A statistical downscaling prediction model for winter temperature over Xinjiang based on the CFSv2 and sea ice forcing

• Published in: International Journal of Climatology
• Main Authors: Liu, Jing, Chen, Lijuan, Liu, Ying
• Other Authors: National Natural Science Foundation of China, National Key Research and Development Program of China
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2022
• Subjects: Barents Sea; Laptev Sea; laptev; Sea ice
• Online Access: http://dx.doi.org/10.1002/joc.7747; https://onlinelibrary.wiley.com/doi/pdf/10.1002/joc.7747; https://onlinelibrary.wiley.com/doi/full-xml/10.1002/joc.7747; https://rmets.onlinelibrary.wiley.com/doi/pdf/10.1002/joc.7747

Abstract Based on a year‐to‐year increment approach, a statistical downscaling model is developed for winter temperature prediction over Xinjiang of northwest China by using the predicted 200‐hPa westerly wind in winter over the Ural Mountains from the Climate Forecast System version 2 (CFSv2) as well as the observed sea ice over the Barents Sea–Laptev Sea in the preceding September. The statistical downscaling hindcasts on the 1983–2018 winter temperature over Xinjiang show that the statistical downscaling method has significantly improved the prediction capability compared with the original CFSv2. Specifically, the temporal correlation coefficients increase from negative to positive in most of Xinjiang, with the values passing the significance test at 90% confidence level at 81% of the stations. The regional averaged root‐mean‐square error reduces by more than 20%. In addition, the anomaly correlation coefficient increases from 0.08 to 0.26, passing the significance test at 95% confidence level. For two typical cases of the extreme cold winter in 2011 and the extreme warm winter in 2016, the spatial distribution characteristics of temperature anomaly are well reproduced, which are generally consistent with the observations. Overall, the statistical downscaling model based on the year‐to‐year increment strategy is a relatively effective method for predicting the winter temperature in Xinjiang.