Comprehensive simulation of the middle atmospheric climate: some recent results

This study discusses the results of comprehensive time-dependent, three-dimensional numerical modelling of the circulation in the middle atmosphere obtained with the GFDL SKYHI troposphere-stratosphere-mesosphere general circulation model (GCM). The climate in a long control simulation with an intermediate resolution version (3° in horizontal) is briefly reviewed. While many aspects of the simulation are quite realistic, the focus in this study is on remaining first-order problems with the modelled middle atmospheric general circulation, notably the very cold high latitude temperatures in the Southern Hemisphere (SH) winter/spring, and the virtual absence of a quasi-biennial oscillation (QBO) in the tropical stratosphere. These problems are shared by other extant GCMs. It was noted that the SH cold pole problem is somewhat ameliorated with increasing horizontal resolution in the model. This suggests that improved resolution increases the vertical momentum fluxes from the explicitly resolved gravity waves in the model, a point confirmed by detailed analysis of the spectrum of vertical eddy momentum flux in the winter SH extratropics. This result inspired a series of experiments with the 3° SKYHI model modified by adding a prescribed zonally-symmetric zonal drag on the SH winter westerlies. The form of the imposed momentum source was based on the simple assumption that the mean flow drag produced by unresolved waves has a spatial distribution similar to that of the Eliassen-Palm flux divergence associated with explicitly resolved gravity waves. It was found that an appropriately-chosen drag confined to the top six model levels (above 0.35 mb) can lead to quite realistic simulations of the SH winter flow (including even the stationary wave fields) through August, but that problems still remain in the late-winter/springtime simulation. While the imposed momentum source was largely confined to the extratropics, it produced considerable improvement in the simulation of the equatorial semiannual oscillation, with both the easterly and westerly phases being somewhat more intense than in the control simulation. A separate experiment was conducted in which the SKYHI model was simplified so that it had no topography and so that the seasonal cycle was frozen in perpetual equinox conditions. These changes result in a model that has much reduced interhemispheric asymmetry. This model spontaneously produces a long period mean flow oscillation of considerable amplitude in the tropical upper stratopause. The implication of this result for the general issue of obtaining a QBO in comprehensive GCMs is discussed.     

Extreme peak value vessel response combinations with wide band spectra

An extension is proposed to the standard peak value distribution for time histories possessing wide banded spectra, which allows the distribution of an associated variable to be calculated at the instant that the peak value occurs, the theory is applicable to a wide range of practical circumstances, where it is desired to specify the value of associated response variables at the instant that the dominant response is a maximum. An example is given of the specification of coincident heave and roll of a barge for input to a structural load case analysis.     

Relative sea‐level observations in western Scotland since the Last Glacial Maximum for testing models of glacial isostatic land movements and ice‐sheet reconstructions

Observations of relative sea‐level change and local deglaciation in western Scotland provide critical constraints for modelling glacio‐isostatic rebound in northern Britain over the last 18 000 years. The longest records come from Skye, Arisaig and Knapdale with a shorter, Holocene, record from Kintail. Biostratigraphic (diatom, pollen, dinoflagellate, foraminifera and thecamoebian), lithological and radiocarbon analyses provide age and elevation parameters for each sea‐level index point. All four sites reveal relative sea‐level change that is highly non‐monotonic in time as the local vertical component of glacio‐isostatic rebound and eustasy (or global meltwater influx) dominate at different periods. Copyright © 2006 John Wiley & Sons, Ltd.     

Seasonal inundation patterns in two large savanna floodplains of South America: the Llanos de Moxos (Bolivia) and the Llanos del Orinoco (Venezuela and Colombia)

Inundation patterns in two of the largest savanna floodplains of South America were studied by analysis of the 37‐GHz polarization difference observed by the Scanning Multichannel Microwave Radiometer (Nimbus‐7 satellite). Flooded area was estimated at monthly intervals for January 1979 through to August 1987 using mixing models that account for the major landscape units with distinctive microwave emission characteristics. Results are presented separately for five subregions in each of the two floodplain regions to show the spatial as well as temporal variability in inundation patterns. The total area inundated during the 9 years varied between 2069 and 78 460 km² in the Llanos de Moxos (also spelled as Mojos; median area, 23 383 km²) and 1278 and 105 454 km² in the Llanos del Orinoco (median, 25 374 km²), not including the open‐water area of permanent lakes and river channels. The correlation between flooded area and river stage was used to extend the inundation records over a 30‐year period in the Moxos (1967–97) and a 58‐year period (1927–85) in the Orinoco. Interannual variability in inundation is greater in the Moxos than the Orinoco. Comparison of these data, however, with a previously published analysis of the Pantanal wetland shows that inundation patterns in these two floodplain regions are not as variable across years as they are in the Pantanal. Copyright © 2004 John Wiley & Sons, Ltd.     

