Présentations Colloques

Session 5.01: Groundwater resources in a world facing climate change
Racha El Kadiri
Long-Term Precipitation Patterns Prediction- A Data Mining and Remote Sensing Based Water Management Approach
We developed and applied an integrated approach to construct predictive tools with lead times of 1 to 12 months to forecast precipitation amounts over the Mediterranean Basin region. The following steps were conducted- (1) acquire, assess and intercorrelate temporal remote sensing-based precipitation products (e.g. The CPC Merged Analysis of Precipitation [CMAP], Integrated Multi-Satellite Retrievals for GPM [IMERG]) throughout the investigation period (1979 to 2016), (2) acquire and assess monthly values for all of the climatic indices influencing the regional and global climatic patterns (e.g., Northern Atlantic Oscillation [NOI], Southern Oscillation Index [SOI], and Tropical North Atlantic Index [TNA])+ and (3) apply data mining methods (e.g. neural networks, principal component analyses) to extract relationships between the observed precipitation and the controlling factors (i.e. climatic indices with multiple lead-time periods) and use predictive tools to forecast monthly precipitation.**Preliminary results indicate that by using the period from January 1998 until August 2012 for model training and the period from September 2012 to January 2016 for testing, precipitation can be successfully predicted with a three-months lead over Northern Morocco with high accuracy (i.e. pearson correlation coefficient- 0.911). Future work will focus on applying this technique for prediction of precipitation over each of the climatically contiguous areas of the Mediterranean region. If our efforts are successful, our findings will lead the way to the development and implementation of long-term water management scenarios for the Mediterranean region.
United States