CoChat

Parsimonious statistical learning models for low flow estimation

Open access
Laimighofer, Johannes, Melcher, Michael, Laaha, Gregor · 2021 · Hydrology and earth system sciences

Preprint — published as 10.5194/hess-26-129-2022

The DOI matches its Crossref record; 75 of the 75 references it cites were found in a registry or an index.

Abstract

Statistical learning methods offer a promising approach for low flow regionalization. We examine seven statistical learning models (lasso, linear and non-linear model based boosting, sparse partial least squares, principal component regression, random forest, and support vector machine regression) for the prediction of winter and summer low flow based on a hydrological diverse dataset of 260 catchments in Austria. In order to produce sparse models we adapt the recursive feature elimination for variable preselection and propose to use three different variable ranking methods (conditional forest, lasso and linear model based boosting) for each of the prediction models. Results are evaluated for the low flow characteristic Q95 (Pr(Q>Q95) = 0.95) standardized by catchment area using a repeated nested cross validation scheme. We found a generally high prediction accuracy for winter (R2CV of 0.66 to 0.7) and summer (R2CV of 0.83 to 0.86). The models perform similar or slightly better than a Top-kriging model that constitutes the current benchmark for the study area. The best performing models are support vector machine regression (winter) and non-linear model based boosting (summer), but linear models exhibit similar prediction accuracy. The use of variable preselection can significantly reduce the complexity of all models with only a small loss of performance. The so obtained learning models are more parsimonious, thus easier to interpret and more robust when predicting at ungauged sites. A direct comparison of linear and non-linear models reveals that non-linear relationships can be sufficiently captured by linear learning models, so there is no need to use more complex models or to add non-liner effects. When performing low flow regionalization in a seasonal climate, the temporal stratification into summer and winter low flows was shown to increase the predictive performance of all learning models, offering an alternative to catchment grouping that is recommended otherwise.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.5194/hess-2021-481

Cites

75 of 75 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Overview of the 260 gauging stations used in the study.
Figure 1 Overview of the 260 gauging stations used in the study.CC-BY · From the PDF
Figure 2 Absolute frequency (histogram and kernel density estimate) of summer q 95 and winter q 95 for all 260 stations.
Figure 2 Absolute frequency (histogram and kernel density estimate) of summer q 95 and winter q 95 for all 260 stations.CC-BY · From the PDF
Figure 3 Prediction error for the unseen test data (left panel) and the training data (right panel). Bootstrapping was performed 100-times where the training set consists of 70 % of the observations.
Figure 3 Prediction error for the unseen test data (left panel) and the training data (right panel). Bootstrapping was performed 100-times where the training set consists of 70 % of the observations.CC-BY · From the PDF
Figure 4 Variable selection for a GAM boosting model based on RMSE graphs. The red vertical line indicates the number of variables selected by a 5 % increase of the min( RMSE ) , the green line is the number of variables obtained by min( RMSE ) . The Black graph is the average over 10 inner CV-folds, and coloured graphs represent individual folds.
Figure 4 Variable selection for a GAM boosting model based on RMSE graphs. The red vertical line indicates the number of variables selected by a 5 % increase of the min( RMSE ) , the green line is the number of variables obtained by min( RMSE ) . The Black graph is the average over 10 inner CV-folds, and coloured graphs represent individual folds.CC-BY · From the PDF
Figure 5 Performance of 10 CV-runs of statistical learning models without variable preselection. Horizontal lines display the median of the CV-runs.
Figure 5 Performance of 10 CV-runs of statistical learning models without variable preselection. Horizontal lines display the median of the CV-runs.CC-BY · From the PDF
Figure 6 Reduction of R 2 CV is computed in respect to the model performance without any variable preselection. Each point is the median number of variables of one CV-run and one prediction model and the related R 2 CV . Each line shows (horizontal = Number of variables, vertical = Reduction in R 2 CV ) the upper and lower quartile for a variable ranking method. Colours can be found in Ram and Wickham (2018)
Figure 6 Reduction of R 2 CV is computed in respect to the model performance without any variable preselection. Each point is the median number of variables of one CV-run and one prediction model and the related R 2 CV . Each line shows (horizontal = Number of variables, vertical = Reduction in R 2 CV ) the upper and lower quartile for a variable ranking method. Colours can be found in Ram and Wickham (2018)CC-BY · From the PDF
Figure 7 Count of all variable rankings of all 1000-iterations on a log-scale for winter low flow. The ten best variables are listed after their averaged rank.
Figure 7 Count of all variable rankings of all 1000-iterations on a log-scale for winter low flow. The ten best variables are listed after their averaged rank.CC-BY · From the PDF
Figure 8 Count of all variable rankings of all 1000-iterations on a log-scale for summer low flow. The ten best variables are listed after their averaged rank.
Figure 8 Count of all variable rankings of all 1000-iterations on a log-scale for summer low flow. The ten best variables are listed after their averaged rank.CC-BY · From the PDF
Figure 9 ANE CV for summer and winter q 95 as a function of the aridity index (a), elevation (b), and catchment area (c). Only the best performing models are shown (winter SVM, summer GAM, both without variable preselection). The boxplots summarize the ANE CV averaged over all 10 CV-runs for each of the 260 stations.
Figure 9 ANE CV for summer and winter q 95 as a function of the aridity index (a), elevation (b), and catchment area (c). Only the best performing models are shown (winter SVM, summer GAM, both without variable preselection). The boxplots summarize the ANE CV averaged over all 10 CV-runs for each of the 260 stations.CC-BY · From the PDF
Image 10
CC-BY · From the PDF
Figure A2. Boxplots of the number of variables selected in each CV-fold (100-times).
Figure A2. Boxplots of the number of variables selected in each CV-fold (100-times).CC-BY · From the PDF
Figure A3. Count of all variable rankings of all 1000-iterations on a log-scale for summer low flow. The ten best variables are listed after their averaged rank.
Figure A3. Count of all variable rankings of all 1000-iterations on a log-scale for summer low flow. The ten best variables are listed after their averaged rank.CC-BY · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 75 references

