Global Solar Radiation Data Research Articles

Near Real-Time Global Solar Radiation Forecasting at Multiple Time-Step Horizons Using the Long Short-Term Memory Network

This paper aims to develop the long short-term memory (LSTM) network modelling strategy based on deep learning principles, tailored for the very short-term, near-real-time global solar radiation (GSR) forecasting. To build the prescribed LSTM model, the partial autocorrelation function is applied to the high resolution, 1 min scaled solar radiation dataset that generates statistically significant lagged predictor variables describing the antecedent behaviour of GSR. The LSTM algorithm is adopted to capture the short- and the long-term dependencies within the GSR data series patterns to accurately predict the future GSR at 1, 5, 10, 15, and 30 min forecasting horizons. This objective model is benchmarked at a solar energy resource rich study site (Bac-Ninh, Vietnam) against the competing counterpart methods employing other deep learning, a statistical model, a single hidden layer and a machine learning-based model. The LSTM model generates satisfactory predictions at multiple-time step horizons, achieving a correlation coefficient exceeding 0.90, outperforming all of the counterparts. In accordance with robust statistical metrics and visual analysis of all tested data, the study ascertains the practicality of the proposed LSTM approach to generate reliable GSR forecasts. The Diebold–Mariano statistic test also shows LSTM outperforms the counterparts in most cases. The study confirms the practical utility of LSTM in renewable energy studies, and broadly in energy-monitoring devices tailored for other energy variables (e.g., hydro and wind energy).

Predictive model and assessment of the potential for wind and solar power in Rayak region, Lebanon

With the increasing consumption of fossil fuel, reducing greenhouse gas (GHG) emissions has become a serious issue that has attracted worldwide attention. Therefore, Lebanon is currently interested in utilizing renewable energy technologies to reduce energy dependence on oil reserves and GHG emissions. The present study is focused on solar and wind power potential and the economic viability of wind/solar systems for the Rayak region in Lebanon for the first time. The input data sources for the study including the Meteorological data for wind speed and NASA database for solar energy. In the assessment of wind energy, a two-parameter Weibull distribution function was used to analyze the characteristics of wind speed. Yearly and seasonal Weibull parameters were calculated for 10 m height using the Maximum likelihood method. In addition, yearly and seasonal wind power density values were calculated. The results showed that the mean annual wind speed and wind power density values were 5.884 m/s and 124.534 W/m2, respectively, during the investigation period. It can be concluded that the value of wind power density at the region was classified as marginal wind power potential and high-scale wind turbines can be used to gather wind energy potential in the region. Furthermore, predicting wind speed depends on various atmospheric factors and random variables. Therefore, 63 ANN models are developed by varying the meteorological parameters to predict the daily wind speed in the selected region. All the models with various combinations are validated and the performances of the models are analyzed using root mean squared error. The results demonstrated that the most relevant input variables for predicting the daily wind speed were found to be temperature, pressure, and relative humidity. In the assessment of wind energy, average monthly global solar radiation data were evaluated. It is found that the annual global solar radiation was 1877.41 kWh/m2, which indicates the selected region has high solar resources and is categorized as an excellent potential class. Moreover, this study provides a comprehensive and integrated feasibility analysis of 100 MW grid-connected wind and solar projects economic projects that can be developed in the country to reduce the electricity crisis and GHG emissions. Several different economic and financial indicators were calculated. The results indicate that the wind farm is a more economical option than the solar plant because of the higher values of NPV, BCR, ALCS, and IRR as well as the lower values of EB, SB, and LCOE.

Modelling of the coupling of desalination plants with the thermal solar energy system

Abstract This work aims to develop models able to describe and predict the thermal solar collector coupled with vacuum membrane distillation (VMD) unit to determine the growth evolution of the temperature in the solar collector and the membrane module of the VMD. These models will be used to evaluate the distillate flow rates during the year. The global solar radiation data recorded on an inclined surface as well as ambient temperature data can be included in a program. First, the models were validated with some experimental data studies from appropriate literature and then they were used to predict the unit's performance under different operating conditions in the City of Bouzaréah (Algiers). The results obtained from the proposed models showed that the distillate flux water reached 10.5 kg/h·m2 in September and around 5.6 kg/h·m2 in December during midday.

