Abstract
Climate data with high spatial and temporal resolution were of great significance for regional environmental management, such as for early response to possible predicted local climate changes and extreme weather. However, the current downscaling targets for CMIP6 climate simulations were mostly medium-resolution (MR) reanalysis data, which were still coarse for local analysis. A two-step downscaling method was proposed for 100 × resolution enhancements of general circulation model (GCM) daily temperature data in this study. First, the historical GCM outputs were 10 × downscaled to a set of dynamically predictable MR data using a deep convolutional neural network (CNN), which included both encode-decode structure and long-short skip connections. Then, using high-resolution (HR) topographic data and MR climate data as auxiliary data, the GCM data were super-resolved to a series of images with spatial resolution of 1 km. A one-step downscaling analysis combined only with HR topographic data was performed as comparison. Seven evaluation metrics were selected to evaluate the prediction accuracy, and the results showed that the overall performance of two-step downscaling method was better than one-step downscaling method. Higher Nash–Sutcliffe efficiency (NSE) and lower mean absolute relative error (MARE) indicated that the two-step method performed better prediction of peak and low values. It was further confirmed by accuracy evaluation on the 10% max and 10% min values of the testing dataset. The introduction of dynamically predictable MR data could provide effective detailed information during the downscaling process and improve the prediction accuracies. Finally, the projected data of four scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5) during 2015–2050 were downscaled to the study area. The complex temporal and spatial variations indicated that there were great differences in temperature changes in a basin, and differentiated management measures should be proposed in advance.
Similar content being viewed by others
Data availability
The data presented in this study are available on request from the corresponding author. Declarations.
References
Adewoyin RA, Dueben P, Watson P, He Y, Dutta R (2021) TRU-NET: a deep learning approach to high resolution prediction of rainfall. Mach Learn 110:2035–2062
Asadollah SBHS, Sharafati A, Shahid S (2022) Application of ensemble machine learning model in downscaling and projecting climate variables over different climate regions in Iran. Environ Sci Pollut Res 29:17260–17279
Baño-Medina J, Manzanas R, Gutiérrez JM (2020) Configuration and intercomparison of deep learning neural models for statistical downscaling. Geosci Model Dev 13:2109–2124
Cai H, Shi H, Liu S, Babovic V (2021) Impacts of regional characteristics on improving the accuracy of groundwater level prediction using machine learning: the case of central eastern continental United States. J Hydrol Reg Stud 37:100930
Chadalawada J, Babovic V (2019) Review and comparison of performance indices for automatic model induction. J Hydroinf 21:13–31
Chou C, Park J, Chou E (2021) Generating high-resolution climate change projections using super-resolution convolutional LSTM neural network. In: 2021 13th International Conference on Advanced Computational Intelligence (ICACI). IEEE, pp 293–298
Deser C, Lehner F, Rodgers KB, Ault T, Delworth TL, DiNezio PN, Fiore A, Frankignoul C, Fyfe JC, Horton DE (2020) Insights from Earth system model initial-condition large ensembles and future prospects. Nat Clim Chang 10:277–286
Fan X, Jiang L, Gou J (2021) Statistical downscaling and projection of future temperatures across the Loess Plateau, China. Weath Clim Extr 32:100328
Ghosh S (2010) SVM-PGSL coupled approach for statistical downscaling to predict rainfall from GCM output. J Geophys Res Atmos 115(D22). https://doi.org/10.1029/2009JD013548
Hassan WH (2021) Climate change projections of maximum temperatures for southwest Iraq using statistical downscaling. Climate Res 83:187–200
Hassan WH, Hashim FS (2020) The effect of climate change on the maximum temperature in Southwest Iraq using HadCM3 and CanESM2 modelling. SN Applied Sciences 2:1–11
