饱和修正系数提高ZZLAS型闪烁仪测量显热通量精度

为了研究饱和效应对大孔径闪烁仪估算区域显热通量的影响,在2014年8—9月期间进行试验,以具有抗饱和性能的双光路BLS900型闪烁仪为参考,以孔径尺寸为0.075 m(文中简记为LAS1)、0.15 m(文中简记为LAS2)的中国产ZZLAS型闪烁仪为研究对象,通过光学传播原理计算出饱和修正系数,并对ZZLAS型闪烁仪的观测结果进行饱和修正分析。结果表明:LAS1的饱和修正系数为1.034,LAS2的饱和修正系数为1.019。试验观测中LAS1饱和率为24.58%,LAS2饱和率为2.04%,进行饱和修正后,LAS1的有效饱和修正率为12.87%。与BLS900相比,LAS1修正后显热通量均方根误差变为25.67 W/m2;LAS2的饱和修正率仅为0.32%,修正前后显热通量均方根误差基本无变化。进行饱和修正前,对BLS900与LAS1观测的显热通量进行F检验,未达到显著水平(P=0.15);通过计算得出的修正系数修正后,达到极显著水平(P=0.004);而利用BLS900的实时修正系数进行修正后,接近显著水平(P=0.06)。利用试验计算得出的饱和修正系数修正后,显热通量与参考标准的误差范围为1.28~53.42 W/m2,比修正前更接近BLS900的观测结果。研究对农田、人工林下垫面条件下的观测结果采用文中的饱和修正方法进行验证,结果也表明,经饱和修正后,闪烁仪观测的显热通量更接近BLS900的观测结果。当ZZLAS型闪烁仪发生明显饱和现象时,利用光学传播原理计算得出的修正系数对饱和数据进行修正效果显著。

Saturation correction factor improving sensible heat flux accuracy measured by ZZLAS scintillometer

Abstract: For the purpose of investigating the influence of saturation effect on the estimation of regional sensible heat flux by using large aperture scintillometer (LAS), a field test was conducted from August to September, 2014. The dual-optical-path BLS900 scintillometer, which possesses saturation resistance, was set as a reference, and the ZZLAS type scintillometer with aperture size of 0.075 m (LAS1) and 0.15 m (LAS2) was selected as research object. In this field test, the LAS1 was set as easily as possible to be saturated. The saturation correction coefficients calculated based on the optical propagation theory, and the real-time saturation correction coefficient calculated by BLS900 were also displayed, and the air structure parameters, heat fluxes measured from scintillometer were corrected with the coefficients in this study. Variables from LAS1 and LAS2 and that from BLS900 were compared and analyzed with fitted line, as well as root mean square error before and after correction, and F-test was also used in the test of sensible heat flux. The saturation of scintillometer is often caused by the turbulence, and the scale of turbulence is thought to the main reason. Based on the optical propagation theory, inner scale of turbulence was also taken into consideration and the saturation correction coefficient of ZZLAS scintillometer was calculated. The results from the observation were corrected, and comparisons and analysis were also made. Here are the findings. Saturation correction coefficients of LAS1 and LAS2 are 1.034 and 1.019 respectively. The real-time correction coefficients given by BLS900 range from 0.70 to 1.15 depending on the developing of turbulence. The calculated coefficient of BLS900 during 10:00-12:30 shows smaller difference with the real-time coefficients, and it is thought to be more suitable for this period. During the observation, the saturation rate of the LAS1 is 24.58% and the effective saturation correction rate is 12.87%. After correction, root mean square error of air refraction index changes from 1.003×10-13to 9.74×10-14 m-2/3, while there is no change occurring in R2. There is no obvious change of the air structure parameters between LAS2 and BLS900 due to that the saturation rate of LAS2 is much less than LAS1. Sensible heat fluxes from ZZLAS type scintillometer are compared to the results from BLS900, and the root mean square error of LAS1 is 25.67 W/m2. By contrast, the saturation rate of LAS2 is 2.04% and the saturation correction rate is only 0.32%, and there is no significant difference for LAS2 before and after correction. It indicates that, the more pronounced the saturation phenomenon, the more pronounced the effect of saturation correction; this is consistent with the opinion that there is no need to do corrections when the saturation data are less than 5%. Sensible heat fluxes measured from LAS1 exceed 50 W/m2, the corrected fluxes are more close to the reference, and the system error decreases. Sensible heat fluxes from BLS900 and LAS1 are analyzed with F-test and the P value obtained is 0.15, which means there is much difference between the 2 datasets. The P value becomes to 0.004 when the saturation data are corrected with the calculated coefficient, while the P value changes to 0.06 when saturation data are corrected with the real-time coefficients. After correcting the saturation correction coefficient obtained by experiment, the error range of sensible heat flux with reference standard is 1.28-53.42 W/m2, which is closer to the reference standard than that before correction. The results of the observation over the farmland and plantation are also verified by the saturation correction method in the paper. Results also showed that the sensible heat flux observed by the scintillometer after saturation correction is closer to that of the BLS900. When saturation data of the ZZLAS type scintillometer exceed 20%, there is significant improvement on sensible heat flux after the correction with the correction coefficient calculated by the optical propagation theory.