Estimation methods developing with remote sensing information for energy crop biomass: A comparative review.

Abstract The rapid development of remote sensing (RS) technology enables an increased usage of high - resolution, spatial, temporal or spectral, data to extract vegetation information, improve model parameters, and estimate energy crop biomass accurately. Five estimation methods developing with RS information for energy crop biomass are summarized in this paper. Firstly, the statistical analysis with vegetation index can be regarded as the commonest method. But it is faulted for the deficiency of sample data and the vulnerability to the influence of many factors such as cloudy days. Secondly, the longer wavelength makes SAR information popular for crop biomass estimation. But many limitations lead to measurement uncertainty and bring about poor classification. Thirdly, NPP can directly reflect accumulated biomass production through photosynthesis and measure the consequences caused by climate change and human activities. But, actual LUE, the results of environmental stresses such as light intensity, temperature, water, and nutrients, is usually considered as a constant. Fourthly, crop height information is vitally important for biomass estimation. Yet the corresponding application is often subject to many factors e.g. crop variety, growth period, and farmland management practices limits. Lastly, the most promising approach lies in the possibility of assimilating state variables from RS data into CGMs. And the advent of latest instruments with better characteristics offer a unique chance to overcome current limitations. In the end, the main challenges and opportunities for energy crop biomass estimation using RS information in the future are listed. Highlights • Increasingly, remote sensing data is being used to estimate energy crops biomass. • Five estimation methods developing with remote sensing information are summarized. • Each estimation method has the strengths and weakness. • As the commonest method, the statistical analysis with VI has obvious drawbacks. • The most promising approach is to assimilate state variables from RS data into CGMs. [ABSTRACT FROM AUTHOR]