七套土地覆被数据在羌塘高原的精度评价

Accuracy evaluation of the seven land cover data in Qiangtang Plateau

基于羌塘高原8个一级土地覆被类型(包括10个二级土地覆被类型)的6851个样本点,采用混淆矩阵方法,从总体精度、制图精度和用户精度角度评价International Geosphere-Biosphere Program's Data and Information System Cover(IGBPDIS)、Global Land cover mapping at30 m resolution(Globe Land 30)、The MODIS Land Cover Type product(MCD12Q1)、Climate Change Initiative Land Cover(CCI-LC)和Global Land Cover 2000(GLC2000)等七套土地覆被数据产品在羌塘高原的精度。结果表明:(1)七套数据产品的一级类型和二级类型总体精度普遍偏低,在相对较高的Globe Land 30和CCI-LC数据中,一级类型总体精度分别为55.09%和53.92%,二级类型分别为46.55%和46.23%;(2)草地、裸地和荒漠三个主要一级类型生产者精度最高的数据对应为:GLC 2000(46.19%)、MCD12Q1(39.20%)和IGBPDIS(84.44%)。而三个主要一级类型的用户精度均低于50%。其他覆被类型中,雪被与冰川类型用户精度最高的数据为CCI-LC(92.80%),漏分比例为19.90%;(3)羌塘高原特殊的高原环境与土地覆被分类系统构成原则和标准是影响遥感解译数据精度的主要原因。

The land cover datasets in the Qiangtang Plateau（QP） have generally been considered as the fundamental data in the studies of local environmental and ecological issues.We evaluated the accuracy of seven land cover datasets in the QP, i.e. International GeosphereBiosphere Program＇s Data and Information System Cover（IGBPDIS）, Global Land cover mapping at 30 m resolution（Globe Land 30）, The MODIS Land Cover Type product（MCD12 Q1）, Climate Change Initiative Land Cover（CCI-LC）, Global Land Cover 2000（GLC2000）, Glob Cover 2009（Glob Cover） and University of Maryland（UMD）. For that, the study used 6851 field samples with first and second level（8 and 10, respectively） land cover types. Three widely used parameters were derived to describe the error matrix of the map and also for the overall user＇s and producer＇s accuracy. The quantitative assessments of the map quality and classification accuracy for the available land cover maps will help to improve the overall accuracy of land cover mapping in future. The overall results of the assessment pointed out that the Globe Land 30 and CCI-LC land cover map have higher accuracy than other data sets. However, they are also just 55.09% and 53.92% in first level assessment and 46.55% and46.23% in the second level accuracy assessment. The best producer＇s accuracies of the three main land cover classes, e.g. alpine grassland, barren land and desert land in the QP were46.19% in GLC 2000, 39.20% in MCD12 Q1, and 84.44% in the IGBPDIS. The user＇s accuracy of the three first level land cover classes were less than 50%. In addition, the accuracy of the CCI-LC data was 92.8%, with omission error at 19.90% in the snow and ice cover. After analysis, we found that the discrepancy of classification system and the typical plateau environment in the QP are the main factors that result in a high level of inaccuracy of the land cover datasets.