Preprints
https://doi.org/10.5194/hess-2024-47
https://doi.org/10.5194/hess-2024-47
20 Feb 2024
 | 20 Feb 2024
Status: a revised version of this preprint was accepted for the journal HESS and is expected to appear here in due course.

Learning Landscape Features from Streamflow with Autoencoders

Alberto Bassi, Marvin Höge, Antonietta Mira, Fabrizio Fenicia, and Carlo Albert

Abstract. Understanding the number and types of signatures that best describe streamflow time series is a crucial objective in hydrological science, serving applications such as catchment classification, hydrological model development and calibration. With the main objective of learning a minimal number of streamflow features, we employ an explicit noise conditional autoencoder (ENCA), which, together with meteorological forcings, allows for an accurate reconstruction of the whole streamflow time series. The ENCA architecture feeds the meteorological forcing to the decoder in order to incentivize the encoder to only learn features that are related to landscape properties. By isolating the effect of meteorology, these features can thus be interpreted as landscape fingerprints. The optimal number of features is found by means of an intrinsic dimension estimator. We train our model on the hydro-meteorologic time series data of 568 catchments of the continental United States from the CAMELS dataset. We compare the reconstruction accuracy with state-of-the-art models that take as input a subset of static catchment attributes (both climate and landscape attributes) along with the meteorological forcing variables. Our results suggest that available static catchment attributes compiled by experts account for almost all the relevant information about the rainfall-runoff relationship. Yet, these catchment attributes can be summarized by only two relevant learnt features (or signatures), while a third one is needed for about a dozen difficult catchments in the central US, mainly characterized by high aridity index and intermittent flow. The principal components of the learnt features strongly correlate with the baseflow index and aridity indicators, which is consistent with the idea that these indicators capture the variability of catchment hydrological response and relate to needed model complexity. The correlation analysis further indicates that soil-related and vegetation attributes are of high importance. Finally, in the attempt to interpret the learnt catchment features, we relate them to typical hydrological model components, with specific reference to the parameters of the GR4J model and their function on the hydrograph.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Alberto Bassi, Marvin Höge, Antonietta Mira, Fabrizio Fenicia, and Carlo Albert

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on hess-2024-47', John Ding, 22 Feb 2024
    • AC1: 'Reply on CC1', Alberto Bassi, 21 Mar 2024
  • RC1: 'Comment on hess-2024-47', Anonymous Referee #1, 15 Mar 2024
    • AC2: 'Reply on RC1', Alberto Bassi, 03 Apr 2024
  • CC2: 'Comment on hess-2024-47', Weibo Liu, 16 Mar 2024
    • AC3: 'Reply on CC2', Alberto Bassi, 03 Apr 2024
  • RC2: 'Comment on hess-2024-47', Anonymous Referee #2, 18 Mar 2024
    • AC4: 'Reply on RC2', Alberto Bassi, 03 Apr 2024
  • RC3: 'Comment on hess-2024-47', Daniel Klotz, 22 Mar 2024
    • RC4: 'Short addendum', Daniel Klotz, 24 Mar 2024
      • AC5: 'Reply on RC3', Alberto Bassi, 09 Apr 2024
    • AC5: 'Reply on RC3', Alberto Bassi, 09 Apr 2024
    • AC6: 'Reply on RC3', Alberto Bassi, 09 Apr 2024

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on hess-2024-47', John Ding, 22 Feb 2024
    • AC1: 'Reply on CC1', Alberto Bassi, 21 Mar 2024
  • RC1: 'Comment on hess-2024-47', Anonymous Referee #1, 15 Mar 2024
    • AC2: 'Reply on RC1', Alberto Bassi, 03 Apr 2024
  • CC2: 'Comment on hess-2024-47', Weibo Liu, 16 Mar 2024
    • AC3: 'Reply on CC2', Alberto Bassi, 03 Apr 2024
  • RC2: 'Comment on hess-2024-47', Anonymous Referee #2, 18 Mar 2024
    • AC4: 'Reply on RC2', Alberto Bassi, 03 Apr 2024
  • RC3: 'Comment on hess-2024-47', Daniel Klotz, 22 Mar 2024
    • RC4: 'Short addendum', Daniel Klotz, 24 Mar 2024
      • AC5: 'Reply on RC3', Alberto Bassi, 09 Apr 2024
    • AC5: 'Reply on RC3', Alberto Bassi, 09 Apr 2024
    • AC6: 'Reply on RC3', Alberto Bassi, 09 Apr 2024
Alberto Bassi, Marvin Höge, Antonietta Mira, Fabrizio Fenicia, and Carlo Albert

Data sets

CAMELS: Catchment Attributes and Meteorology for Large-sample Studies A. Newman, K. Sampson, M. P. Clark, A. Bock, R. J. Viger, and D. Blodgett https://ral.ucar.edu/solutions/products/camels

Alberto Bassi, Marvin Höge, Antonietta Mira, Fabrizio Fenicia, and Carlo Albert

Viewed

Total article views: 885 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
659 173 53 885 34 36
  • HTML: 659
  • PDF: 173
  • XML: 53
  • Total: 885
  • BibTeX: 34
  • EndNote: 36
Views and downloads (calculated since 20 Feb 2024)
Cumulative views and downloads (calculated since 20 Feb 2024)

Viewed (geographical distribution)

Total article views: 862 (including HTML, PDF, and XML) Thereof 862 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 11 Nov 2024
Download
Short summary
The goal is to remove the impact of meteorological drivers in order to uncover the unique landscape fingerprints of a catchment from streamflow data. Our results reveal an optimal two-feature summary for most catchments, with a third feature needed for challenging cases, associated with aridity and intermittent flow. Baseflow index, aridity, and soil/vegetation attributes strongly correlate with learned features, indicating their importance for streamflow prediction.