Spatial scale effects on retrieval accuracy of surface solar radiation using satellite data

Hou Jiang, Ning Lu, Guanghui Huang, Ling Yao, Jun Qin and Hengzi Liu

Applied Energy, 2020, vol. 270, issue C, No S0306261920306905

Abstract: The presence of nonhomogeneous clouds and their induced radiation interactions result in significant horizontal photon transport and spatially adjacent effects on surface solar radiation (SSR), making spatial estimation scale-dependent. Overlooking scale effects during SSR retrieval from satellite data is responsible for variations in retrieval accuracy and deviations in associated applications. In this paper, the spatial scale effects on SSR retrieval accuracy are investigated using multivariate linear regression and artificial neural network and convolutional neural network models. Scale effects are quantified through changes in retrieval accuracy under varying satellite data input size and compared among different models to reveal the merits and defects of classic linear, ordinary nonlinear, and spatially nonlinear models. The results show that scale effects have considerable impacts on retrieval accuracy in each of the three models for both site-specific and general conditions. The maximum improvement in terms of the root mean square error can reach up to 9% after involving scale information. The performance of site-specific models is continually enhanced with the expansion of spatial scale, while that of general models will drop to some extent beyond a particular threshold. Approximate distances of 20 km and 40 km from the central location are identified as the optimal scale for artificial neural and convolutional neural networks, respectively. This study also concludes that the robustness of general models is relevant to various atmospheric factors, providing perspectives for further improvements including the fusion of time series images, integration with physical modules, and the combination of multi-resolution data.

Keywords: Scale effect; Surface solar radiation; Convolutional neural network; Artificial neural network; Multivariate linear regression (search for similar items in EconPapers)
Date: 2020
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (6)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0306261920306905
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:appene:v:270:y:2020:i:c:s0306261920306905

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/405891/bibliographic
http://www.elsevier. ... 405891/bibliographic

DOI: 10.1016/j.apenergy.2020.115178

Access Statistics for this article

Applied Energy is currently edited by J. Yan

More articles in Applied Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().