Relative sea‐level changes,glacial isostatic modelling and ice‐sheet reconstructions from the British Isles since the Last Glacial Maximum

While contributing <1 m equivalent eustatic sea‐level rise the British Isles ice sheet produced glacio‐isostatic rebound in northern Britain of similar magnitude to eustatic sea‐level change, or global meltwater influx, over the last 18 000 years. The resulting spatially variable relative sea‐level changes combine with observations from far‐field locations to produce a rigorous test for quantitative models of glacial isostatic adjustment, local ice‐sheet history and global meltwater influx. After a review of the attributes of relative sea‐level observations significant for constraining large‐scale models of the isostatic adjustment process we summarise long records of relative sea‐level change from the British Isles and far‐field locations. We give an overview of different global theoretical models of the isostatic adjustment process before presenting intercomparisons of observed and predicted relative sea levels at sites in the British Isles and far‐field for a range of Earth and ice model parameters in order to demonstrate model sensitivity and the resolving power available from using evidence from the British Isles. For the first time we show a good degree of fit between relative sea‐level observations and predictions that are based upon global Earth and ice model parameters, independently derived from analysis of far‐field data, with a terrain‐corrected model of the British Isles ice sheet that includes extensive glaciation of the North Sea and western continental shelf, that does not assume isostatic equilibrium at the Last Glacial Maximum and keeps to trimline constraints of ice surface elevation. We do not attempt to identify a unique solution for the model lithosphere thickness parameter or the local‐scale detail of the ice model in order to provide a fit for all sites, but argue that the next stage should be to incorporate an ice‐sheet model that is based on quantitative, glaciological model simulations. We hope that this paper will stimulate this debate and help to integrate research in glacial geomorphology, glaciology, sea‐level change, Earth rheology and quantitative modelling. Copyright © 2006 John Wiley & Sons, Ltd.     

The dark side of Iapetus

This paper presents new photometric and spectrophotometric observations of the dark (leading) hemisphere of Saturn's satellite Iapetus. Spectrophotometry from 0.3–1.0 um (May 1979) shows the dark hemisphere to be very red, similar to a few asteroids and the Moon, but with no spectral features attributable to olivine or pyroxene. Near-infrared spectrophometry in the regions 1.4–2.5 um (May 1981) and 3.0–3.8 um (February 1981) reveals water ice absorption bands, probably resulting from the polar caps intruding onto the dark hemisphere. The reflectance of Iapetus is unlike that of carbonaceous chondrites or C-type asteroids and most closely resembles the reflectance (and low albedo) of carbonaceous (organic) residue from the Murchison C2 carbonaceous chondrite. The Murchison material has the same red slope and a probable spectral feature near 0.6 um seen in Iapetus data. Three hypotheses for the formation of the dark hemisphere are discussed in light of the observational data. The favored hypothesis is that debris from Phoebe or other unknown outer satellites of Saturn impacts the dark hemisphere of Iapetus as Poynting-Robertson drag causes the debris to spiral toward Saturn. The high-velocity impacts preferentially remove ice from the satellite's surface, causing enrichment of included carbonaceous material intrinsic to Iapetus. The reflectance of Phoebe itself is significantly different from that of Iapetus, suggesting that relatively little Phoebe debris lies on the dark hemisphere. There remains the possibility that the impacting debris originates from another body of composition similar to the Murchison residue and that this material is exposed on the surface of Iapetus.     

Some tests of the Vaníček Method of spectral analysis

One of the chief claims of the Vaníček Method of spectrum analysis is its, ability to remove ‘systematic noise’ from a time series with minimal distortion of the spectrum of the remaining series. In this paper, comparisons are made between the Vaníček Method and the more usual method consisting of preliminary removal of the systematic noise followed by fourier analysis. Formulas relating the two methods are developed and a series of comparative plots of simple spectra are presented. Results indicate that the Vaníček Method does not displace a spectral peak falling close to the systematic noise, that it distorts the amplitude of the peak less than the conventional method of analysis, and that it tends to intensify the side lobes.