Estimation of solar radiation using stepwise multiple linear regression with principal component analysis in Algeria

For the modelling of solar systems, reliable and complete time series of solar radiation are required. However, solar radiation data from ground measurements or satellite images may be available only for limited time periods and often have data quality issues or/and data gaps. It is possible to get reasonably accurate radiation estimates of solar radiation from meteorological parameters, which can complement or extend existing data. In this paper, a model based on eight meteorological parameters (evaporation, temperature, wind speed, visibility, cloud cover, sunshine duration, sunshine ratio and global solar radiation) is developed to predict the global solar radiation using 29-year (1986–2014) data from Oran radiometric station in Algeria. Two versions of the multiple regression analysis are used: the first with original observation variables (called as manifest variables) and the second by PCs called as latent variables or (common) factors. Separate analyses have been carried out for two scenarios: overall and partial study. The first scenario uses all climatic dataset, but the last scenario uses three subsets which are sunny, partly cloudy and cloudy days. This classification has been obtained using certain statistical properties that we have considered as thresholds. Stepwise multiple linear regression analysis is used to fit global solar radiation data using meteorological variables as predictors. A variable selection method based on PCA technique are used to obtain the subsets of predictors to be included in the regression model of global solar radiation data. Hence, the proposed relationships may be considered useful for predicting global solar radiation rate in other sites of Algeria that have climatic conditions similar to the study area.

Global Solar Radiation Characteristics at Calabar and Port Harcourt Cities in Nigeria

This study analyzed the inter-annual variability in solar radiation at Port Harcourt and Calabar, aiming at improving knowledge of solar resources. For the investigation, monthly mean global solar radiation data for fifteen years (2000 – 2014) was collected from Nigerian Meteorological Agency (NIMET), and the monthly mean extraterrestrial solar radiation was determined using globally recognized standard relation. The clearness index parameter was employed for characterizing the spatial variability of solar radiation for Calabar and Port Harcourt. The statistics of the monthly mean solar radiation deviations of Port Harcourt and Calabar was tested using the Kolmogorov–Smirnov method. The test results showed that they are normally distributed random variables. Furthermore, the analysis of sequential properties showed that the coefficients of the auto-correlation with lag 1 are significant for both stations. The auto-correlation coefficients with lag 1, though usually not significant, are negative for both stations. The auto regression lag 1 (AR-1) is the recommended procedure (model equation) for generating monthly solar radiation synthetic time series, with auto-correlation coefficients varying from 0.30 to 0.47 for both stations in the South-South of Nigeria. Citation: Amadi, S. (2020). Global Solar Radiation Characteristics at Calabar and Port Harcourt Cities in Nigeria. Trends in Renewable Energy, 6, 101-120. DOI: 10.17737/tre.2020.6.2.00114

Time series forecasting on multivariate solar radiation data using deep learning (LSTM)

Energy management is an emerging problem nowadays and utilization of renewable energy sources is an efficient solution. Solar radiation is an important source for electricity generation. For effective utilization, it is important to know precisely the amount from different sources and at different horizons: minutes, hours, and days. Depending on the horizon, two main classes of methods can be used to forecast the solar radiation: statistical time series forecasting methods for short to midterm horizons and numerical weather prediction methods for medium- to long-term horizons. Although statistical time series forecasting methods are utilized in the literature, there are a limited number of studies that utilize deep artificial neural networks. In this study, we focus on statistical time series forecasting methods for short-term horizons (1 h). The aim of this study is to discover the effect of using multivariate data on solar radiation forecasting using a deep learning approach. In this context, we propose a multivariate forecast model that uses a combination of different meteorological variables, such as temperature, humidity, and nebulosity. In the proposed model, recurrent neural network (RNN) variation, namely a long short-term memory (LSTM) unit is used. With an experimental approach, the effect of each meteorological variable is investigated. By hyperparameter tuning, optimal parameters are found in order to construct the best models that fit the global solar radiation data. We compared the results with those of previous studies and we found that the multivariate approach performed better than the previous univariate models did. In further experiments, the effect of combining the most effective parameters was investigated and, as a result, we observed that temperature and nebulosity are the most effective parameters for predicting future solar radiance.

Estimating Daily Global Solar Radiation with No Meteorological Data in Poland

The aim of the study was to calibrate coefficients and evaluate performance of simple, day-of-the-year, global solar radiation (H) models nominated from the literature. Day-of-the-year models enable estimation of global solar radiation when no meteorological data is available. The study used 16-year-long data series of daily H, taken at 15 actinometric stations located in various parts of Poland. The goodness-of-fit of the models to the actual long-term monthly average daily global solar radiation data expressed by determination coefficient (R2) ranges from 0.94 to 0.97. Depending on statistical indicators analysis (root mean square error—RMSE, mean absolute bias error—MABE, mean average percentage error—MAPE) the best model was selected. The averaged values of H computed by the recommended model deviate from those measured by 4.16% to 8.71%. Locally calibrated, day-of-the-year model provides satisfactory accuracy and—where meteorological data is unavailable—can be used to estimate mean monthly daily global solar radiation in Poland and similar climate conditions.

