Premium
South American palaeoclimate model simulations: how reliable are the models?
Author(s) -
Valdes Paul J.
Publication year - 2000
Publication title -
journal of quaternary science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.142
H-Index - 94
eISSN - 1099-1417
pISSN - 0267-8179
DOI - 10.1002/1099-1417(200005)15:4<357::aid-jqs547>3.0.co;2-8
Subject(s) - last glacial maximum , climatology , climate model , holocene , general circulation model , geology , latitude , storm , climate change , paleoclimatology , environmental science , oceanography , geodesy
Modelling climate change over South America represents a special challenge for computer climate models. The coarse resolution of most general circulation models (GCM) means that the high but narrow relief of the Andes is poorly represented. In addition, changes in the tropical lowlands will be strongly dependent on the ability of the models to accurately simulate convection, cloud cover and land surface processes. Both marine and terrestrial data also will be relatively sparse compared with northern latitudes. This paper examines the modelling uncertainties by discussing results from the Palaeoclimate Model Intercomparison Project (PMIP), which simulated the mid‐Holocene (6 kyr BP) and the Last Glacial Maximum (LGM, 21 kyr BP). The PMIP mid‐Holocene experiments show a reasonable amount of agreement, especially for surface air temperature changes. The PMIP LGM simulations show a greater degree of scatter, which is partly associated with uncertainties in the treatment of the ocean surface. This suggests that uncertainties in the boundary conditions can be as important or more important than uncertainties due to the internal physics within the GCMs. In addition, we consider the results from a new set of simulations for 15 kyr BP, 12 kyr BP and 9 kyr BP, using the climate model of the UK Universities Global Atmospheric Modelling Programme (UGAMP). We compare these results with previous simulations by COHMAP. The higher resolution of the UGAMP model allows us to examine mid‐latitude storm‐track changes. The changes in the storm track are largest upstream of South America and reach their largest at around 15 kyr. The model also predicts a positive snow mass balance in the Southern Andes up until 12 kyr BP. The associated changes in mid‐latitude temperatures and precipitation over South America, however, are generally small and are broadly in agreement with previous studies. All of these results suggest that the basic elements of the GCM predictions of climate change are relatively robust, and the most important uncertainties probably arise from lack of detailed knowledge of the required boundary conditions. The next generation of climate models will alleviate many of these problems. Copyright © 2000 John Wiley & Sons, Ltd.