A machine learning approach to estimation of downward solar radiation from satellite-derived data products: An application over a semi-arid ecosystem in the U.S

Integrating Reanalysis and Satellite Cloud Information to Estimate Surface Downward Long-Wave Radiation

The estimation of downward long-wave radiation (DLR) at the surface is very important for the understanding of the Earth’s radiative budget with implications in surface–atmosphere exchanges, climate variability, and global warming. Theoretical radiative transfer and observationally based studies identify the crucial role of clouds in modulating the temporal and spatial variability of DLR. In this study, a new machine learning algorithm that uses multivariate adaptive regression splines (MARS) and the combination of near-surface meteorological data with satellite cloud information is proposed. The new algorithm is compared with the current operational formulation used by the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) Satellite Application Facility on Land Surface Analysis (LSA-SAF). Both algorithms use near-surface temperature and dewpoint temperature along with total column water vapor from the latest European Centre for Medium-range Weather Forecasts (ECMWF) reanalysis ERA5 and satellite cloud information from the Meteosat Second Generation. The algorithms are trained and validated using both ECMWF-ERA5 and DLR acquired from 23 ground stations as part of the Baseline Surface Radiation Network (BSRN) and the Atmospheric Radiation Measurement (ARM) user facility. Results show that the MARS algorithm generally improves DLR estimation in comparison with other model estimates, particularly when trained with observations. When considering all the validation data, root mean square errors (RMSEs) of 18.76, 23.55, and 22.08 W·m−2 are obtained for MARS, operational LSA-SAF, and ERA5, respectively. The added value of using the satellite cloud information is accessed by comparing with estimates driven by ERA5 total cloud cover, showing an increase of 17% of the RMSE. The consistency of MARS estimate is also tested against an independent dataset of 52 ground stations (from FLUXNET2015), further supporting the good performance of the proposed model.

Using supervised learning to develop BaRAD, a 40-year monthly bias-adjusted global gridded radiation dataset

Diffuse solar radiation is an important, but understudied, component of the Earth's surface radiation budget, with most global climate models not archiving this variable and a dearth of ground-based observations. Here, we describe the development of a global 40-year (1980-2019) monthly database of total shortwave radiation, including its diffuse and direct beam components, called BaRAD (Bias-adjusted RADiation dataset). The dataset is based on a random forest algorithm trained using Global Energy Balance Archive (GEBA) observations and applied to the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) dataset at the native MERRA-2 resolution (0.5° by 0.625°). The dataset preserves seasonal, latitudinal, and long-term trends in the MERRA-2 data, but with reduced biases than MERRA-2. The mean bias error is close to 0 (root mean square error = 10.1 W m-2) for diffuse radiation and -0.2 W m-2 (root mean square error = 19.2 W m-2) for the total incoming shortwave radiation at the surface. Studies on atmosphere-biosphere interactions, especially those on the diffuse radiation fertilization effect, can benefit from this dataset.

An Innovative Intelligent System with Integrated CNN and SVM: Considering Various Crops through Hyperspectral Image Data

Generation of a thematic map is important for scientists and agriculture engineers in analyzing different crops in a given field. Remote sensing data are well-accepted for image classification on a vast area of crop investigation. However, most of the research has currently focused on the classification of pixel-based image data for analysis. The study was carried out to develop a multi-category crop hyperspectral image classification system to identify the major crops in the Chiayi Golden Corridor. The hyperspectral image data from CASI (Compact Airborne Spectrographic Imager) were used as the experimental data in this study. A two-stage classification was designed to display the performance of the image classification. More specifically, the study used a multi-class classification by support vector machine (SVM) + convolutional neural network (CNN) for image classification analysis. SVM is a supervised learning model that analyzes data used for classification. CNN is a class of deep neural networks that is applied to analyzing visual imagery. The image classification comparison was made among four crops (paddy rice, potatoes, cabbages, and peanuts), roads, and structures for classification. In the first stage, the support vector machine handled the hyperspectral image classification through pixel-based analysis. Then, the convolution neural network improved the classification of image details through various blocks (cells) of segmentation in the second stage. A series of discussion and analyses of the results are presented. The repair module was also designed to link the usage of CNN and SVM to remove the classification errors.

