Why do drought indices overestimate the drought-related impacts of
global warming - in models and in reality?
Abstract
In global climate models, CO2-driven warming causes strong and very
widespread mean drying trends in climatic wetness indices like the
Palmer Drought Severity Index, Aridity Index and Standardized
Precipitation-Evapotranspiration Index. Yet, these same simulations also
predict that runoff will not decline over most of Earth’s surface, that
root-zone soil moisture and evaporative fraction will decline only
regionally, and that vegetation will become broadly greener. Thus,
actual drought impacts of warming in these models are far less broad and
severe than implied by the drought indices. Here, I probe why this
“index-impact gap” occurs, and whether it is a feature of reality as
well as models. In particular, I show that the discrepancies are not
just limited to simulations which assume a substantial direct CO2 effect
on vegetation, but are also large in greenhouse-only simulations,
implying that they occur for fundamental climatic reasons rather than
via CO2-physiological pathways. I also review key observational evidence
that the index-impact gap has also been large over the historical era
and was very evident for the last glacial-to-interglacial warming,
lending much additional credence to the model output.