This paper evaluates in a realistic context the local contributions of direct atmospheric forcing and intrinsic oceanic processes on interannual sea level anomalies (SLAs). A ¼° global ocean–sea ice general circulation model, driven over 47 yr by the full range of atmospheric time scales, is quantitatively assessed against altimetry and shown to reproduce most observed features of the interannual SLA variability from 1993 to 2004. Comparing this simulation with a second driven only by the climatological annual cycle reveals that the intrinsic part of the total interannual SLA variance exceeds 40% over half of the open-ocean area and exceeds 80% over one-fifth of it. This intrinsic contribution is particularly strong in eddy-active regions (more than 70%–80% in the Southern Ocean and western boundary current extensions) as predicted by idealized studies, as well as within the 20°–35° latitude bands. The atmosphere directly forces most of the interannual SLA variance at low latitudes and in most midlatitude eastern basins, in particular north of about 40°N in the Pacific. The interannual SLA variance is almost entirely due to intrinsic processes south of the Antarctic Circumpolar Current in the Indian Ocean sector, while half of this variance is forced by the atmosphere north of it. The same simulations were performed and analyzed at 2° resolution as well: switching to this laminar regime yields a comparable forced variability (large-scale distribution and magnitude) but almost suppresses the intrinsic variability. This likely explains why laminar ocean models largely underestimate the interannual SLA variance.

No abstract

No abstract

Using the SIBFA polarizable molecular mechanics procedure, we analyze the binding energy of a bimetallic Mg(II)/Zn(II) enzyme, isopentenyl diphosphate isomerase, to an inhibitor built up of a trianionic diphosphate and of a cationic ethyldimethylammonium (EDMA) moiety. The analyses are performed on the protein recognition site, which totals 13 residues, as well as on some "mutants" in which one selected residue is removed at a time. They are also carried out for the individual recognition sites, namely, EDMA, Mg(II), and Zn(II). Comparisons are done with ab initio quantum chemistry (QC) results on all considered sites, with different basis sets and at different levels of correlation. The SIBFA computations reproduce the evolutions of the QC interaction energies in the recognition site and its "mutants". For such sites, small (<2-3%) relative errors are found after the BSSE correction is done. Such close agreements can conceal, however, some shortcomings found in the individual binding sites, which QC energy decomposition analyses can identify.