W. Van Snyder

The Earth observing system microwave limb sounder (EOS MLS) on the aura Satellite

The Earth Observing System Microwave Limb Sounder measures several atmospheric chemical species (OH, HO/sub 2/, H/sub 2/O, O/sub 3/, HCl, ClO, HOCl, BrO, HNO/sub 3/, N/sub 2/O, CO, HCN, CH/sub 3/CN, volcanic SO/sub 2/), cloud ice, temperature, and geopotential height to improve our understanding of stratospheric ozone chemistry, the interaction of composition and climate, and pollution in the upper troposphere. All measurements are made simultaneously and continuously, during both day and night. The instrument uses heterodyne radiometers that observe thermal emission from the atmospheric limb in broad spectral regions centered near 118, 190, 240, and 640 GHz, and 2.5 THz. It was launched July 15, 2004 on the National Aeronautics and Space Administrations Aura satellite and started full-up science operations on August 13, 2004. An atmospheric limb scan and radiometric calibration for all bands are performed routinely every 25 s. Vertical profiles are retrieved every 165 km along the suborbital track, covering 82/spl deg/S to 82/spl deg/N latitudes on each orbit. Instrument performance to date has been excellent; data have been made publicly available; and initial science results have been obtained.

Encyclopedia of software components

Intelligent browsing through a collection of reusable software components is facilitated with a computer having a video monitor and a user input interface such as a keyboard or a mouse for transmitting user selections, by presenting a picture of encyclopedia volumes with respective visible labels referring to types of software, in accordance with a metaphor in which each volume includes a page having a list of general topics under the software type of the volume and pages having lists of software components for each one of the generic topics, altering the picture to open one of the volumes in response to an initial user selection specifying the one volume to display on the monitor a picture of the page thereof having the list of general topics and altering the picture to display the page thereof having a list of software components under one of the general topics in response to a next user selection specifying the one general topic, and then presenting a picture of a set of different informative plates depicting different types of information about one of the software components in response to a further user selection specifying the one component.

EOS MLS forward model polarized radiative transfer for Zeeman-split oxygen lines

This work supplements the Earth Observing System (EOS) Microwave Limb Sounder (MLS) clear-sky unpolarized forward model with algorithms for modeling polarized emission from the Zeeman-split 118.75-GHz O/sub 2/ spectral line. The model accounts for polarization-dependent emission and for correlation between polarizations with complex, 2/spl times/2 intensity and absorption matrices. The oxygen line is split into three Zeeman components by the interaction of oxygens electronic spin with an external magnetic field, and the splitting is of order /spl plusmn/0.5 MHz in a typical geomagnetic field. Zeeman splitting is only significant at pressures low enough that collisional broadening (/spl sim/1.6 MHz/hPa) is not very large by comparison. The polarized forward model becomes significant for MLS temperature retrievals at pressure below 1.0 hPa and is crucial at pressures below /spl sim/0.03 hPa. Interaction of the O/sub 2/ molecule with the radiation field depends upon the relative orientation of the radiation polarization mode and the geomagnetic field direction. The model provides both limb radiances and the derivatives of these radiances with respect to atmospheric temperature and composition, as required by MLS temperature retrievals. EOS MLS views the atmospheric limb at 118.75 GHz with a pair of linear-cross-polarized, 100-kHz-resolution, 10-MHz-wide spectrometers. The antennas of the associated receivers are scanned to view rays with tangent heights from the Earths surface to 0.001 hPa. Comparisons of the modeled MLS radiances with measurements show generally good agreement in line positions and strengths, however residuals in the line centers at the highest tangent heights are larger than desired and still under investigation.

The exact dot product as basic tool for long interval arithmetic

Computing with guarantees is based on two arithmetical features. One is fixed (double) precision interval arithmetic. The other one is dynamic precision interval arithmetic, here also called long interval arithmetic. The basic tool to achieve high speed dynamic precision arithmetic for real and interval data is an exact multiply and accumulate operation and with it an exact dot product. Pipelining allows to compute it at the same high speed as vector operations on conventional vector processors. Long interval arithmetic fully benefits from such high speed. Exactitude brings very high accuracy, and thereby stability into computation. This document, which has been incorporated into the draft standard for interval arithmetic being developed by IEEE P1788, specifies the implementation of an exact multiply and accumulate operation.