Developing a new model for predicting global solar radiation on a horizontal surface located in Southwest Region of Algeria

ABSTRACT Global solar radiation data is essential in the design of solar energy conversion devices. The most widely used model in the estimation of the global solar radiation coming to the earth’s surface is the Angtröm-Prescott model. In this model, relative sunshine duration plays an important role. In the current study, it was attempted to review different models and to determine the best performing model developed for the city of Bechar and two locations of Algerian Big South desert for solar energy projects and compare them with the different established models available in the literature up until now, based on measured sunshine hours data. The new models give high accuracies on the application, where (R2) ranges between 0.70 and 0.99. In addition, MBE and RMSE values are very low in the studied stations. The results presented in this paper are quite useful for quick estimation of global solar radiation data required for performance evaluation of solar collector used for different applications of solar energy.

Coupling meteorological variables with Moderate Resolution Imaging Spectroradiometer atmospheric products for estimating global solar radiation

Global solar radiation is a crucial variable for scientific researches and solar energy application, while it is measured at very few sites mainly due to the technical and fiscal obstacles. Developing robust and accurate models for estimating global solar radiation had been being a focus for many studies. This study was conducted to develop integrated models combining Moderate Resolution Imaging Spectroradiometer atmospheric products and meteorological variables. 43 empirical models based on the meteorological variables were collected. A total of 645 integrated models incorporating atmospheric constituents into the empirical models were developed. The researched models were evaluated and compared at Chongqing in Three Gorges Reservoir Area, China. The results showed that the integrated models outperformed the empirical models. The best integrated model had the root mean square error of 0.817 MJ m−2 and relative root mean square error of 8.11%. On average, the integrated models had the root mean square error of 1.071 MJ m−2, 15.6% smaller than the empirical models. The results suggest that coupling Moderate Resolution Imaging Spectroradiometer atmospheric products with meteorological variables can enhance the performance of the conventional empirical models, which may provide a promising alternative to generate global solar radiation data with better accuracy.

Global Solar Radiation Models in Minas Gerais, Southeastern Brazil

Accurate and complete global solar radiation (Hs) data at a specific region are crucial for regional climate assessment, crop growth modeling, and all operations that use solar energy. However, in the Minas Gerais state, Southeastern Brazil (SEB), the number of weather stations that measure global solar radiation is scarce, and when it is available, it presents gaps in the time series. An attractive alternative to solve the data gap problem is to estimate global solar radiation using empirical models. In this study, thirteen models based on maximum and minimum air temperatures, precipitation, sunshine duration, and extraterrestrial solar radiation were compared in the daily estimation ofHs.Data from 10 weather stations, from 1999 to 2017, located in Minas Gerais were used. Also, cluster analysis was used to group the localities (weather stations) with similar patterns of model performance, climatic classification (Köppen–Geiger and Thornthwaite), and seasonal data variability, considering minimum and maximum air temperatures, precipitation, sunshine duration, and global solar radiation. Although it is apparently simple, studies on this subject are scarce and the few existing ones in Minas Gerais have flaws, which justifies this study. The models were evaluated by root mean square error (RMSE), mean absolute percentage error (MAPE), mean bias error (BIAS), Willmott’s index of agreement (d), and performance index (c-index). Models based on sunshine duration, such as those proposed by Ertekin and Yaldiz and by Newland, showed the best performance (averagec-index = 0.71). Models based on temperature and precipitation showed the worst results (averagec-index = 0.41). Cluster analysis showed that there is a similar pattern between the performance of the models, climatic classification, and seasonal variability of data among the localities of Minas Gerais. In general, models that presented extremely poor performance were formed with weather stations located in the dry zone, but with different climate classification, and models that presented very good (and good) performance were composed by weather stations located in the humid zone (dry subhumid) with the same climate classification and similar seasonal variability. Furthermore, the models based on temperature have a tendency to overestimate radiation values below 10 MJ·m−2 day−1and to underestimate values higher than 25 MJ·m−2 day−1. This point is a limitation of the model for estimating global solar radiation below and above these levels, showing the influence of atmospheric systems and atmospheric attenuation mechanisms of global solar radiation.

