Satellite Remote Sensing Of Surface Net Radiation

Net radiation at the Earth's surface drives the process of evaporation, photosynthesis, and heating of soil and air. Net radiation and the overall surface energy budget are important for the development of the planetary boundary-layer. Its quantification over heterogenous land surfaces is crucial to study land-atmosphere interactions. Quantifying net radiation with high accuracy is significant for the whole physical environment and decision-making based on GIS. Undoubtedly, worldwide ground based measurements can provide the most accurate observations of net radiation, but given that direct measurements are expensive and limited, it is desirable to obtain global measurements of surface radiation fluxes using data provided by satellites.At present, the major satellite-estimate datasets of surface net radiation or its components are generally with low spatial resolution and low accuracy. While most of the methodologies to estimate surface net radiation using moderate/high spatial resolution satellite data are relying on meteorology data and only suitable for clear-sky conditions. Application of existing methodologies to estimate net radiation for cloudy-sky conditions from remote sensing sensors remains a significant challenge. Therefore, in this dissertation, we comprehensively evaluate the major satellite-estimate net radiation datasets and algorithms using ground measurements, present a framework to estimate instantaneous and daytime total net radiation under all sky conditions using MODIS data. Below are the major work conducted in this dissertation:Firstly, three satellite-estimated surface shortwave radiation data sets:1) GEWEX-SRB; 2) ISCCP-FD; and 3) CERES-FSW from 2000-2002, were evaluated using ground measurements at 36 sites around the world. Though the statistical errors are quite variable for the individual sites, systematic biases and tendencies may be identified. Results show that significant biases for some sites exceed the acceptable values of research community.Secondly, three satellite-estimated surface longwave radiation data sets:1) GEWEX-SRB; 2) ISCCP-FD; and 3) CERES-FSW of 2003, were evaluated using ground measurements at 15 sites around the world. Results show that, in general, the accuracy of these three datasets can meet the accuracy requirement of research community, and some uncertainties found in this evaluation can be explained by the elevation differences and input data accuracy.Thirdly, three algorithms of estimating surface radiation using MODIS data:1) Look-up table method of all-sky downwelling shortwave radiation; 2) Hybrid method of clear-sky downwelling longwave radiation; 3) Hybrid method of clear-sky upwelling longwave radiation, were evaluated using ground measurements. Results show that, these three methods have high accuracy and can meet the accuracy requirement of research community. Comparing with GEWEX-SRB, ISCCP-FD and CERES-FSW, these three MODIS algorithms outperform at most statistical indexes and agree better with the ground measurements.Fourthly, this dissertation presents a framework to estimate instantaneous net radiation under all sky conditions using only MODIS data. This framework consists of two parts:clear-sky algorithm and cloudy-sky algorithm, which establish a steady foundation for future research.Fifthly, this dissertation presents specific methods to estimate instantaneous net radiation using MODIS data, especially developed a new method to estimate instantaneous net radiation under cloudy-sky condition. Through validation, we found using the framework and specific algorithms proposed here can obtain better net radiation results than GEWEX-SRB, ISCCP-FD and CERES-FSW, besides, MODIS has higher spatial resolution.At last, this dissertation presents a method to estimate daytime total net radiation under all sky conditions using MODIS data. We applied'adjusted sinusoid interpolation'to the estimation of daytime total net radiation and validated using MODIS. Results show that, daytime total net radiation estimated by using this method agree well with the ground-measurements.