Hassan WH, Nile BK (2021) Climate change and predicting future temperature in Iraq using CanESM2 and HadCM3 modeling. Model Earth Syst Environ 7:737–748
Hassan WH, Hussein H, Nile BK (2022) The effect of climate change on groundwater recharge in unconfined aquifers in the western desert of Iraq. Groundw Sustain Dev 16:100700
He K, Zhang X, Ren S, Sun J (2016) Identity mappings in deep residual networks. European conference on computer vision. Springer, Berlin, pp 630–645
He J, Yang K, Tang W, Lu H, Qin J, Chen Y, Li X (2020) The first high-resolution meteorological forcing dataset for land process studies over China. Sci Data 7:1–11
Jiang Y, Yang K, Shao C, Zhou X, Zhao L, Chen Y, Wu H (2021) A downscaling approach for constructing high-resolution precipitation dataset over the Tibetan Plateau from ERA5 reanalysis. Atmos Res 256:105574
Jiang S, Zheng Y, Wang C, Babovic V (2022) Uncovering flooding mechanisms across the contiguous United States through interpretive deep learning on representative catchments. Water Resour Res 58:e2021WR030185
Khan A, Sohail A, Zahoora U, Qureshi AS (2020) A survey of the recent architectures of deep convolutional neural networks. Artif Intell Rev 53:5455–5516
Kling H, Fuchs M, Paulin M (2012) Runoff conditions in the upper Danube basin under an ensemble of climate change scenarios. J Hydrol 424:264–277
Knoben WJ, Freer JE, Woods RA (2019) Inherent benchmark or not? Comparing Nash-Sutcliffe and Kling-Gupta efficiency scores. Hydrol Earth Syst Sci 23:4323–4331
Kumar B, Chattopadhyay R, Singh M, Chaudhari N, Kodari K, Barve A (2021) Deep learning–based downscaling of summer monsoon rainfall data over Indian region. Theoret Appl Climatol 143:1145–1156
Kun Y, Jie H (2019) China meteorological forcing dataset (1979-2018). In: National Tibetan Plateau Data, C. (Ed.). National Tibetan Plateau Data Center. https://doi.org/10.11888/AtmosphericPhysics.tpe.249369.file
Li X, Babovic V (2019) Multi-site multivariate downscaling of global climate model outputs: an integrated framework combining quantile mapping, stochastic weather generator and Empirical Copula approaches. Clim Dyn 52:5775–5799
Li P, Shi C, Li Z, Muller J-P, Drummond J, Li X, Li T, Li Y, Liu J (2013) Evaluation of ASTER GDEM using GPS benchmarks and SRTM in China. Int J Remote Sens 34:1744–1771
Li X, Sha J, Wang Z-L (2019) Comparison of daily streamflow forecasts using extreme learning machines and the random forest method. Hydrol Sci J 64:1857–1866
Liu Y, Ganguly AR, Dy J (2020) Climate downscaling using YNet: A deep convolutional network with skip connections and fusion. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. pp. 3145–3153. https://doi.org/10.1145/3394486.3403366
Mao X, Shen C, Yang Y-B (2016) Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connections. Adv Neural Inf Process Syst 29:2802–2810
Misra S, Sarkar S, Mitra P (2018) Statistical downscaling of precipitation using long short-term memory recurrent neural networks. Theoret Appl Climatol 134:1179–1196
Pan B, Hsu K, AghaKouchak A, Sorooshian S (2019) Improving precipitation estimation using convolutional neural network. Water Resour Res 55:2301–2321
Pham HX, Shamseldin AY, Melville BW (2021) Projection of future extreme precipitation: a robust assessment of downscaled daily precipitation. Nat Hazards 107:311–329
Pichuka S, Maity SS, Maity R (2022) Benefit of time-varying downscaling model for the assessment of urban temperature rise. Model Earth Syst Environ 8:2431–2447
Sachindra D, Ahmed K, Rashid MM, Shahid S, Perera B (2018) Statistical downscaling of precipitation using machine learning techniques. Atmos Res 212:240–258
Stengel K, Glaws A, Hettinger D, King RN (2020) Adversarial super-resolution of climatological wind and solar data. Proc Natl Acad Sci 117:16805–16815
Sun L, Lan Y (2021) Statistical downscaling of daily temperature and precipitation over China using deep learning neural models: localization and comparison with other methods. Int J Climatol 41:1128–1147
Tulbure A-A, Tulbure A-A, Dulf E-H (2022) A review on modern defect detection models using DCNNs–deep convolutional neural networks. J Adv Res 35:33–48