On the importance of relative humidity for prediction of diffuse solar radiation in warm and humid climate of Iran

ABSTRACTIn numerous applications of the solar energy, diffuse solar radiation plays crucial role while their availability in many countries is scarce. Many models have been developed to estimate the diffuse solar radiation that each of them is a function of different metrological parameters. The objective of this study is obtaining a model to estimate the diffuse solar radiation in warm and humid climate in Iran and specify the importance of metrological parameters especially relative humidity. Many empirical models in three categories are selected. By using the long-term data for horizontal global solar radiation and sunshine duration and by employing the regression analysis, nine new models have been created. Then, by entering two parameters, the monthly mean daily ambient temperature and relative humidity in the empirical models, nine other new models have been developed. Performances of all the models are evaluated by means of different statistical indicators. Obtained results showed adding of the ambient temperature and especially relative humidity parameters cause increasing a few accuracy even though in warm and humid climate of Iran. The result shown in the two-degree model, in terms of the clearness index and relative sunshine duration, is the best model.

Solcast: Validation of a satellite-derived solar irradiance dataset

Solcast (https://solcast.com/) is a global solar forecasting and historical solar irradiance data company. Their datasets are important to the scientific community as substantial amounts of historical data is made freely available to researchers. Solcast’s hourly averages and sub-hourly instantaneous (10/15-min) global horizontal irradiance (GHI) estimates are validated at 48 BSRN ground stations that contain 1,014,792 valid hourly average measurements and 2,932,976 valid instantaneous measurements. The data are validated globally and across four climates under clear-sky, clouded-sky and all-sky periods, repeated for ‘all zenith angles’ (θ_z<85°) and high-sun zenith angles (θ_z<50°). A site-by-site validation is also performed for both hourly and instantaneous datasets. All validations and recommendations are impartial and unbiased.The hourly dataset validated similarly for both zenith angle scenarios, which is remarkable due to the added complexities at low-sun angles. In terms of rRMSE, the global performances of hourly GHI under clear-sky, clouded-sky and all-sky were 3.4%, 25.6% and 16.9%, respectively. With each ground station equally weighted, the overall bias was 0.33%, and standard deviation of biases at 1.7%. Solcast’s instantaneous GHI product validations were at times similar to the hourly validations though is naturally harder with increasing temporal resolution.The historical datasets from Solcast are found to be of good quality, thus, they are recommended for use by researchers. Identified areas for improvement herein will guide future developments of Solcast’s methodology.

New approach to estimate 5-min global solar irradiation data on tilted planes from horizontal measurement

This paper presents a new methodology to estimate global tilted irradiation in 5-min steps using only global irradiation on the horizontal plane. The methodology is based on a combination of the two well-known conventional Perrin Brichambaut and Liu and Jordan models. Tilted irradiation is of great importance for the design and short-term performance assessment of fixed-tilt flat-plate collectors and photovoltaic (PV) systems. Intermediate key parameters of the state of the sky, referred to here as cloudiness factors, are determined and introduced to transform isotropic models into an anisotropic model. The results of the proposed model for all sky conditions with regard to normalized root mean square error (nRMSE), relative percentage error (RPE), normalized mean absolute error (nMAE) and coefficient of correlation (R2) range from 4.7 to 6.41%, 5.5–5.9%, 3.07–4.73% and 0.97 to 0.99, respectively, which are very accurate results, especially for such a short time step. A comparison with the best conventional models and even artificial neural network (ANN) models described in the literature has confirmed that the developed model has smaller errors.

New temperature dependent models for estimating global solar radiation across the midland climatic zone of Nigeria

Authentic information of the availability of global solar radiation is significant to agro/hydro meteorologists, atmospheric Physicists and solar energy engineers for the purpose of local and international marketing, designs and manufacturing of solar equipment. In this study, five new proposed temperature dependent models were evaluated using measured monthly average daily global solar radiation, maximum and minimum temperature meteorological data during the period of thirty one years (1980-2010). The new models were compared with three existing temperature dependent models (Chen et al., Hargreaves and Samani and Garcia) using seven different statistical validation indicators of coefficient of determination (R2), Mean Bias Error (MBE), Root Mean Square Error (RMSE), Mean Percentage Error (MPE), t – test, Nash – Sutcliffe Equation (NSE) and Index of Agreement (IA) to ascertain the suitability of global solar radiation estimation in five different locations (Zaria, Bauchi, Jos, Minna and Yola) situated in the Midland climatic zone of Nigeria. In each location, the result shows that a new empirical regression model was found more accurate when compared to the existing models and are therefore recommended for estimating global solar radiation in the location and regions with similar climatic information where only temperature data are available. The evaluated existing Hargreaves and Samani and Garcia temperature based models for Jos were compared to those available in literature and was found more suitable for estimating global solar radiation for the location. The comparison between the measured and estimated temperature dependent models depicts slight overestimation and underestimation in some months with good fitting in the studied locations. However, the recommended models give the best fitting.  Â

