the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Revisiting parameter sensitivities in the Variable Infiltration Capacity model
- 1Department of Civil Engineering, Universidad de Chile, Santiago, Chile
- 2Advanced Mining Technology Center, Universidad de Chile, 5 Santiago, Chile
- 3National Center for Atmospheric Research, Boulder, CO, USA
- 1Department of Civil Engineering, Universidad de Chile, Santiago, Chile
- 2Advanced Mining Technology Center, Universidad de Chile, 5 Santiago, Chile
- 3National Center for Atmospheric Research, Boulder, CO, USA
Abstract. Despite the Variable Infiltration Capacity (VIC) model being used for decades in the hydrology community, there are still model parameters whose sensitivities remain unknown. Additionally, understanding the factors that control spatial variations in parameter sensitivities is crucial given the increasing interest to obtain spatially coherent parameter fields over large domains. In this study, we investigate the sensitivities of 43 soil, vegetation and snow parameters in the VIC model for 101 catchments spanning the diverse hydroclimates of continental Chile. We implement a hybrid local-global sensitivity analysis approach, using eight model evaluation metrics to quantify sensitivities, with four of them formulated from runoff time series; two characterizing snow processes, and the remaining two based on evaporation processes. Our results confirm an over-parameterization for the processes analysed here, with only 12 (i.e., 28 %) parameters found as sensitive, distributed among soil (7), vegetation (2) and snow (3) model components. Correlation analyses show that climate variables – in particular, mean annual precipitation and aridity index – are the main controls on parameter sensitivities. Additionally, our results highlight the influence of the leaf area index on simulated hydrologic processes – regardless on the dominant climate types – and the relevance of hard-coded snow parameters. Based on correlation results and the interpretation of spatial sensitivity patterns, we provide guidance on the most relevant parameters for model calibration according to the target processes and the prevailing climate type. Overall, the results presented here contribute to improved understanding of model behaviour across watersheds with diverse physical characteristics that encompass a wide hydroclimatic gradient from hyper-arid to humid systems.
Ulises Sepúlveda et al.
Status: closed
-
RC1: 'Comment on hess-2021-550', Anonymous Referee #1, 14 Dec 2021
This paper studies the sensitivity of the VIC model to 43 soil, vegetation, and snow parameters using the DELSA sensitivity analysis method in about 5500, 0.05 degree grid cells in Chile. The authors find that different goodness-of-fit metrics (the authors try eight of these) have more or less sensitivity to different parameters. Precipitation and aridity are also found to control parameter sensitivity. Recommendations are provided for how to select calibration parameters based on climatology and which variable is of interest.
Summary of findings:
- VIC is overparameterized - only 12 of the 43 parameters are sensitive – 7 soil parameters, 2 vegetation parameters, and 3 snow parameters.
- Mean annual precipitation and aridity control which parameters are sensitive.
- Leaf area index and hard-coded snow parameters are sensitive.
- Provides guidance on the most relevant parameters for model calibration depending on the target process (runoff, snow, or ET) and climate type (humid/arid).
This paper is well-written, and I think it will be a useful resource for VIC modelers, as it describes the VIC parameters in depth, where they came from, and what good min/max values are for calibration. Its point that some of the hard-coded snow model parameters are sensitive and perhaps should be exposed to users is well-taken.
My main criticism of this paper is that some of the findings, e.g. that VIC is overparameterized and certain parameters are more sensitive under certain conditions – are well-known from other studies (such as Demaria et al., 2007 and Gou et al., 2020). On the other hand, examining parameter interactions, which the authors say is possible using the DELSA method, might be more interesting.
Some other critiques:
- I wonder whether the amount of peak SWE is a useful goodness-of-fit parameter. I imagine that many combinations of parameters could give the same peak SWE, since it is an integrated measure of the entire season’s snowfall. This would be worth mentioning in Section 3.4 performance metrics.
- You study 101 catchments throughout Chile, but only a few catchments are highlighted in the figures. Is there any justification for why these catchments are spotlighted?
- In Figure 3, indicate whether the rows are organized by latitude (they appear to be, but it would help readers interpret the figure if this were more clear).
- L315: SnowLegth typo
- In Section 4.3.1, it is claimed that humid environments enable recharge of the lower soil layers and thus cause increased sensitivity of baseflow parameters. Is this true for all wet catchments, or does it depend whether precipitation occurs as rainfall or snowfall? I imagine that snowy catchments will have more sensitivity to baseflow parameters, as water will have more time to infiltrate into the soils. (I’m assuming here that snowmelt runoff is generated more gradually than rainfall runoff.)
- L380: syntax should be “LAI, Rmin, etc. being the most important parameters.”