Vandal T, Kodra E, Ganguly AR (2019) Intercomparison of machine learning methods for statistical downscaling: the case of daily and extreme precipitation. Theoret Appl Climatol 137:557–570
Vaughan A, Tebbutt W, Hosking JS, Turner RE (2022) Convolutional conditional neural processes for local climate downscaling. Geosci Model Dev 15(1): 251–268. https://doi.org/10.5194/gmd-15-251-2022
Wang F, Tian D, Lowe L, Kalin L, Lehrter J (2021) Deep learning for daily precipitation and temperature downscaling. Water Resour Res 57:e2020WR029308
Willmott CJ, Robeson SM, Matsuura K (2012) A refined index of model performance. Int J Climatol 32:2088–2094
Wootten AM, Dixon KW, Adams-Smith DJ, McPherson RA (2021) Statistically downscaled precipitation sensitivity to gridded observation data and downscaling technique. Int J Climatol 41:980–1001
Xiaodong Z, Wenbin T, Lirong D, Jin M, Xu Z (2021) Daily 1-km all-weather land surface temperature dataset for the Chinese landmass and its surrounding areas (TRIMS LST; 2000-2021). In: National Tibetan Plateau Data, C. (Ed.). National Tibetan Plateau Data Center. https://doi.org/10.11888/Meteoro.tpdc.271252
Xu K, Zhang M, Jegelka S, Kawaguchi K (2021) Optimization of graph neural networks: implicit acceleration by skip connections and more depth. arXiv preprint arXiv:2105.04550.
Yang K, He J, Tang W, Qin J, Cheng CC (2010) On downward shortwave and longwave radiations over high altitude regions: observation and modeling in the Tibetan Plateau. Agric for Meteorol 150:38–46
Yu T, Kuang Q, Zheng J, Hu J (2022) Deep Precipitation Downscaling. IEEE Geosci Remote Sens Lett 19: 1–5. https://doi.org/10.1109/LGRS.2021.3049673
Zeleznik R, Foldyna B, Eslami P, Weiss J, Alexander I, Taron J, Parmar C, Alvi RM, Banerji D, Uno M (2021) Deep convolutional neural networks to predict cardiovascular risk from computed tomography. Nat Commun 12:1–9
Zhang X, Zhou J, Göttsche F-M, Zhan W, Liu S, Cao R (2019) A method based on temporal component decomposition for estimating 1-km all-weather land surface temperature by merging satellite thermal infrared and passive microwave observations. IEEE Trans Geosci Remote Sens 57:4670–4691
Zhang X, Zhou J, Liang S, Wang D (2021a) A practical reanalysis data and thermal infrared remote sensing data merging (RTM) method for reconstruction of a 1-km all-weather land surface temperature. Remote Sens Environ 260:112437
Zhang Z, Park CY, Theesfeld CL, Troyanskaya OG (2021b) An automated framework for efficiently designing deep convolutional neural networks in genomics. Nat Mach Intell 3:392–400
Zhou J, Zhang X, Zhan W, Göttsche F-M, Liu S, Olesen F-S, Hu W, Dai F (2017) A thermal sampling depth correction method for land surface temperature estimation from satellite passive microwave observation over barren land. IEEE Trans Geosci Remote Sens 55:4743–4756
Funding
This work is supported by Open Research Fund Program of State key Laboratory of Hydroscience and Engineering (sklhse-2021-B-07).
Author information
Authors and Affiliations
Contributions
All the authors contributed to the study conception and design. Data analysis, methodology, software, and code writing were performed by Xue Li and Man Zhang. Data curation and validation were performed by Sha Jian and Zhou Yingyin. Conceptualization and supervision were performed by Zhang Man and Wang Zhong-Liang. The first draft of the manuscript was written by Xue Li, and all the authors commented on the previous versions of the manuscript. All the authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Ethical approval
Not applicable.
Consent to participate
All the authors agreed to participate in this study.
Consent for publication
All the authors read and approved the final manuscript.
Competing interests
The authors declare no competing interests.
Additional information
Responsible Editor: Philippe Garrigues
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Li, X., Zhou, Y., Zhang, M. et al. A two-step downscaling method for high-scale super-resolution of daily temperature — a case study of Wei River Basin, China. Environ Sci Pollut Res 30, 32474–32488 (2023). https://doi.org/10.1007/s11356-022-24422-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11356-022-24422-6