Volume 14, Issue 27 (8-2023)
J Watershed Manage Res 2023, 14(27): 75-85 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Ershadfath F, Shahnazari A, Raeini Sarjaz M, Eivind Olese J. (2023). Combining Bias Correction Methods for Simulated Temperature and Precipitation by CMIP6 Models in Hamedan-Bahar Plain. J Watershed Manage Res. 14(27), 75-85. doi:10.61186/jwmr.14.27.75
URL: http://jwmr.sanru.ac.ir/article-1-1201-en.html
URL: http://jwmr.sanru.ac.ir/article-1-1201-en.html
1- Sari University of Agriculture and Natural Sciences
2- Aarhus University
2- Aarhus University
Abstract: (1626 Views)
Extended Abstract
Introduction and Objective: Applying the raw data of regional climate models in assessing the impact of climate change is not advisable due to possible biases. Therefore, correcting the bias of these data is necessary before using them for climate scenarios of the future. The aim of this study is to evaluate the performance and introduce the best combination of bias correction methods for precipitation and minimum and maximum temperature simulated by three CMIP6 climate models.
Material and Methods: Five bias correction methods including linear scaling, variance scaling, local intensity scaling, power transformation and distribution mapping were investigated using root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), correlation coefficient (r) and Student's t test for the historical period (1990-2014). Afterward, the combination of the best bias correction methods was used to project precipitation and temperature under SSP2-4.5 and SSP5-8.5 scenarios in the future period (2051-2075).
Results: Based on the results, three methods of variance scaling, local intensity scaling and power transformation for correcting the bias of the investigated data had weaker performances compared to the other methods. Two methods of linear scaling and distribution mapping had the lowest RMSE and highest r and NSE. Projecting the future climate using the combination of these two selected methods showed that the average annual precipitation in the Hamedan-Bahar region will decrease by 28 and 37 percent under scenarios of SSP2-4.5 and SSP5-8.5, respectively. Furthermore, the annual average of the maximum and minimum temperature will increase by 0.7, 0.9 under scenarios of SSP2-4.5 and 1.4 and 1.5 °C, under scenarios of SSP5-8.5, respectively. In addition, the highest seasonal decrease in precipitation (19.8 mm) compared to the baseline period will occur in the spring under the SSP5-8.5 scenario. Moreover, the highest seasonal increase of maximum and minimum temperature compared to the baseline period was projected in winter (1.6°C) and spring (1.7°C), respectively, under the SSP5-8.5 scenario.
Conclusion: Two methods of linear scaling and distribution mapping are suitable for reducing the bias of the CMIP6 models in the Hamedan-Bahar plain. Also, considering the projected increase in temperature and decrease in precipitation in this region, this study can provide useful information for policy-makers of water resources and agriculture to decide about the rainwater harvesting, recharging of aquifers, crop selection, cultivating period, crop rotation and management methods to reduce the impact of future climate change.
Introduction and Objective: Applying the raw data of regional climate models in assessing the impact of climate change is not advisable due to possible biases. Therefore, correcting the bias of these data is necessary before using them for climate scenarios of the future. The aim of this study is to evaluate the performance and introduce the best combination of bias correction methods for precipitation and minimum and maximum temperature simulated by three CMIP6 climate models.
Material and Methods: Five bias correction methods including linear scaling, variance scaling, local intensity scaling, power transformation and distribution mapping were investigated using root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), correlation coefficient (r) and Student's t test for the historical period (1990-2014). Afterward, the combination of the best bias correction methods was used to project precipitation and temperature under SSP2-4.5 and SSP5-8.5 scenarios in the future period (2051-2075).
Results: Based on the results, three methods of variance scaling, local intensity scaling and power transformation for correcting the bias of the investigated data had weaker performances compared to the other methods. Two methods of linear scaling and distribution mapping had the lowest RMSE and highest r and NSE. Projecting the future climate using the combination of these two selected methods showed that the average annual precipitation in the Hamedan-Bahar region will decrease by 28 and 37 percent under scenarios of SSP2-4.5 and SSP5-8.5, respectively. Furthermore, the annual average of the maximum and minimum temperature will increase by 0.7, 0.9 under scenarios of SSP2-4.5 and 1.4 and 1.5 °C, under scenarios of SSP5-8.5, respectively. In addition, the highest seasonal decrease in precipitation (19.8 mm) compared to the baseline period will occur in the spring under the SSP5-8.5 scenario. Moreover, the highest seasonal increase of maximum and minimum temperature compared to the baseline period was projected in winter (1.6°C) and spring (1.7°C), respectively, under the SSP5-8.5 scenario.
Conclusion: Two methods of linear scaling and distribution mapping are suitable for reducing the bias of the CMIP6 models in the Hamedan-Bahar plain. Also, considering the projected increase in temperature and decrease in precipitation in this region, this study can provide useful information for policy-makers of water resources and agriculture to decide about the rainwater harvesting, recharging of aquifers, crop selection, cultivating period, crop rotation and management methods to reduce the impact of future climate change.
Keywords: Bias correction, CMIP6, Climate change, Distribution mapping, Linear scaling, SSPs scenarios
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