- AC1: 'Reply on RC1', Pablo Mendoza, 07 Feb 2022
-
RC2: 'Comment on hess-2021-550', Anonymous Referee #2, 23 Dec 2021
The authors assessed application results of the VIC model on 101 basins in Chile to test how sensitive model results were to variations in 43 calibration parameters. They explained how some parameters could be adjusted by model users and others were hidden inside the coding of the model. Their results showed that 12 parameters exhibited significant sensitivity in soil, vegetation, and snow input variables. Their work seems to provide insight into the inner workings of the model and to contribute useful guidance for future advancement in the popular VIC model.
The paper is long and detailed with an extensive set of graphs and tables that will be of interest to modelers working on the finest details of the VIC model. The writing and presentation are excellent.
The paper would be more useful to hydrologists and modelers who are not part of the mainline VIC users if the authors would include brief information about the model’s history and accessibility. Very little information of this nature is included. See Line 142 as an example of how the model is introduced.
- AC2: 'Reply on RC2', Pablo Mendoza, 07 Feb 2022
Status: closed
-
RC1: 'Comment on hess-2021-550', Anonymous Referee #1, 14 Dec 2021
This paper studies the sensitivity of the VIC model to 43 soil, vegetation, and snow parameters using the DELSA sensitivity analysis method in about 5500, 0.05 degree grid cells in Chile. The authors find that different goodness-of-fit metrics (the authors try eight of these) have more or less sensitivity to different parameters. Precipitation and aridity are also found to control parameter sensitivity. Recommendations are provided for how to select calibration parameters based on climatology and which variable is of interest.
Summary of findings:
- VIC is overparameterized - only 12 of the 43 parameters are sensitive – 7 soil parameters, 2 vegetation parameters, and 3 snow parameters.
- Mean annual precipitation and aridity control which parameters are sensitive.
- Leaf area index and hard-coded snow parameters are sensitive.
- Provides guidance on the most relevant parameters for model calibration depending on the target process (runoff, snow, or ET) and climate type (humid/arid).
This paper is well-written, and I think it will be a useful resource for VIC modelers, as it describes the VIC parameters in depth, where they came from, and what good min/max values are for calibration. Its point that some of the hard-coded snow model parameters are sensitive and perhaps should be exposed to users is well-taken.
My main criticism of this paper is that some of the findings, e.g. that VIC is overparameterized and certain parameters are more sensitive under certain conditions – are well-known from other studies (such as Demaria et al., 2007 and Gou et al., 2020). On the other hand, examining parameter interactions, which the authors say is possible using the DELSA method, might be more interesting.
Some other critiques:
- I wonder whether the amount of peak SWE is a useful goodness-of-fit parameter. I imagine that many combinations of parameters could give the same peak SWE, since it is an integrated measure of the entire season’s snowfall. This would be worth mentioning in Section 3.4 performance metrics.
- You study 101 catchments throughout Chile, but only a few catchments are highlighted in the figures. Is there any justification for why these catchments are spotlighted?
- In Figure 3, indicate whether the rows are organized by latitude (they appear to be, but it would help readers interpret the figure if this were more clear).
- L315: SnowLegth typo
- In Section 4.3.1, it is claimed that humid environments enable recharge of the lower soil layers and thus cause increased sensitivity of baseflow parameters. Is this true for all wet catchments, or does it depend whether precipitation occurs as rainfall or snowfall? I imagine that snowy catchments will have more sensitivity to baseflow parameters, as water will have more time to infiltrate into the soils. (I’m assuming here that snowmelt runoff is generated more gradually than rainfall runoff.)
- L380: syntax should be “LAI, Rmin, etc. being the most important parameters.”
- AC1: 'Reply on RC1', Pablo Mendoza, 07 Feb 2022
-
RC2: 'Comment on hess-2021-550', Anonymous Referee #2, 23 Dec 2021
The authors assessed application results of the VIC model on 101 basins in Chile to test how sensitive model results were to variations in 43 calibration parameters. They explained how some parameters could be adjusted by model users and others were hidden inside the coding of the model. Their results showed that 12 parameters exhibited significant sensitivity in soil, vegetation, and snow input variables. Their work seems to provide insight into the inner workings of the model and to contribute useful guidance for future advancement in the popular VIC model.
The paper is long and detailed with an extensive set of graphs and tables that will be of interest to modelers working on the finest details of the VIC model. The writing and presentation are excellent.
The paper would be more useful to hydrologists and modelers who are not part of the mainline VIC users if the authors would include brief information about the model’s history and accessibility. Very little information of this nature is included. See Line 142 as an example of how the model is introduced.
- AC2: 'Reply on RC2', Pablo Mendoza, 07 Feb 2022
Ulises Sepúlveda et al.
Ulises Sepúlveda et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 616 | 188 | 11 | 815 | 5 | 4 |
- HTML: 616
- PDF: 188
- XML: 11
- Total: 815
- BibTeX: 5
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1