Flashcard library · Earth & Climate Science · PaperFren

Neural nets map CMIP6 weather into Greenland melt

ANNs trained on CESM2 translate CMIP6 atmospheres into Greenland surface-melt increases of about 79–264% by 2100 depending on SSP.

Keep what you just learned

Save these 11 cards to your NoteFren library. Your first review starts right away, and each card links back to the study it came from.

What sea-level potential of the GrIS is cited?

7.34 m if fully melted.

What is the ANN trained to emulate?

CESM2’s explicit Greenland surface-melt calculation from large-scale atmosphere fields.

Which five summer predictors were tried?

T2m, 500 hPa height, cloud cover, incoming radiation, and snowfall.

How was training data split?

90% train, 10% cross-validation of CESM2 years.

How many CMIP6 runs feed the projections?

345 historical plus 123–147 members across four SSPs from dozens of models.

Which ANNs work best versus MAR?

T2m and snowfall; Z500, clouds, and radiation did not perform well.

What 21st-century melt increases are projected?

79% for SSP1-2.6 to 264% for SSP5-8.5 versus 1979–1998.

What absolute SSP melt increases are given?

414±276 to 1,378±555 Gt yr−1 from SSP1-2.6 to SSP5-8.5.

How close is the SSP5-8.5 ANN mean to MAR?

1.6% bias, with MAR inside the ANN spread.

When do scenarios diverge?

Between 2030 and 2050.

Why are these melt increases higher than older RCP studies?

CMIP6 models have higher climate sensitivity than CMIP5.