Estimation of Surface Downward Shortwave Radiation over China from Himawari-8 AHI Data Based on Random Forest

Downward shortwave radiation (RS) drives many processes related to atmosphere–surface interactions and has great influence on the earth’s climate system. However, ground-measured RS is still insufficient to represent the land surface, so it is still critical to generate high accuracy and spatially continuous RS data. This study tries to apply the random forest (RF) method to estimate the RS from the Himawari-8 Advanced Himawari Imager (AHI) data from February to May 2016 with a two-km spatial resolution and a one-day temporal resolution. The ground-measured RS at 86 stations of the Climate Data Center of the Chinese Meteorological Administration (CDC/CMA) are collected to evaluate the estimated RS data from the RF method. The evaluation results indicate that the RF method is capable of estimating the RS well at both the daily and monthly time scales. For the daily time scale, the evaluation results based on validation data show an overall R value of 0.92, a root mean square error (RMSE) value of 35.38 (18.40%) Wm−2, and a mean bias error (MBE) value of 0.01 (0.01%) Wm−2. For the estimated monthly RS, the overall R was 0.99, the RMSE was 7.74 (4.09%) Wm−2, and the MBE was 0.03 (0.02%) Wm−2 at the selected stations. The comparison between the estimated RS data over China and the Clouds and Earth’s Radiant Energy System (CERES) Energy Balanced and Filled (EBAF) RS dataset was also conducted in this study. The comparison results indicate that the RS estimates from the RF method have comparable accuracy with the CERES-EBAF RS data over China but provide higher spatial and temporal resolution.

Spatial Assessment of Solar Radiation by Machine Learning and Deep Neural Network Models Using Data Provided by the COMS MI Geostationary Satellite: A Case Study in South Korea

Although data-driven methods including deep neural network (DNN) were introduced, there was not enough assessment about spatial characteristics when using limited ground observation as reference. This work aimed to interpret the feasibility of several machine learning approaches to assess the spatial distribution of solar radiation on Earth based on the Communication, Ocean, and Meteorological Satellite (COMS) Meteorological Imager (MI) geostationary satellite. Four data-driven models were selected (artificial neural network (ANN), random forest (RF), support vector regression (SVR), and DNN), to compare their accuracy and spatial estimating performance. Moreover, we used a physical model to probe the ability of data-driven methods, implementing hold-out and k-fold cross-validation approaches based on pyranometers located in South Korea. The results of analysis showed the RF had the highest accuracy in predicting performance, although the difference between RF and the second-best technique (DNN) was insignificant. Temporal variations in root mean square error (RMSE) were dependent on the number of data samples, while the physical model showed relatively less sensitivity. Nevertheless, DNN and RF showed less variability in RMSE than the others. To examine spatial estimation performance, we mapped solar radiation over South Korea for each model. The data-driven models accurately simulated the observed cloud pattern spatially, whereas the physical model failed to do because of cloud mask errors. These exhibited different spatial retrieval performances according to their own training approaches. Overall analysis showed that deeper layers of networks approaches (RF and DNN), could best simulate the challenging spatial pattern of thin clouds when using satellite multispectral data.

Examining Interactions Between and Among Predictors of Net Ecosystem Exchange: A Machine Learning Approach in a Semi-arid Landscape

Net ecosystem exchange (NEE) is an essential climate indicator of the direction and magnitude of carbon dioxide (CO₂) transfer between land surfaces and the atmosphere. Improved estimates of NEE can serve to better constrain spatiotemporal characteristics of terrestrial carbon fluxes, improve verification of land models, and advance monitoring of Earth's terrestrial ecosystems. Spatiotemporal NEE information developed by combining ground-based flux tower observations and spatiotemporal remote sensing datasets are of potential value in benchmarking land models. We apply a machine learning approach (Random Forest (RF)) to develop spatiotemporally varying NEE estimates using observations from a flux tower and several variables that can potentially be retrieved from satellite data and are related to ecosystem dynamics. Specific variables in model development include a mixture of remotely sensed (fraction of photosynthetically active radiation (fPAR), Leaf Area Index (LAI)) and ground-based data (soil moisture, downward solar radiation, precipitation and mean air temperature) in a complex landscape of the Reynolds Creek Experimental Watershed (RCEW) in southwest Idaho, USA. Predicted results show good agreement with the observed data for the NEE (r² = 0.87). We then validate the temporal pattern of the NEE generated by the RF model for two independent years at the two sites not used in the development of the model. The model development process revealed that the most important predictors include LAI, downward solar radiation, and soil moisture. This work provides a demonstration of the potential power of machine learning methods for combining a variety of observational datasets to create spatiotemporally extensive datasets for land model verification and benchmarking.