Evaluation of temperature-based machine learning and empirical models for predicting daily global solar radiation

Accurate global solar radiation data are fundamental information for the allocation and design of solar energy systems. The current study compared different machine learning and empirical models for global solar radiation prediction only using air temperature as inputs. Four machine learning models, e.g., hybrid mind evolutionary algorithm and artificial neural network model, original artificial neural network, random forests and wavelet neural network, as well as four empirical temperature-based models (Hargreaves-Samani model, Bristow-Campbell model, Jahani model, and Fan model) were applied for prediction of daily global solar radiation in temperate continental regions of China. The results indicated the hybrid mind evolutionary algorithm and artificial neural network model provided better estimations, compared with the existing machine learning and empirical models. Thus, the temperature-based hybrid model is highly recommended to predict global solar radiation in temperate continental regions of China when only air temperature data are available. Combining the hybrid model with future air temperature forecasts, we can get the accurate information of future solar radiation, which is of great importance to management and operation of solar energy systems.

Daily reference evapotranspiration prediction based on hybridized extreme learning machine model with bio-inspired optimization algorithms: Application in contrasting climates of China

Reliable and accurate prediction of reference evapotranspiration (ETo) is a precondition for the efficient management and planning of agricultural water resources as well as the optimal design of irrigation scheduling. This study evaluated the performances of four bio-inspired algorithm optimized extreme learning machine (ELM) models, i.e. ELM with genetic algorithm (ELM-GA), ELM with ant colony optimization (ELM-ACO), ELM with cuckoo search algorithm (CSA) and ELM with flower pollination algorithm (ELM-FPA), for predicting daily ETo across China by using a five-fold cross-validation approach. These models were further compared with the classical ELM model parameterized by the grid search method to demonstrate their capability and efficiency. Daily maximum and minimum ambient temperatures, wind speed, relative humidity and global solar radiation data during 2001–2015 collected from eight meteorological stations in contrasting climates of China were utilized to train, validate and test the models. The results showed that ETo values predicted by all ELM models agreed well with the corresponding FAO-56 Penman–Monteith values, with R2, RMSE, NRMSE and MAE ranging 0.9766–0.9967, 0.0896–0.2883 mm day−1, 3.2910%-11.7653% and 0.0708–0.1998 mm day−1, respectively. The ELM-FPA model (R2 = 0.9930, RMSE = 0.1589 mm day−1, NRMSE = 5.5406% and MAE = 0.1188 mm day−1) slightly outperformed the ELM-CSA model (R2 = 0.9922, RMSE = 0.1619 mm day−1, NRMSE = 5.6864% and MAE = 0.1200 mm day−1) during testing, both of which were superior to the ELM-ACO (R2 = 0.9912, RMSE = 0.1730 mm day−1, NRMSE = 6.0816% and MAE = 0.1254 mm day−1) and ELM-GA (R2 = 0.9889, RMSE = 0.1895 mm day−1, NRMSE = 6.7197% and MAE = 0.1310 mm day−1) models, followed by the standalone ELM model (R2 = 0.9856, RMSE = 0.2104 mm day−1, NRMSE = 7.1693% and MAE = 0.1408 mm day−1). The four hybrid ELM models exhibited higher improvements in daily ETo prediction in the temperate monsoon and (sub)tropical monsoon climates (with average decrease in RMSE of 14.0%, 25.1%, 31.4% and 33.1%, respectively), compared with those in the temperate continental and mountain plateau climates (with average decrease in RMSE of 5.2%, 9.8%, 12.9% and 14.1%, respectively). The results advocated the capability of bio-inspired optimization algorithms, especially the FPA and CSA algorithms, for improving the performance of the conventional ELM model in daily ETo prediction in contrasting climates of China.

