Logo Lanfrica

derimy/drought-modeling-probabilistic-forecasting-chelif

Domaine:

climate
Créateur:
der
Hôte:
Bayesian drought modelling and probabilistic forecasting framework for the Chélif basin (Algeria), combining SPI, scPDSI, CMIP6 projections, and uncertainty-aware time-series analysis. # Drought Modelling and Probabilistic Forecasting Framework — Chélif Basin Bayesian drought modelling and probabilistic forecasting framework for the Chélif basin (Algeria), combining SPI, scPDSI, CMIP6 projections, and uncertainty-aware time-series analysis. --- **Author**: Imene DERRAR **Supervisor**: Dr. Abdelillah OTMANE CHERIF **Institutions**: NHSM — National Higher School of Mathematics · IHFR — Institut Hydrométéorologique de Formation et de Recherche **Academic year**: 2025–2026 **Project**: Final-year engineering thesis in Applied Mathematics — Specialization: Statistics, Econometrics and Actuarial Science --- ## 1. Project overview This repository accompanies a final-year engineering thesis on drought modelling and probabilistic forecasting in the Chélif basin, a semi-arid catchment in north-west Algeria. The work builds two complementary drought indices on the same precipitation observations and propagates structural, parameter, and scenario uncertainty into probabilistic statements about future drought conditions. Two operational deliverables are documented here: 1. A **Bayesian probabilistic forecast** of the Standardised Precipitation Index (SPI) at the six-month accumulation scale, calibrated on station data from 1990–2014 and projected forward under SSP2-4.5 and SSP5-8.5. 2. A **complementary physically-interpretable drought signal** based on the self-calibrating Palmer Drought Severity Index (scPDSI), driven by TerraClimate and ERA5-Land forcings. ## 2. Scientific motivation The Chélif basin is one of the most water-stressed catchments in North Africa. Reservoir management, agricultural planning, and water supply all depend on credible drought information at horizons of months to decades. The thesis asks two related questions: - Can a probabilistic monthly forecast of SPI-6 produce calibrated drought-risk statements at short horizons given the limited observational record? - How do CMIP6 projections of precipitation and temperature, afte …