MODIS产品反演与WRF模式默认植被覆盖度对比研究

A comparison study of fraction of green vegetation coverage estimated using MODIS products and default data used in WRF model

WRF模式气候态植被覆盖度GVF数据集的制作源于卫星遥感资料,由于制作时间较早代表性逐渐变差,最新的MODIS遥感产品可用于WRF模式下垫面数据集的更新。本文利用2013年MODIS卫星遥感产品反演的植被覆盖度更新WRF模式默认的气候态数据集,进一步利用WRF模式计算与GVF相关的地表参数,并与WRF模式默认数据进行对比。结果表明:利用MODIS更新后的GVF在2013年不同季节均较WRF模式默认的气候态GVF偏高,尤其是中国南方和东北地区表现的"更绿";MODIS制作GVF的原数据空间分辨率为1 km,较WRF模式默认数据集的分辨率更高,适合高分辨率中尺度模式下垫面数据集的更新,更新后的GVF数据更细致,尤其是夏季城市下垫面GVF较周围偏低非常明显;通过分析GVF变化引起的叶面积指数、地表反照率和地表发射率的变化可知,农作物植被类型叶面积指数的变化最显著。东北地区森林覆盖虽较广,但由于森林的叶面积指数年季变化相对较小,叶面积指数增量并没有农作物的明显。地表反照率和地表发射率也与WRF模式默认值有一定的差异。通过对比可见,更新后的GVF对WRF模式下垫面描述有一定的影响,需进一步开展季节尺度的模拟试验,对比植被覆盖度的变化对WRF模式模拟的影响。

The data sets of fraction of green vegetation coverage estimated by remote sensing and used in WRF （Weather Research and Forecasting Model）, was produced many years ago, and its representation is gradually deteriorating. The latest MODIS （ Moderate Resolution Imaging Spectroradiometer） products can be used to update the data sets in WRF. In this work, the green vegetation coverage fraction estimated using the MODIS products was compared with the default data sets in WRF along with the other relevant parameters. The results show that the green vegetation coverage fraction estimated by MODIS products is higher than that of default, especially in the south and northeast of China. Because of its high resolution of MODIS production, it is very suitable to estimate the underlying surface data sets in Numerical Weather Prediction （NWP）. Green vegetation coverage fraction in urban is significant lower than in non-urban area at summer. Analysis of increment of leaf area index, surface albedo and surface emission by update green vegetation coverage fraction suggests that leaf area index increment of crop is the largest among all vegetation. Although forest coverage is large in the northeast China, its annual and seasonal variation of leaf area index is relatively small, and its increment is not as lager as that of crop. Thus, surface albedo and emission have a certain difference compared with default by the WRF. In all, green vegetation coverage fraction has a propounding influence in NWP, and some simulation work such as seasonal simulation should be carded out to evaluate the influence results in NWP.