Score-driven threshold ice-age models: Benchmark models for long-run climate forecasts

Climate variables are known to be subject abrupt changes when some threshold levels surpassed. We use data for the last 798,000 years on global ice volume (Ice), atmospheric carbon dioxide level (CO2), and Antarctic land surface temperature (Temp) model measure those long-run nonlinear climate effects. The have very long asymmetric cycles, created by periods of upward trends, followed downward trends driven exogenous orbital variables. considered Milankovitch cycles eccentricity Earth’s orbit, obliquity, precession equinox. show that our new score-driven ice-age models improve statistical inference forecasting performance competing from literature. drawback using 1000-year frequency observations, is we cannot effects humanity during 250 years, which generated structural in climate, due unprecedented high CO2 Temp, low Ice volume. On other hand, advantage low-frequency they allow us obtain forecasts what would occurred if had not burned fossil fuels since start Industrial Revolution. These can serve as benchmarks evaluation impact Without predict existence turning points evolution three next 5,000 years: an trend volume, temperature.