AbstractThe length of record (LOR) method is an evaluation method that provides quantitative advice for the amount of computational data required for use of the indicators of hydrological alteration (IHA). The use of multi-index hydrological indicators to reflect river hydrological–ecological characteristics is the essence of the IHA method, while the LOR evaluation result using a single index does not have practical application value in the absence of IHA data volume. In this paper, we expand the LOR method from single index version into multi-index version, apply it to comprehensively analyze the credibility of hydrological alteration (HA) multi-indicators under different data volumes, and explore the relationship between multi-index LOR results and data requirements. Combined with the hydrological–ecological relationship, the practical application criteria of LOR dimension reduction under the condition of multiple HA indicators is given. The results show that the LOR results corresponding to each group of indicators in IHA have different data requirements, so an in-depth understanding of the hydrological–ecological relationship is the key to LOR's application of IHA data dimension reduction. In addition, we discuss the limitations of the LOR method of multi-index dimension reduction and its application value in IHA calculations.
Cumulative Median Estimation for Sufficient Dimension Reduction