Threshold level method is well-known for drought identification with the advantage of its simplicity. However, there is no criterion for classification of drought classes in this method. Therefore, a K index based on threshold level method was proposed and verified to perform well in drought assessment in the Luanhe River basin, China. Meteorological data and remote sensing data were used to calculate the net primary productivity (NPP) of the basin by the CASA model. Results showed that the model had a good performance in comparison with the downloaded yearly remote sensing NPP and monthly PsnNet. The NPP in the basin tended to increase slowly during the year of 2000–2010, and the NPP in the downstream was generally larger than upstream. In three selected representative drought years, drought reduced NPP by about 15–25%, and during the growth season, drought may reduce NPP by about 4.6%, 2.7%, and 1.9% in July, August, and September, respectively, due to reduced precipitation. NPP on grasslands and agricultural land were more susceptible to drought than forests. The result of gray incidence analysis showed that the effects of drought on NPP had a certain delay with about 5 months.

阈值水平方法以其简单性而闻名于干旱识别。但是，这种方法没有分类干旱类别的标准。因此，提出并验证了基于阈值水平法的K指数在to河流域的干旱评估中表现良好。利用气象数据和遥感数据，通过CASA模型计算了盆地的净初级生产力（NPP）。结果表明，与下载的年度遥感NPP和每月的PsnNet相比，该模型具有良好的性能。流域的NPP在2000–2010年期间趋于缓慢增长，下游的NPP通常大于上游的NPP。在选定的三个代表性干旱年中，干旱使NPP降低了约15–25％，在生长季节，由于降水减少，干旱可能会使7月，8月和9月的NPP分别降低约4.6％，2.7％和1.9％。草原和农业用地上的核电厂比森林更容易受到干旱的影响。灰色关联度分析结果表明，干旱对NPP的影响有一定的延迟，大约5个月。