Monitoring drought occurrence through a count-based statistical modeling

Rizwan Farooq

doi:10.1016/j.pce.2026.104426

Monitoring drought occurrence through a count-based statistical modeling

2026-04-08

0

PRE

AI

R

Rizwan Farooq

M

Maysaa Elmahi Abd Elwahab

Mohammed M.A. Almazah

I

Ijaz Hussain

DOI：10.1016/j.pce.2026.104426

Original

Original request for help

Save

Abstract

En 中文

• Dual-threshold SPI framework to model mild and severe drought occurrence counts. • 1981–2021 monthly data from 24 stations in Punjab, Pakistan. • Poisson GEE applied for population-averaged modeling of drought counts. • Significant spatial clustering confirmed using Moran’s I and Geary’s C. • Lasso-based selection improved prediction and reduced multicollinearity. • Severity-dependent drivers differ between mild and extreme drought conditions. • Temporal dependence modeled using GEE with AR(1) and independence structures.

Keywords:

drought occurrence

count-based modeling

Poisson GEE

spatial clustering

Lasso selection

Monitoring drought occurrence through a count-based statistical modeling

Abstract

Journal

Organization