# Climate Extrapolations in Hydrology: The Expanded Bluecat Methodology

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Bluecat and Its Expansion to Deal with Climate Projections

- To appropriately correct the D-model bias, which may differ for different ranges of $Q$;
- To infer the model uncertainty in terms of confidence bands, whose width may also differ for different ranges of $Q$.

## 3. Data

## 4. Results

## 5. Discussion and Conclusions

- It appropriately corrects the D-model bias.
- It infers the model uncertainty in terms of confidence bands.

- If the conditioning value $Q$ lies within the range of “observations”, the conditional distribution is inferred using a subset of pairs $\left(q,Q\right)$ with adjacent $Q$ values. Namely, those pairs satisfying $Q-\Delta {Q}_{1}\le \underset{\_}{Q}\le Q+\Delta {Q}_{2}$ are chosen, where $\Delta {Q}_{1}$ and $\Delta {Q}_{2}$ are properly specified to form a sample of pairs $\left(q,Q\right)$ that allows the empirical estimation of the conditional probabilities ${F}_{q|Q}\left(q|Q\right)\approx P\left\{\underset{\_}{q}\le q|Q-\Delta {Q}_{1}\le \underset{\_}{Q}\le Q+\Delta {Q}_{2}\right\}$.
- Otherwise, an extrapolation is made by the following method: A constant $c\ne 1$ is chosen so that $cQ$ lie within the range of “observations” and the above method be applicable by replacing or $Q$ with $cQ$. Then, the conditional distribution is estimated by enrolling the relationship ${F}_{q|Q}\left(q|Q\right)\approx {F}_{q|Q}\left(q+a\left(1-1/c\right)cQ|cQ\right)$, where $a$ is a parameter representing a regression slope between $Q$ and $q$. This parameter is locally determined for the high or low values of $Q$, depending on the direction of the extrapolation.

- The scales of the application of the Bluecat’s S-model are not necessarily different from that of the climate D-model. They could be precisely the same, as in the present application to Italy. In this respect, Bluecat makes the D-model consistent with reality irrespective of spatial scale. Apparently, the case where the scales are different (e.g., smaller area of the S-model with respect to that corresponding to the D-model) is also served by the proposed methodology, without any change with respect to what is described above.
- The Bluecat methodology considers both the time sequence and the amplitude of the D-model and actual series. It does not make a lumped fitting on the entire set of past data to find unique parametric relationships to be applied to the adaptation of the future values (as usually conducted in downscaling techniques). Nor does it regard the future values as correct ones that only need downscaling. Rather, it treats past and future values produced by the D-model in the same manner—as representing a model and not the truth.
- In addition to modifying a D-model series, correcting it for bias, Bluecat advances it to a stochastic representation, thus characterizing the uncertainty that is illustrated in this paper in terms of confidence bands.

