Observing at-surface irradiance and albedo from space: the Tibet experiment

Abstract

Monitoring the solar radiation budget on a daily basis is a prerequisite to study land surface processes, especially in climatology and hydrology, and in derived applications like drought early warning. Current space-born radiometers can provide daily observations to derive surface radiative fluxes at a spatial resolution of one square kilometer at best, which is not enough to properly take into account surface heterogeneity in many regions of the globe. As part of a major scientific initiative to study the hydrology of the Tibetan Plateau -especially known for its rough topography- this thesis focuses on developing a method to adequately estimate at-surface daily solar radiation budget over this particular area. Following a first effort to produce a time series of the radiative budget from existing satellite data products, it appeared that it is necessary to consider terrain and clouds spatial variability at the sub-pixel level when working over heterogeneous areas such as the Tibetan Plateau. Thus, the impact of spatial and temporal variability of clouds on solar radiation was investigated through a case study conducted on the field whose results demonstrate that the surface irradiance estimation would benefit from using cloud distribution instead of cloud fraction to account for the cloud cover. Furthermore, a high temporal resolution cloud cover leads to a better temporal average of the radiative fluxes. Regarding the effects of the terrain, a new sub-pixel topographic correction method is proposed and applied. It demonstrates that the integration of the sub-pixel topographic effects using high resolution DEM improves the irradiance as well as the albedo retrieval. The temporal resolution of the latter is also to consider and the use of geostationary satellite allows to increase the retrieval frequency. Based on multi-source and multi-resolution remote sensing data, the developed method provides a usable solar radiation budget dataset. Combined with the outcomes of the in-depth researches on clouds and topography, it paves the way for a new operational methodology which adequately accounts for sub-pixel heterogeneity when producing large area time series of solar radiation budget at the surface.