the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Jul 2025
Using evapotranspiration signatures to assess evapotranspiration realism in rainfall-runoff models
Abstract. Conceptual rainfall-runoff models are typically evaluated using streamflow data. Many studies have shown the benefits of moving from a conventional single-objective function to consideration of multiple signatures of streamflow behaviour, allowing more specific diagnosis of model deficiencies. In this study, we extend this approach beyond streamflow, using actual evapotranspiration (AET) signatures to assess AET dynamics in traditionally calibrated conceptual rainfall-runoff models. We calibrated models to streamflow only and separately to streamflow jointly with AET, then evaluated AET realism using AET signatures defined across various temporal scales. This was done for five models at 14 Australian sites spanning a variety of conditions, with each site co-located with a flux tower. Our results show that incorporating AET data into calibration significantly improves aspects of AET dynamics in models, particularly monthly variability and the degree to which AET and potential evapotranspiration are synchronous or asynchronous. This improvement extends even to the independent evaluation period in split sample testing. However, other signatures were not well improved, including aspects of seasonal and event-scale timing, in addition to interannual variability. Future research could explore a wider range of calibration strategies to assess whether these deficiencies can be calibrated away or are inherent to the models. Overall, these findings suggest that commonly used conceptual rainfall-runoff models struggle with many aspects of AET dynamics, even when AET information is included in the calibration. We recommend that future model evaluations examine a wider range of measures, aiming to characterise performance against non-streamflow variables in a more holistic manner.
- Preprint
(2550 KB) - Metadata XML
-
Supplement
(3441 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-3373', Anonymous Referee #1, 03 Nov 2025
-
RC2: 'Comment on egusphere-2025-3373', Anonymous Referee #2, 02 Dec 2025
This paper by Gardiya Weligamage et al. deals with the issue of evapotranspiration in conceptual rainfall-runoff models. This is a very important topic for the community of hydrological modelling. Evapotranspiration is in general overlooked, although it represents a significant (and sometimes dominant) term of catchment water balance. To this date, only very few studies have proposed a proper evaluation of evapotranspiration in models, so the approach described here is extremely promising. The paper is also well organised and written very clearly. A companion paper, also submitted to HESS, presents the AET signatures in more details, but enough information is given here so that the present paper can be read independently, which is appreciated.
1/ However, given the title of the paper and the main part of the introduction, I was expecting a study on the evaluation of realism of AET in the tested models, containing interpretations and conclusions about these models and their structures. Unfortunately, apart from brief comments here and there, nothing such appears in the paper, that turns out to be a simple (though very carefully thought-out and well executed) comparison study of two calibration methods (with streamflow only of with streamflow + AET data), which is much more limited in terms of interest and impact. This is really regrettable, and I think it would not be a big amount of work to add these elements. In particular:
- Methods section: when describing the models used (p7, l 142) and mentioning their “various complexities with various evapotranspiration representations”: describe them + add a table to sum up how the calculations of AET are made in the models.
- Results section : develop comments and interpretation according to model structures and AET calculations, for example l 250, l 290, l 319-320
- Add a full paragraph in the Discussion section
2/ Another discussion point that seems overlooked is the interest of the AET signatures for model assessment. Some of the signatures seem to be not discriminant between models for example, which raises questions about their added value (see Fig 11: annual median, water stress in particular). I believe this point could be developed in the paper as well.
3/ Figures could also be improved. A lot of them (Figs 3 to 10) are very similar and almost indistinguishable. Some of them don’t seem to bring a lot of information and could probably be transferred to supplementary material (Figs 4, 5, 6). For all of them, the symbols for the different models are indistinguishable, which limits the interpretation of AET deficiencies (eg in Fig 6: there are comments about some of the models but they can’t really be seen on the graphs). On the other hand, the colours for the catchments don’t seem to be very useful, as they are almost never discussed / commented. And finally, as the point of interest is the comparison between two calibration methods, maybe just merging both graphs into one representing the difference between the calibration methods would be more synthetic and straightforward?
These remarks apply also to Figure 12.
4/ additional remarks
- I am surprised by the use of a single flux tower for a 4800 km2 catchment (p3, l90). There are some considerations about the spatial representativeness of the flux tower data (section 2.1 + supplementary material), but they address a slightly different question as they present a comparison between nearest neighbour flux tower data and remotely sensed data, without consideration for the catchments themselves (no mention of the catchments can be found in Table S1). There is some important information missing, in particular about the climate characteristics of the catchments (climate zones such as Koppen-Geiger as mentioned in Table S1 or any relevant classification), and the dominant vegetation. Representativeness of the flux tower data over such large distances depends on the spatial variability of climate and vegetation in the study area. This is important for the potential generalization of the method. How would it apply to another part of the world?
- PET is never properly defined (I don’t think that the acronym is even developed once). I don’t think that readers can take the definition for granted, as there is a bit of confusion around this term with several definitions. For example Allen et al (1998) in the FAO reference define ET0 (= reference evapotranspiration) rather than PET. Is it the same?
- performance of models 3.1 + Fig 2: Are the improvement/ degradation of performance statistically significant? Or is it only visual?
Citation: https://doi.org/10.5194/egusphere-2025-3373-RC2
Data sets
Assessing evapotranspiration realism in rainfall-runoff models using evapotranspiration signatures Hansini Gardiya Weligamage, Keirnan Fowler, Margarita Saft, Tim Peterson, Dongryeol Ryu, Murray C. Peel https://doi.org/10.5281/zenodo.14228809
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 691 | 68 | 22 | 781 | 65 | 26 | 30 |
- HTML: 691
- PDF: 68
- XML: 22
- Total: 781
- Supplement: 65
- BibTeX: 26
- EndNote: 30
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript addresses an important topic in hydrological modeling by exploring joint calibration of rainfall–runoff models using both streamflow and flux tower evapotranspiration (AET) data. The study is relevant and has the potential to contribute to improving process representation and multi-variable calibration approaches. However, the paper would benefit from clearer methodological explanations and stronger conceptual consistency. The rationale for including the annual variability signature (CVannual) is unclear given that the adopted objective function explicitly ignores the annual timescale, and the description of calibration periods for discharge and AET is confusing and should more clearly distinguish between the fixed simulation window and site-specific data availability. Additionally, the use of the Nash–Sutcliffe Efficiency (NSE) to evaluate spatial patterns across catchments is not conceptually appropriate, as NSE is designed for temporal rather than spatial performance assessment. Some statements (e.g., regarding over- or underestimation of interannual variability) are internally inconsistent and should be revised for clarity. Overall, while the study is timely and methodologically relevant, it requires better alignment between the objectives, metrics, and interpretations to fully realize its potential contribution to the field. I hope the detailed comments provided in the attached pdf will help improve the clarity, structure, and scientific rigor of the manuscript.