A ROBUST PLS PROCEDURE

A robust partial least squares(PLS)regression algorithm is developed.This is achieved by substitutionof the univariate regression steps in the iterative PLS2 algorithm by a robust alternative.The anglebetween loading vectors from both perturbed and unperturbed solutions is used as a measure ofrobustness.By means of a perturbation study on a structure-activity data set,it is demonstrated thatthe stability of the robust method is superior to standard PLS.     

Modes of synoptic development within the Polar Basin

The patterns of advective development are classified for twelve synoptic sequences in the Polar Basin during the June through October interval of 1979. The intent is to develop models of the advective contribution to synoptic evolution.Fifteen categories are derived from 94 cases. Five are significant classes in that they represent at least 6 members each. While there is a predominance of patterns which are interpreted as weakening or occuluded, there are significant examples of intensification and frontogenesis within the Basin. One case is over the central Arctic Ocean. In other classes, there is development as a result of radiative or diabatic influences despite the frontolyic nature of the advective patterns.This is the first attempt to classify synoptic systems on the basis of a dynamic index, in contrast to pressure patterns. This approach provides new insight into the processes responsible for the evolution of the systems.List of symbols C_p specific heat of dry air - g acceleration due to gravity - H diabatic heating rate - p pressure - R gas constant - V_g geostrophic wind velocity - W vertical velocity - static stability - f _o Criolis parameter - geopotential height - approximate vertical velocity — omega - p density of air     

The significance of surface temperature patterns on the energy balance of a small lake in the Canadian Shield

Abstract

Temperature patterns of a small lake in the Canadian Shield are examined by means of thermal imagery. The effect of the different surface temperatures on the daytime energy balance is examined for two points over the lake at the time the thermal imagery was taken. For 14 June 1979, two distinct energy balance regimes are noted. At the lake centre, where the deeper water registers a relatively cool thermal signature, a boundary‐layer inversion is observed. The downward sensible heat flux augments the net radiation, and the latent heat flux is 105% of the radiant input. Along the lake margins, the shallows register warm thermal signatures and a lapse profile is observed. The sensible heat flux is an energy sink and the latent heat flux is diminished to 88% of the net radiation. This difference indicates that a single point estimate may introduce a bias if it is assumed to be representative of the lake average for the purposes of studying lake evaporation.

The calculation of the latent heat flux and evaporation is very sensitive to the value of the surface temperature. When the spatial patterns of surface temperature are considered in an estimate of the lake evaporation, the spatially integrated value differs by —6% from the estimate based upon a single point observation at the lake centre for a mid‐day in June and by +10% from the estimate based upon observations collected over the warm shallows.     

The role of local heat sources in synoptic activity within the polar basin

Abstract

A quasi‐geostrophic model of the atmosphere is used to determine the significance of the surface enthalpy flux in synoptic activity within the Polar Basin. Of primary interest is whether the enthalpy flux from open water in the seasonal sea‐ice zone is the predominant contributing mechanism or whether the advective fields of vorticity and thickness are controlling factors. This is of importance in discussions of the feedback processes between the atmosphere, cryosphere and ocean.

For a case selected in the Laptev Sea near the end of the fall period of ice growth, the surface enthalpy flux is as significant a contribution to synoptic activity as the vorticity advection is. The enthalpy flux is a relatively insignificant factor at this time in the Beaufort Sea, however, because of the smaller area of open water and the lower wind speeds associated with the weaker synoptic systems in this region. It is also relatively insignificant at both locations at the beginning of the fall freeze‐up interval and in June, during the melt period.     

Remote sensing of coral reefs and their physical environment

There has been a vast improvement in access to remotely sensed data in just a few recent years. This revolution of information is the result of heavy investment in new technology by governments and industry, rapid developments in computing power and storage, and easy dissemination of data over the internet. Today, remotely sensed data are available to virtually anyone with a desktop computer. Here, we review the status of one of the most popular areas of marine remote sensing research: coral reefs. Previous reviews have focused on the ability of remote sensing to map the structure and habitat composition of coral reefs, but have neglected to consider the physical environment in which reefs occur. We provide a holistic review of what can, might, and cannot be mapped using remote sensing at this time. We cover aspects of reef structure and health but also discuss the diversity of physical environmental data such as temperature, winds, solar radiation and water quality. There have been numerous recent advances in the remote sensing of reefs and we hope that this paper enhances awareness of the diverse data sources available, and helps practitioners identify realistic objectives for remote sensing in coral reef areas.     

Enhanced Arctic Ice Concentration Estimation Merging MODIS Ice Surface Temperature and SSM/I Sea-Ice Concentration

Ice concentration data from passive microwave sensors are not reliable during the summer melt season. In this study passive microwave ice concentration estimates are improved upon through the assimilation of ice surface temperature data from the Moderate-Resolution Imaging Spectroradiometer (MODIS). Ice concentration from the analysis is converted to ice extent and compared with data from the Interactive Multisensor Snow and Ice Mapping System (IMS). Ice concentration analysis from data assimilation is in better agreement with the IMS data than the original passive microwave data, with the largest improvements occurring during sea-ice melt.     

A TEST OF SIGNIFICANCE FOR PARTIAL LEAST SQUARES REGRESSION

Partial least squares (PLS) regression is a commonly used statistical technique for performingmultivariate calibration, especially in situations where there are more variables than samples. Choosingthe number of factors to include in a model is a decision that all users of PLS must make, but iscomplicated by the large number of empirical tests available. In most instances predictive ability is themost desired property of a PLS model and so interest has centred on making this choice based on aninternal validation process. A popular approach is the calculation of a cross-validated r~2 to gauge howmuch variance in the dependent variable can be explained from leave-one-out predictions. Using MonteCarlo simulations for different sizes of data set, the influence of chance effects on the cross-validationprocess is investigated. The results are presented as tables of critical values which are compared againstthe values of cross-validated r~2 obtained from the user's own data set. This gives a formal test forpredictive ability of a PLS model with a given number of dimensions.