In a recent study published in Remote Sensing of Environment, a research team led by Prof. LI Rui from University of Science and Technology of China (USTC) developed a new satellite-based light use effective (LUE) model coupled with a passive microwave vegetation index (emissivity difference vegetation index, EDVI) to monitor the gross primary production (GPP) of terrestrial ecosystems. It is the first step toward the integration of microwave-derived variables into LUE model for daily GPP estimation.
Terrestrial ecosystems fix CO2 in the atmosphere through plant photosynthesis and generate GPP. This process is essential in the carbon cycle between land and atmosphere, affecting global carbon balance and climate change. It represents nature’s maximum ability to deal with carbon in the atmosphere, greatly influencing the way humans generate and use energy.
However, due to the complexity of terrestrial geographical systems, regional GPP is impossible to obtain on the ground. Thus it is necessary to use satellite-based remote sensing technology to observe this factor.
Most satellite methods use optical vegetation indices to describe the characteristics of vegetation. Specific models were input to estimate the exchange rate of carbon between the atmosphere and vegetation.
The problem with this traditional mode is that optical sensing relies upon reflected sunlight, rendering it powerless at night or in cloudy weather. There also exists the case of saturation in thick forests.
In contrast, microwave remote sensing monitors the heat radiation of the ground and the vegetation itself, regardless of sunlight. Also, microwaves have a good penetration effect on clouds, enabling it to ‘see clearly’ through the atmosphere.
Due to the fact that there are no previous studies using microwave remote sensing for this purpose, the research team independently developed the index EDVI, and creatively proposed EDVI-LUE, a light utilizing efficiency model based on microwave EDVI. This model has been successfully applied to typical forest, grassland and farmland vegetation in China.
Based on data gathered from seven flux tower sites (four forests, two grasslands and one cropland) of ChinaFLUX network, the results show that the new model has advantages over traditional methods. In certain environments, EDVI surpasses traditional methods in coping with complex weather and estimate accuracy.
To be specific, normalized EDVI, an indicator of canopy-scale leaf development and biomass change, was used as a proxy of fraction of photosynthetically active radiation. Compared with optical indicators, normalized EDVI (nEDVI) based FPAR is less likely to saturate in forests. nEDVI has higher sensitivity, and shows better relativity to GPP observed at sites. Also, compared with widely used products, this research provides a smaller estimate variation in forest and cropland sites, and matches other products in accuracy on heterogeneous grassland underlying surfaces.
On the other hand, in the situation of changing cloud condition, over ever-green broadleaf forests, the estimate accuracy indicator of EDVI-LUE performs capably and stably, presenting the advantage the model has in accurately estimating GPP in cloud-covered thick forest areas.
The novelty of this research has gained praise from RSE deputy editor, recognized vegetation remote scientist, Prof. Pablo. J. Zarco-Tejada, ‘…The three reviews indicate that the manuscript is innovative, as there are no studies using microwave data for assessing LUE to derive GPP. …’. This research offers a new way of quantifying the carbon fixing ability of terrestrial ecosystems in China. It provides scientific background for the ‘carbon neutrality’ and ‘carbon emission peak’ strategy.
Locations of seven flux tower sites. Land cover types are obtained from the International Geosphere-Biosphere Programme (IGBP) types of MCD12C1 product (Image by WANG Yipu et al.)
（Written by ZHAO Ziyue, edited by LI Xiaoxi, USTC News Center）