Alternatives for the assessment of the solar resource in Argentina

In Argentina, from 1972 on, several maps were presented which reported the distribution of global solar irradiation received on a horizontal plane placed at ground level and which used different time bases and information quality, whether estimates obtained from correlations established with other meteorological variables or direct irradiation measurements.n a paper by Grossi Gallegos (1998) 12 charts were presented with the monthly distribution of the mean value of daily global irradiation and one with their annual distribution, using all available information up to that moment in Argentina, whether from daily irradiation data obtained with Argentina s Solarimetric Network pyrano meters or sunshine hours measured by the National Meteorological Service (SMN) Network; the error due to the inclusion of estimates and interpolations was evaluated as lower than 10%.Argentina's Solarimetric Network underwent a continuous decrease in the number of operational stations due to the lack of resources for supporting them.In view of this situation, different alternatives were gradually evaluated which would make it possible to improve the already mentioned available global solar irradiation charts.In this sense, a statistical survey of the adjustment of satellite irradiation data available in Internet (in the base known as Surface Solar Energy (SSE) Data Set, Version 1.00) to the ground values.The objective was evaluating the possibility of using them as a complement to the data that had already been used and their application in order to obtain contour maps in homogeneous zones such as the Pampa Humeda, using geostatistical methods for drawing the irradiation isolines.Root-mean-square errors (RMSE) range from 3.7% to 24.8% depending on the inhomogeneity of the area. Nevertheless, the available temporal series are limited in time and thus their climatic representativity can be questioned.Given the shortage of solar irradiation measured data accurate enough to fulfill statistical conditions to be considered representative in time and space, there appeared alternatives for estimating them on the basis of the existing meteorological information, and one of the values used was the number of hours of bright sunshine.It was thus necessary to attempt rescuing the available information from almost two hundred heliographs distributed throughout the country.In some cases, such information dates back to the beginning of the XX century and a data base for the last 30 years, representative of the existing variations within the country, could easily be extracted from it.The use of relative sunshine hours in a relationship with global solar irradiation valued determines the need for taking the relative radiation at a certain value also; the linear relationship is thus more simple to state, as was proposed by Angstrom in 1924.Once maps with the temporal-spatial distribution of mean values of bright sunshine hours were available (that have been drawn applying geostatistical methods using a reasonable data base of bright), the following step is to attempt to obtain monthly charts from them with the spatial distribution of the average of daily global solar irradiation.The monthly values of the a and b constants used for each location were obtained and then, using geostatistical interpolation methods, their spatial distribution were determined in order to draw contour maps for them.Those maps, made using a limited number of stations where sunshine hours and solar radiation were simultaneously measured, make it possible to apply the linear equation throughout the country. Results so obtained were verified using available data of global irradiation giving mean bias error (MBE) of -4% and RMSE of 9.4%.In 2000 Suehrcke proposed a new relation between solar irradiation and sunshine hours, requiring the knowledge of the clear sky radiation.Were determined monthly mean values of clear sky irradiation from argentine solarimetric stations in 2005 and, using geostatistical methods for interpolation, spatial distribution of this values were obtained for each month. Verifying with surfaces MBE of 3.8% and RMSE of 9.4% were obtained. Another alternative that is being analyzed is the estimation of solar irradiation with Tarpley model in five locations of Argentina employing GOES 8 images.The constants of the model were locally calculated using ground pyrano meters data; the model has shown RMSE between 13% and 24% for daily values.Finally, a method for estimating global solar irradiation proposed by Hargreaves et al.in 1985, based on differences between maximun and minimum temperatures, was evaluated in the same way, concluding that is a valid alternative giving uncertainties similar to those of more known models (MBE of -3.7% and RMSE of 11%).Therefore, authors consider that these methodologies constitute a valuable alternative for countries with a high number of stations for measuring solar sunshine hours and temperature, with sufficiently wide registers, and few stations for simultaneous measurement of global solar irradiation using pyrano meters