Algorithm 699: a new representation of Patterson's quadrature formulae

Patterson [1, 2, 3, 41 describes a generalization of the Kronrod [51 scheme to derive a family of quadri~ture formulae 11, Iz, etc., in whichl 11 is the midpoint rule, and Iz is the three point Gauss rule. 2 k new points are added to the points used for Ik to produce lk + ~, with the new abscissae selected so as to maximize the degree of algebraic precision. Based on our experience [61, we believe these formulae to be the most effective available for use in an automatic quadrature routine. We present here a method to reduce the number of coefficients necessary to represent Patterson’s quadrature formulae, to reduce the amount of storage necessary for storing function values and to achieve a slightly smaller error when the integrand has singular behavior at the end of the interval. Let ~~,1 = O, and let &~,k be the ith new positive abscissa needed to form the kth formula from the k – lst. Define Yo,l = f(o) and ‘Yi,k = f(gi,k) + fl – fi,k) for k > ~. The formulae can be given in the form

Testing functions of one and two arguments

Software to evaluate functions of one or two arguments is too often tested or certified, or the accuracy assessed, by ad hoc methods. Systematic methods to test, certify or assess the accuracy of single- or double-precision functions of one or two arguments, and supporting software, are described here.

Multilevel EXIT and CYCLE aren't so bad

The literature is replete with arguments pro and con EXIT and CYCLE control structures. We agree with the arguments pro, and hold here that the most common argument con is specious. The most important argument pro multilevel EXIT and CYCLE is that they promote efficient use of computing resources. The most frequently cited argument con multilevel EXIT and CYCLE is that they obscure intent. We argue that absence of multilevel EXIT and CYCLE obscures intent, given adequate programming support tools. By way of example, consider an object x and a set S, and the two abstract statements xES and x~S. When posed in this abstract form, they seem innocuously similar. But when cast into a programming language that does not allow EXIT, or allows EXIT only from LOOP structures, one requires the use of boolean variables or obscuration of the major loop control strategy to discover x~S. Consider the case in which S is represented by a mapping of the N elements of S onto consecutive elements of an array A. By allowing EXIT to apply to BLOCK as well as an arbitrary LOOP, we can respond to x~S simply:

Algorithm 982: Explicit Solutions of Triangular Systems of First-Order Linear Initial-Value Ordinary Differential Equations with Constant Coefficients

A method to compute explicit solutions of homogeneous triangular systems of first-order linear initial-value ordinary differential equations with constant coefficients is described. It is suitable for the limited case of well separated eigenvalues, or for multiple zero eigenvalues provided the entire column corresponding to a zero eigenvalue is zero. The solution for the case of constant inhomogeneity is described. The method requires only the computation of a constant matrix using a simple recurrence. Computing the solutions of the system from that matrix, for values of the independent variable, requires one to exponentiate only the diagonal of a matrix. It is not necessary to compute the exponential of a general triangular matrix. Although this work was motivated by a study of nuclear decay without fission or neutron absorption, which is used throughout as an example, it has wider applicability.

Scientific programming in Fortran

The Fortran programming language was designed by John Backus and his colleagues at IBM to reduce the cost of programming scientific applications. IBM delivered the first compiler for its model 704 in 1957. IBMs competitors soon offered incompatible versions. ANSI (ASA at the time) developed a standard, largely based on IBMs Fortran IV in 1966. Revisions of the standard were produced in 1977, 1990, 1995 and 2003. Development of a revision, scheduled for 2008, is under way. Unlike most other programming languages, Fortran is periodically revised to keep pace with developments in language and processor design, while revisions largely preserve compatibility with previous versions. Throughout, the focus on scientific programming, and especially on efficient generated programs, has been maintained.

Remark on algorithm 723: Fresnel integrals

Algorithm 723: Fresnel Integrals has been improved to provide more precise results for x >> 0.