Identifying the links among poverty, hydroenergy and water use using data mining methods

Fuyou Tian, Bingfang Wu*, Hongwei Zeng, Shukri Ahmed, Nana Yan, Ian White, Miao Zhang, A. Stein

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Water is fundamental to human well-being, social development and the environment. Water development, particularly hydropower, provides an important source of renewable energy. Water development is strongly affected by poverty, but only few attempts have been made to understand the links between water development and poverty from a global water development point of view. In this work, this linkage was explored using reservoir construction, hydroenergy and water use data along with six derived indicators. We used association rule mining and classification and regression trees (CART) to identify the links. Random forests were employed to search for factors sensitive to poverty. This study shows that the reservoir density is significantly related to poverty, and reservoir densities are lower in countries with higher poverty rates. Countries with a higher use of small hydropower (SHP) systems are generally more prosperous as follows: an SHP utilization rate above 27% corresponds to a poverty rate below 4.9%. The ratio of water utilization, water availability per capita (WAPC) and reservoir density were essential for the prediction of the poverty class. All three ratios could be related to poverty alleviation as they enable the identification of the potential for water resource development and their constraints. This study concludes that water development in poor countries needs to receive more attention.
Original languageEnglish
Pages (from-to)1725-1741
Number of pages17
JournalWater resources management
Volume34
Issue number5
DOIs
Publication statusPublished - 25 Mar 2020

Keywords

  • ITC-ISI-JOURNAL-ARTICLE

Fingerprint Dive into the research topics of 'Identifying the links among poverty, hydroenergy and water use using data mining methods'. Together they form a unique fingerprint.

Cite this