Simplified model to correct thermopile pyranometer solar radiation measurements for photovoltaic module yield estimation

This article presents and evaluates the performance of a simplified model to generate 10-min global horizontal synthetic solar radiation data that would correspond to the measurements of a calibrated photovoltaic monocrystalline cell. The model, which only requires global horizontal solar radiation data measured with a thermopile pyranometer as input, is based on the characterization of the relation between the data measured with a thermopile pyranometer and a calibrated cell as a function of the sky condition and the solar elevation. We have used an extensive solar radiation database for the location of Seville (Spain) for the training of the model that has been tested in Seville and Lancaster (USA), showing satisfactory results and suggesting a global applicability with no local adaptation or calibration requirement.The model shows the best results for high levels of solar radiation and solar elevations and decreases its performance on days with high levels of diffuse irradiation and for very low solar elevation angles. We obtain a daily RMSD between measured and synthetic data of 1.9% in Seville and 5.2% in Lancaster. The frequency distribution of the synthetic datasets shows a KSI of 3.7 W/m2 in Seville and 8.6 W/m2 in Lancaster. We also evaluate the ramp rates of measured and synthetic sets through the KSI of the measured and synthetic ramp rates sets, obtaining 0.11 W/m2۰min in Seville and 0.20 W/m2۰min in Lancaster.

Review on mathematical models for the prediction of solar radiation

Global Solar Radiation (GSR) data is important for all solar energy based applications, mainly to forecast the output power of solar PV system in case of renewable energy integration in to the existing grid. The solar radiation components are measured using pyranometer, solarimeter, pyroheliometer and so on. It is not practically possible to install this radiation measuring instruments at all the locations due to the cost and difficulty in measurements. Hence the availability of solar radiation data is limited to few meteorological stations especially in the developing country like India. Therefore, it is necessary to develop mathematical models to predict the solar radiation to eliminate the costly pyranometer. In this paper, the review of mathematical models using trigonometric functions for the prediction of global solar radiation is presented. The mathematical models are applicable wherever the radiation data is unavailable. From the review results, it is concluded that mathematical model with both sine and cosine wave equation gives good prediction accuracy with correlation coefficient of 0.95

Econometric estimates of Earth’s transient climate sensitivity

How sensitive is Earth’s climate to a given increase in atmospheric greenhouse gas (GHG) concentrations? This long-standing question in climate science was recently analyzed by dynamic panel data methods using extensive spatio-temporal data of global surface temperatures, solar radiation, and GHG concentrations over the last half century to 2010 (Storelvmo et al, 2016). Those methods revealed that atmospheric aerosol effects masked approximately one-third of the continental warming due to increasing GHG concentrations over this period, thereby implying greater climate sensitivity to GHGs than previously thought. The present study provides regularity conditions and asymptotic theory justifying the use of time series cointegration-based methods of estimation when there are both stochastic process and deterministic trends in the global forcing variables, such as GHGs, and station-level trend effects from such sources as local aerosol pollutants. The asymptotics validate estimation and confidence interval construction for econometric measures of Earth’s transient climate sensitivity (TCS). The methods are applied to observational data and to data generated from several groups of global climate models (GCMs) that are sampled spatio-temporally and aggregated in the same way as the empirical observations for the time period 1964–2005. The findings indicate that 7 out of 9 of the GCM reported TCS values lie within the 95% empirical confidence interval computed econometrically from the GCM output. The analysis shows the potential of econometric methods to provide empirical estimates and confidence limits for TCS, to calibrate GCM simulation output against observational data in terms of the implied TCS estimates obtained via the econometric model, and to reveal the respective sensitivity parameters (GHG and non-GHG related) governing GCM temperature trends.

Machine learning regressors for solar radiation estimation from satellite data

In this paper we evaluate the performance of several Machine Learning regression techniques in a problem of global solar radiation estimation from geostationary satellite data. Different types of neural networks, Support Vector Regression and Gaussian Processes have been selected as regression techniques to be evaluated, due to their good performance in similar problems in the past. The study area is located in the surroundings of the radiometric station of Toledo, Spain. In order to train the regression techniques considered, one complete year of hourly global solar radiation data is used as the target of the experiments, and different input variables are considered: a cloud index, a clear-sky solar radiation model and several reflectivity values from Meteosat visible images. To assess the results obtained by the Machine Learning algorithms, we have selected as a reference three different physical-based methods, a model based on the Heliosat-2 method (Heliosat-2), the Copernicus Atmosphere Monitoring Service (CAMS) and the SolarGIS model (Soevaluate the performance of Machine Learning regressors when the physical models are included as input variables, in a class of post-processing of these physical approaches. The results obtained show the capacity of Machine Learning regressors to obtain reliable global solar radiation estimation by using satellite measurements.