the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Two sets of bias-corrected regional UK Climate Projections 2018 (UKCP18) of temperature, precipitation and potential evapotranspiration for Great Britain
Abstract. The United Kingdom Climate Projections 2018 (UKCP18) regional climate model (RCM) 12 km regional perturbed physics ensemble (UKCP18-RCM-PPE) is one of the three strands of the latest set of UK national climate projections produced by the UK Met Office. It has been widely adopted in climate impact assessment. In this study, we report biases in the raw UKCP18-RCM simulations that are significant and are likely to deteriorate impact assessments if they are not adjusted. Two methods were used to bias-correct UKCP18-RCM: non-parametric quantile mapping using empirical quantiles and a variant developed for the third phase of the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) designed to preserve the climate change signal. Specifically, daily temperature and precipitation simulations for 1981 to 2080 were adjusted for the 12 ensemble members. Potential evapotranspiration was also estimated over the same period using the Penman-Monteith formulation and then bias-corrected using the latter method. Both methods successfully corrected biases in a range of daily temperature, precipitation and potential evapotranspiration metrics, and reduced biases in multi-day precipitation metrics to a lesser degree. An exploratory analysis of the projected future changes confirms the expectation of wetter, warmer winters and hotter, drier summers, and shows uneven changes in different parts of the distributions of both temperature and precipitation. Both bias-correction methods preserved the climate change signal almost equally well, as well as the spread among the projected changes. The change factor method was used as a benchmark for precipitation, and we show that it fails to capture changes in a range of variables, making it inadequate for most impact assessments. By comparing the differences between the two bias-correction methods and within the 12 ensemble members, we show that the uncertainty in future precipitation and temperature changes stemming from the climate model parameterisation far outweighs the uncertainty introduced by selecting one of these two bias-correction methods. We conclude by providing guidance on the use of the bias-corrected data sets. The data sets bias adjusted with ISIMIP3BA are publicly available in the following repositories: https://doi.org/10.5281/zenodo.6337381 for precipitation and temperature (Reyniers et al., 2022a) and https://doi.org/10.5281/zenodo.6320707 for potential evapotranspiration (Reyniers et al., 2022b) . The datasets bias-corrected using the quantile mapping method are available at https://doi.org/10.5281/zenodo.8223024 (Zha et al., 2023) .
- Preprint
(9683 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2024-132', Anonymous Referee #1, 14 Oct 2024
Reyniers et al. 2024 evaluated and bias-corrected the UKCP18-RCM-PPE dataset and discussed the performance of both the original and bias-corrected data. I acknowledge that bias correction (BC) is important for impact studies. Effective BC should ideally reduce model biases while preserving key model benefits, such as climate change trend projections and spatial details that might not be captured by observations. However, this study does not appear to introduce a new BC method or provide in-depth insights into the limitations of exiting BC approaches. Similar BC techniques have already been widely applied in many impact studies, which somewhat limits the novelty of this study and the dataset produced.
I have several suggestions for improvement:
- Line 35-40: This is a good and important discussion on key limitations of bias correction. However, the methodology used in this study did not addresses these challengesyou’re your BC approach did help to “address the origin of model errors”, that would be a significant contribution to the community. As it stands, the manuscript does not seem to offer solutions for tackling the root causes of model errors.
- Line 66: “simple BC methods used in CHESS-SCAPE”: what methods they use?
- The data you corrected is from a perturbed physics ensemble, which is designed to both explore the influence of parameter variations on simulations and to reduce uncertainty. It would be highly valuable to examine how bias correction affects the ensemble results. For example, does using a single observation dataset to correct multiple ensemble runs impact the ensemble’s representation of uncertainty? Additionally, it would be useful to see a discussion on the performance of individual ensemble members before and after bias correction. Such an analysis could provide insights into how bias correction interacts with model parameterizations and the ensemble spread, and provide physical interpretation of your results.
- The overestimation of dry-day frequency is a common issue in climate models, often linked to the drizzle effect. I recommend addressing the drizzle effect first, as this might improve the accuracy of subsequent bias corrections and separate this systematic bias from others.
Citation: https://doi.org/10.5194/essd-2024-132-RC1 -
RC3: 'Reply on RC1', Anonymous Referee #1, 21 Oct 2024
Correction: "you’re your BC approach did help to ...”'- "if your BC approach did help to ..."
Citation: https://doi.org/10.5194/essd-2024-132-RC3 - AC1: 'Reply on RC1', qianyu zha, 06 Nov 2024
-
RC2: 'Comment on essd-2024-132', Anonymous Referee #2, 21 Oct 2024
Two simple bias correction methods have been addressed to investigate their impact on the climate model projections. Overall, these are interesting results. However, I think some major concerns and questions need to be addressed. Here, I list some of the comments/questions that were unclear.
