue of small sample sizes to describe different sources in hydrological mixing applications. Using composite likelihood functions, the model accounts for an often overlooked bias arising due to unweighted mixing. We test the model efficacy using a series of statistical benchmarking tests and demonstrate its real-life applicability by applying it to a Swiss Alpine catchment to obtain the proportion of groundwater recharged from rain vs. snow.