Deep-Learning-Based Harmonization and Super-Resolution of Near-Surface Air Temperature from CMIP6 Models (1850–2100)

Future global temperature change would have significant effects on society and ecosystems. Earth system models (ESM) are the primary tools to explore the future climate change. However, ESMs still exist great uncertainty and often run at a coarse spatial resolution (The majority of ESMs at about 2 degree). Accurate temperature data at high spatial resolution are needed to improve our understanding of the temperature variation and for many applications. We innovatively apply the deep-learning(DL) method from the Super resolution (SR) in the computer vision to merge 31 ESMs data and the proposed method can perform data merge, bias-correction and spatial-downscaling simultaneously. The SR algorithms are designed to enhance image quality and outperform much better than the traditional methods. The CRU TS (Climate Research Unit gridded Time Series) is considered as reference data in the model training process. In order to find a suitable DL method for our work, we choose five SR methodologies made by different structures. Those models are compared based on multiple evaluation metrics (Mean square error(MSE), mean absolute error(MAE) and Pearson correlation coefficient(R)) and the optimal model is selected and used to merge the monthly historical data during 1850–1900 and monthly future scenarios data (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) during 2015–2100 at the high spatial resolution of 0.5 degree. Results showed that the merged data have considerably improved performance than any of the individual ESM data and the ensemble mean (EM) of all ESM data in terms of both spatial and temporal aspects. The MAE displays a great improvement and the spatial distribution of the MAE become larger and larger along the latitudes in north hemisphere, presenting like a ‘tertiary class echelon’ condition. The merged product also presents excellent performance when the observation data is smooth with few fluctuations in time series. Additionally, this work proves that the DL model can be transferred to deal with the data merge, bias-correction and spatial-downscaling successfully when enough training data are available. Data can be accessed at https://doi.org/10.5281/zenodo.5746632 (Wei et al., 2021).

Improving the Spatial Resolution of GRACE-Derived Terrestrial Water Storage Changes in Small Areas Using the Machine Learning Spatial Downscaling Method

Gravity Recovery and Climate Experiment (GRACE) satellites can effectively monitor terrestrial water storage (TWS) changes in large-scale areas. However, due to the coarse resolution of GRACE products, there is still a large number of deficiencies that need to be considered when investigating TWS changes in small-scale areas. Hence, it is necessary to downscale the GRACE products with a coarse resolution. First, in order to solve this problem, the present study employs modeling windows of different sizes (Window Size, WS) combined with multiple machine learning algorithms to develop a new machine learning spatial downscaling method (MLSDM) in the spatial dimension. Second, The MLSDM is used to improve the spatial resolution of GRACE observations from 0.5° to 0.25°, which is applied to Guantao County. The present study has verified the downscaling accuracy of the model developed through the combination of WS3, WS5, WS7, and WS9 and jointed with Random Forest (RF), Extra Tree Regressor (ETR), Adaptive Boosting Regressor (ABR), and Gradient Boosting Regressor (GBR) algorithms. The analysis shows that the accuracy of each combined model is improved after adding the residuals to the high-resolution downscaled results. In each modeling window, the accuracy of RF is better than that of ETR, ABR, and GBR. Additionally, compared to the changes in the TWS time series that are derived by the model before and after downscaling, the results indicate that the downscaling accuracy of WS5 is slightly more superior compared to that of WS3, WS7, and WS9. Third, the spatial resolution of the GRACE data was increased from 0.5° to 0.05° by integrating the WS5 and RF algorithm. The results are as follows: (1) The TWS (GWS) changes before and after downscaling are consistent, decreasing at −20.86 mm/yr and −21.79 mm/yr (−14.53 mm/yr and −15.46 mm/yr), respectively, and the Nash–Sutcliffe efficiency coefficient (NSE) and correlation coefficient (CC) values of both are above 0.99 (0.98). (2) The CC between the 80% deep groundwater well data and the downscaled GWS changes are above 0.70. Overall, the MLSDM can not only effectively improve the spatial resolution of GRACE products but also can preserve the spatial distribution of the original signal, which can provide a reference scheme for research focusing on the downscaling of GRACE products.

Research Progress on Precipitation Accuracy Verification and Statistical Post-Processing in Ensemble Numerical Forecast System

The temporal and spatial accuracy of precipitation of ensemble numerical forecast systems is an important factor that affects the level of meteorological and hydrological coupled forecasting. This article focuses on the current research of verification of precipitation accuracy and statistical post-processing. The verification of forecast precipitation accuracy mainly focuses on the probabilistic characteristics such deterministic accuracy, the resolution, the forecasting skills and the degree of dispersion. Some mainstream statistical post-processing methods have strong performance of spatial downscaling and error correction, but they commonly have the defect of destroying the temporal and spatial dependent structure of precipitation. A comprehensive statistical post-processing method integrated the three functions is the development direction in the future. At the same time, statistical post-processing methods to improve the certainty and probabilistic accuracy of forecast precipitation need to be systematically identified. Its impact on the spatio-temporal dependence structure also needs to be improved.

Spatial Downscaling of MODIS Snow Cover Observations Using Sentinel-2 Snow Products

Understanding those processes in which snow dynamics has a significant influence requires long-term and high spatio-temporal resolution observations. While new optical space-borne sensors overcome many previous snow cover monitoring limitations, their short temporal length limits their application in climatological studies. This work describes and evaluates a probabilistic spatial downscaling of MODIS snow cover observations in mountain areas. The approach takes advantage of the already available high spatial resolution Sentinel-2 snow observations to obtain a snow probability occurrence, which is then used to determine the snow-covered areas inside partially snow-covered MODIS pixels. The methodology is supported by one main hypothesis: the snow distribution is strongly controlled by the topographic characteristics and this control has a high interannual persistence. Two approaches are proposed to increase the 500 m resolution MODIS snow cover observations to the 20 m grid resolution of Sentinel-2. The first of these computes the probability inside partially snow-covered MODIS pixels by determining the snow occurrence frequency for the 20 m Sentinel-2 pixels when clear-sky conditions occurred for both platforms. The second approach determines the snow probability occurrence for each Sentinel-2 pixel by computing the number of days in which snow was observed on each grid cell and then dividing it by the total number of clear-sky days per grid cell. The methodology was evaluated in three mountain areas in the Iberian Peninsula from 2015 to 2021. The 20 m resolution snow cover maps derived from the two probabilistic methods provide better results than those obtained with MODIS images downscaled to 20 m with a nearest-neighbor method in the three test sites, but the first provides superior performance. The evaluation showed that mean kappa values were at least 10% better for the two probabilistic methods, improving the scores in one of these sites by 25%. In addition, as the Sentinel-2 dataset becomes longer in time, the probabilistic approaches will become more robust, especially in areas where frequent cloud cover resulted in lower accuracy estimates.