Main comments
1. This paper evaluates two relatively simple statistical methods for correcting climate model output directly. The results indicate that bias correction techniques can mitigate biases related to the indices utilized in this study. Nonetheless, my principal concern is the paper's originality. Numerous studies have already been published that compare various statistical and machine learning methods across different domains and temporal contexts, ranging from historical data to future projections. Consequently, I do not believe this paper contributes significantly new information relative to existing literature, particularly given that available published data could offer additional insights. While the journal is dedicated to data publication, it is essential for this paper to either introduce a novel methodology or provide a comprehensive analysis of each method, highlighting the advantages and disadvantages of their application in specific circumstances. Merely comparing simple methods that produce comparable results for future projections may not provide substantial value.
2. The manuscript lacks a comprehensive explanation for the preference of two simpler methods over more sophisticated approaches. The selection of the degree of bias correction should depend on the specific application, as advanced techniques have been shown to produce greater improvements, particularly in the context of extreme events. While quantile mapping and simple climatological mean correction have demonstrated their advantages in preventing excessive correction based on observational data, they may still permit the persistence of biases related to low-frequency variability, which can complicate the direct correction of surface variables in climate model outputs.
Numerous studies have employed sophisticated bias correction across various time scales to adjust the outputs of GCM and RCM or the boundary conditions for RCM inputs, with the objective of enhancing the accuracy of simulations for extreme and compound events. It would be advantageous to incorporate additional references that pertain to bias correction prior to addressing the limitations of existing studies. Furthermore, it is essential to provide detailed information and explanations regarding using simpler methods in the introduction or conclusion to ensure a thorough understanding of the techniques ranging from simple to sophisticated methods.
- Correcting outputs
Wood AW, Leung LR, Sridhar V, Lettenmaier D (2004) Hydrologic implications of dynamical and statistical approaches to downscaling climate model outputs. Clim Change 62:189–216
Cannon, A. J. (2018). Multivariate quantile mapping bias correction: an N-dimensional probability density function transform for climate model simulations of multiple variables. Climate dynamics, 50(1), 31-49.
- Correcting RCM input full variable fields
Bruyere CL, Done JM, Holland GJ, Fredrick S (2014) Bias corrections of global models for regional climate simulations of high-impact weather. Clim Dyn 43:1847–1856
Kim, Y, Evans, JP, Sharma, A (2023). Can Sub‐Daily Multivariate Bias Correction of Regional Climate Model Boundary Conditions Improve Simulation of the Diurnal Precipitation Cycle?. Geophysical Research Letters, 50(22), p.e2023GL104442.
- Software for correcting climate model variables
Cannon, A.J. (2016). Multivariate bias correction of climate model output: Matching marginal distributions and intervariable dependence structure. Journal of Climate, 29(19), pp.7045-7064.
Kim, Y., Evans, J. P., & Sharma, A. (2023). A software for correcting systematic biases in RCM input boundary conditions. Environmental Modelling & Software, 168, 105799.
3. The authors have undertaken corrections for three variables: precipitation, temperature, and potential evapotranspiration (PET). It would be beneficial to clarify the methodology employed for correcting PET. Did the authors correct PET derived from the raw variables directly, or did they utilize the adjusted variables in the PET calculations? As indicated by the authors, PET is influenced by several variables generated from the climate model, including specific humidity, pressure, and temperature, which can be modified by adjusting surface variables interconnected with these factors. Correcting these variables statistically, without accounting for the physical relationships among them, may lead to inconsistencies and produce unrealistic results.
Specific comments
L66. “the simple methods …” It would be beneficial to provide more details about what these methods entail.L68. “the quantile mapping (QM) method outperforms … the standards deviation and percentiles.” Quantile mapping (QM) can outperform simple mean correction for variance since the latter does not address the standard deviation or percentiles. I recommend incorporating more details about bias correction techniques, including methods like simple mean and standard deviation correction. This will help justify the use of empirical QM when publishing datasets for broader applications.
L82. “the trend-preserving BC method.” It is essential to justify the trend-preserving method for future projections, as climate models also contain biases in trends.
Figures 9 and 10. I recommend modifying the figures to use a bias map instead of presenting each one individually, as the differences are minimal.
Citation: https://doi.org/10.5194/essd-2024-132-RC2 - AC2: 'Reply on RC2', qianyu zha, 19 Nov 2024
Data sets
UKCP18 RCM precipitation and temperature bias corrected using non-parametric quantile mapping method Qianyu Zha, Nele Reyniers, Nans Addor, Timothy J. Osborn, Yi He, and Nicole Forstenhäusler https://doi.org/10.5281/zenodo.8223024
UKCP18 RCM precipitation and temperature bias corrected using ISIMIP3BA change-preserving quantile mapping Nele Reyniers, Nans Addor, Qianyu Zha, Timothy J. Osborn, Nicole Forstenhäusler, and Yi He https://doi.org/10.5281/zenodo.6337381
Projected changes in droughts and extreme droughts in Great Britain are strongly influenced by the choice of drought index: UKCP18-based bias adjusted potential evapotranspiration Nele Reyniers, Timothy J. Osborn, Nans Addor, and Geoff Darch https://doi.org/10.5281/zenodo.6320707
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
545 | 149 | 29 | 723 | 24 | 27 |
- HTML: 545
- PDF: 149
- XML: 29
- Total: 723
- BibTeX: 24
- EndNote: 27
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1