loading page

Improving seasonal drought predictions by conditioning on ENSO states
  • Patrick Pieper,
  • André Düsterhus,
  • Johanna Baehr
Patrick Pieper
University Hamburg

Corresponding Author:[email protected]

Author Profile
André Düsterhus
Author Profile
Johanna Baehr
Universität Hamburg, Center for Earth System Research and Sustainability
Author Profile


Significant hindcast skill for the 3-month standardized precipitation index (SPI$_{3M}$) has been so far limited to one lead month. To increase that lead time, we propose to exploit well-known El Ni\~no-Southern Oscillation (ENSO)–precipitation teleconnections through ENSO-state conditioning. We condition initialized seasonal SPI$_{3M}$ hindcasts, derived from the Max-Planck-Institute Earth System Model over the period 1982-2013, on ENSO states by exploring significant agreements between two complementary analyses: hindcast skill ENSO–composites, and observed ENSO–precipitation correlations. Predictions conditioned on autumn (ASO)-ENSO states demonstrate significant and reliable winter (DJF) drought hindcast skill up to lead month 4 in equatorial South- and southern North America. The area of reliable drought hindcast skill is further enlarged when the respective region’s dry ENSO phase is already present in the antecedent summer (JJA-ENSO-state-conditioned). In contrast to previous studies, our evaluation separates predictions and observations. Thereby, ENSO-state conditioning demonstrates genuine hindcast skill up to lead month 4.