中红外波段地表发射率特性分析

陆表温度(LST)在地-气相互作用过程中扮演着重要的角色，是全球变化研究的关键参数。陆表发射率是陆表温度反演的关键输入参数之一。中红外谱区(3~5 μm)介于可见光-近红外谱区(0.38~2.5 μm)与热红外谱区(8~14 μm)之间，地物的发射率在该谱区表现出独特的光谱特性，可用于霜冻监测、矿物成分分析等研究。由于传感器在中红外谱区探测到的能量既有来自于地物自身发射的热辐射能量，又有反射的太阳辐射能量，这两部分的能量分离机理比较复杂，因此中红外谱区发射率特性分析的相关文献较少。本文针对单一均匀地表和具有混合像元的复杂地表计算了MODIS红外通道的有效发射率，发现通道有效发射率在单一均匀地表下与温度的耦合效应不强烈；但在复杂地表下，通道有效发射率与混合像元内的成分比例以及成分的地表温度具有耦合效应。在误差允许的范围内，混合像元的有效发射率可以忽略成分地表温度的影响。发射率误差对陆表温度反演精度的敏感性随着波长的变化而变化。在热红外波段，敏感性是其在中红外波段的2倍左右，说明利用中红外波段进行陆表温度反演具有一定的优势。

Analyses of Land Surface Emissivity Characteristics in Mid-Infrared Bands

Land surface temperature (LST) plays an important role in the process of ground-air interaction and is an parameter in global change research. At the same time, the emissivity of land surface is the key input parameter of LST inversion. Mid-infrared spectrum (3~5 μm) is between visible-near infrared (0.38~2.5 μm) and thermal infrared spectrum (8~14 μm). The emissivities of terrestrial materials exhibit unique characteristics in mid-infrared spectrum, which can be used for frost monitoring and mineral composition analysis et al. Energy detected with sensor in the mid-infrared region, however, is a combination of emitted radiation from terrestrial materials and reflected radiation due to sun irradiance. The energy separation mechanism of these two parts is complicated. Therefore, there are few relevant literatures about researches on emissivities of terrestrial materials in mid-infrared spectrum. In this paper, the effective emissivities of MODIS infrared channels were calculated for a single uniform surface and a complex region with mixed pixel. It is found that the effect of surface temperature is insignificant for effective emissivity calculation in a single uniform surface. Under complex surface, the effective emissivity has a coupling effect with composition ratio of materials in the mixed pixel and surface temperatures of these materials. Within the allowable range of error, the effective emissivity of mixed pixel can ignore the effect of materials’ surface temperature. The sensitivity of emissivity error to the precision of LST inversion varies with wavelength. An accuracy of 1 K in LST retrieval requires emissivity error be constrained to within 0.04 in mid-infrared region, while within 0.02 in thermal infrared region. It can be seen that mid-infrared spectrum has much potentials in LST retrieval.