Abstract
A spectrally integrated clear-sky and three-layer cloudy-sky models were developed to determine atmospheric transmittances and instantaneous surface insolation. Half-hourly observations at 8 km spatial resolution in optical and thermal infrared bands from an Indian geostationary satellite (Kalpana-1) Very High Resolution Radiometer (VHRR) sensor were used to provide inputs to these models in addition to global 8 day aerosol optical depth and columnar ozone. Sensitivity analysis of the clear-sky model showed a higher influence of aerosol on global insolation, diffuse insolation, and its fraction as compared with water vapour and ozone. The root mean square error (RMSE) of insolation estimates of the daily integral was found to be 2.05 MJ m−2 (∼11.2% of measured mean) with a high correlation coefficient (r = 0.93) when compared with in situ measurements during 1 August 2008 to 31 March 2010 over six locations in India. The errors were found to reduce to 7.5% over 3 to 5 day averages. The comparison of annual estimates and equivalent reanalysis fields showed a mean difference of the order of ±1.7 MJ m−2 over the majority of the Indian landmass.
Acknowledgements
This work was carried out under the ISRO-GBP EMEVS project. We are grateful to A.S. Kirankumar, Director, Space Applications Centre (ISRO), for his constant support in this work. The first author is thankful to Dr S.S. Ray, Head, Agro-Ecosystems Division and Dr S. Panigrahy, Group Director for their critical suggestions while revising the manuscript. Thanks are due to Miss Tarjani Vyas of Gota Engineering College, Gujarat, for prototype scripting of KIRAN and Mr Sunayan Saha of IARI, New Delhi, for his help during validation. The help from Dr Manoj Kumar, BIT, Ranchi, for providing in situ data is gratefully acknowledged. The suggestions given by anonymous reviewers helped improve the quality of the article.