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## References

- Milly, P.C.D.; Betancourt, J.; Falkenmark, M.; Hirsch, R.M.; Kundzewicz, Z.W.; Lettenmaier, D.P.; Stouffer, R.J. Stationarity is dead: Whither water management? Science
**2008**, 319, 573–574. [Google Scholar] [CrossRef] [PubMed] - Montanari, A.; Koutsoyiannis, D. Modeling and mitigating natural hazards: Stationarity is immortal! Water Resour. Res.
**2014**, 50, 9748–9756. [Google Scholar] [CrossRef] - Koutsoyiannis, D.; Montanari, A. Negligent killing of scientific concepts: The stationarity case. Hydrol. Sci. J.
**2015**, 60, 1174–1183. [Google Scholar] [CrossRef] - Padrón, R.S.; Gudmundsson, L.; Seneviratne, S.I. Observational constraints reduce likelihood of extreme changes in multidecadal land water availability. Geophys. Res. Lett.
**2019**, 46, 736–744. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Moon, H.; Gudmundsson, L.; Seneviratne, S.I. Drought persistence errors in global climate models. J. Geophys. Res. Atmos.
**2018**, 123, 3483–3496. [Google Scholar] [CrossRef] [PubMed] - Koutsoyiannis, D.; Efstratiadis, A.; Georgakakos, K. Uncertainty assessment of future hydroclimatic predictions: A comparison of probabilistic and scenario-based approaches. J. Hydrometeorol.
**2007**, 8, 261–281. [Google Scholar] [CrossRef] - Koutsoyiannis, D.; Efstratiadis, A.; Mamassis, N.; Christofides, A. On the credibility of climate predictions. Hydrol. Sci. J.
**2008**, 53, 671–684. [Google Scholar] [CrossRef] - Anagnostopoulos, G.G.; Koutsoyiannis, D.; Christofides, A.; Efstratiadis, A.; Mamassis, N. A comparison of local and aggregated climate model outputs with observed data. Hydrol. Sci. J.
**2010**, 55, 1094–1110. [Google Scholar] [CrossRef] - Kundzewicz, Z.W.; Stakhiv, E.Z. Are climate models “ready for prime time” in water resources management applications, or is more research needed? Editorial. Hydrol. Sci. J.
**2010**, 55, 1085–1089. [Google Scholar] [CrossRef] - Koutsoyiannis, D.; Christofides, A.; Efstratiadis, A.; Anagnostopoulos, G.G.; Mamassis, N. Scientific dialogue on climate: Is it giving black eyes or opening closed eyes? Reply to “A black eye for the Hydrological Sciences Journal” by D. Huard. Hydrol. Sci. J.
**2011**, 56, 1334–1339. [Google Scholar] [CrossRef] [Green Version] - Tyralis, H.; Koutsoyiannis, D. On the prediction of persistent processes using the output of deterministic models. Hydrol. Sci. J.
**2017**, 62, 2083–2102. [Google Scholar] [CrossRef] - Montanari, A.; Koutsoyiannis, D. A blueprint for process-based modeling of uncertain hydrological systems. Water Resour. Res.
**2012**, 48, W09555. [Google Scholar] [CrossRef] - Nearing, G. Comment on “A blueprint for process-based modeling of uncertain hydrological systems” by Alberto Montanari and Demetris Koutsoyiannis. Water Resour. Res.
**2014**, 50, 6260–6263. [Google Scholar] [CrossRef] - Montanari, A.; Koutsoyiannis, D. Reply to comment by G. Nearing on “A blueprint for process-based modeling of uncertain hydrological systems”. Water Resour. Res.
**2014**, 50, 6264–6268. [Google Scholar] [CrossRef] - Sikorska, A.; Montanari, A.; Koutsoyiannis, D. Estimating the uncertainty of hydrological predictions through data-driven resampling techniques. J. Hydrol. Eng. (ASCE)
**2015**, 20, A4014009. [Google Scholar] [CrossRef] - Papacharalampous, G.; Koutsoyiannis, D.; Montanari, A. Quantification of predictive uncertainty in hydrological modelling by harnessing the wisdom of the crowd: Methodology development and investigation using toy models. Adv. Water Resour.
**2020**, 136, 103471. [Google Scholar] [CrossRef] [Green Version] - Papacharalampous, G.; Tyralis, C.; Koutsoyiannis, D.; Montanari, A. Quantification of predictive uncertainty in hydrological modelling by harnessing the wisdom of the crowd: A large-sample experiment at monthly timescale. Adv. Water Resour.
**2020**, 136, 103470. [Google Scholar] [CrossRef] [Green Version] - Koutsoyiannis, D.; Montanari, A. Bluecat: A local uncertainty estimator for deterministic simulations and predictions. Water Resour. Res.
**2022**, 58, e2021WR031215. [Google Scholar] [CrossRef] - Koutsoyiannis, D. Stochastics of Hydroclimatic Extremes—A Cool Look at Risk; Kallipos Open Academic Editions: Athens, Greece, 2021; 333p, ISBN 978-618-85370-0-2. Available online: https://www.itia.ntua.gr/2000/ (accessed on 13 March 2022).
- Koutsoyiannis, D.; Montanari, A. A brisk local uncertainty estimator for hydrologic simulations and predictions (blue cat). In Proceedings of the EGU General Assembly Conference Abstracts, Online, 4–8 May 2020; p. 10125. Available online: https://www.itia.ntua.gr/2044 (accessed on 13 March 2022).
- Kalnay, E.; Kanamitsu, M.; Kistler, R.; Collins, W.; Deaven, D.; Gandin, L.; Iredell, M.; Saha, S.; White, G.; Woollen, J.; et al. The NCEP/NCAR 40-year reanalysis project. B. Am. Meteorol. Soc.
**1996**, 77, 437–472. [Google Scholar] [CrossRef] [Green Version] - Hausfather, Z.; Marvel, K.; Schmidt, G.A.; Nielsen-Gammon, J.W.; Zelinkadoi, M. Climate simulations: Recognize the ‘hot model’ problem. Nature
**2022**, 605, 26–29. [Google Scholar] [CrossRef] [PubMed]

**Figure 2.**Comparison of the actual and D-model temperature data: (

**upper**) original monthly data; (

**middle**) after the aggregation to annual scale; (

**lower**) after the subtraction of monthly means to reduce the effect of periodicity.

**Figure 3.**Comparison of actual and D-model precipitation data: (

**upper**) original monthly data; (

**lower**) after aggregation to annual scale.

**Figure 4.**True values vs. D-model predictions for temperature, along with the S-model predictions (median and 90% confidence limits). The vertical dotted lines define the area out of which extrapolation is necessary.

**Figure 5.**True values and predictions by D-model and S-model (median and 90% confidence limits) of temperature in Italy: (

**upper**) entire period; (

**middle**and

**lower**) focus on 20-year periods.

**Figure 6.**As Figure 5, but with S-model results (median and 90% confidence limits) replaced by running averages over 12 months.

**Figure 7.**Modified Figure 5 for two hypothetical cases, in which the D-model values are reconstructed so as to be (

**upper**) very close to reality, by replacing the D-model series with the weighted sum of true and D-model values with weights of 0.75 and 0.25, respectively, and (

**lower**) completely irrelevant to reality, by randomly rearranging the time order the true values.

**Figure 8.**True values vs. D-model predictions for precipitation, transformed for normalization, along with S-model predictions (median and 90% confidence limits). The vertical dotted lines define the area out of which extrapolation is necessary.

**Figure 9.**True values and predictions by D-model and S-model (median and 90% confidence limits) of precipitation in Italy: (

**upper**) entire period; (

**middle**and

**lower**) focus on 20-year periods. (Nb. A normalizing transformation by Equation (10) with $\lambda =2$ mm/d is used, while the plotted values are back transformed).

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Koutsoyiannis, D.; Montanari, A.
Climate Extrapolations in Hydrology: The Expanded Bluecat Methodology. *Hydrology* **2022**, *9*, 86.
https://doi.org/10.3390/hydrology9050086

**AMA Style**

Koutsoyiannis D, Montanari A.
Climate Extrapolations in Hydrology: The Expanded Bluecat Methodology. *Hydrology*. 2022; 9(5):86.
https://doi.org/10.3390/hydrology9050086

**Chicago/Turabian Style**

Koutsoyiannis, Demetris, and Alberto Montanari.
2022. "Climate Extrapolations in Hydrology: The Expanded Bluecat Methodology" *Hydrology* 9, no. 5: 86.
https://doi.org/10.3390/hydrology9050086