Research works printed but not linked: 2

  1. https://doi.org/10.5194/hess-2021-481 Preprint. Discussion started: 22 September 2021 c Author(s) 2021. CC BY 4.0 License. · doi:10.5194/hess-2021-481preprint.discussionstarted:22september2021cauthor
  2. Wold, H.: Estimation of principal components and related models by iterative least squares, Multivariate analysis, pp. 391-420, 1966. https://doi.org/10.5194/hess-2021-481 Preprint. Discussion started: 22 September 2021 c Author(s) 2021. CC BY 4.0 License. · doi:10.5194/hess-2021-481preprint.discussionstarted:22september2021cauthor

Cited by

See all 30

Metadata

Authors
Laimighofer, J., Melcher, M., & Laaha, G.
Affiliations
  • BOKU University (AT)
  • FH JOANNEUM University of Applied Sciences (AT)
Corresponding author
  • Laimighofer, Johannes
  • Laaha, Gregor
Venue
Hydrology and earth system sciences
Published
2021-09-22
Volume
26
Issue
1
Pages
129-148
ISSN
1027-5606 (ISSN-L)
OpenAlex
W4225623739
Cited by
30
References
65
Citations by year
2026: 1 · 2025: 6 · 2024: 5 · 2023: 6 · 2022: 12
FWCI
2.12
Type
journal article
Language
en
Open access
Locations
  • Hydrology and earth system sciences (publishedVersion, cc-by)
  • FH JOANNEUM ePUB (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Österreichischen Akademie der Wissenschaften (25819)
Topics
  • Physical Sciences › Environmental Science › Water Science and Technology › Hydrology and Watershed Management Studies
  • Physical Sciences › Environmental Science › Environmental Engineering › Hydrological Forecasting Using AI
  • Physical Sciences › Environmental Science › Global and Planetary Change › Hydrology and Drought Analysis
Keywords
statistical learning models, recursive feature elimination, feature selection, feature ranking, support vector regression, sparse partial least squares, principal component regression, random forest, ungauged basins, seasonal stratification

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Parsimonious statistical learning models for low flow estimation. https://app.cochat.ai/papers/10.5194/hess-2021-481
The paper
Laimighofer, J., Melcher, M., & Laaha, G. (2021). Parsimonious statistical learning models for low flow estimation. Hydrology and earth system sciences. https://doi.org/10.5194/hess-2021-481

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.