Comparing the H-alpha Intensity
and Radio Wave Scattering
on Eight Low-Latitude Lines of Sight

(A Poster presented at the 187th meeting of the AAS, Jan 1996)

John H. Simonetti, Brian Dennison, and Gregory A. Topasna
Martin Observatory, Institute for Particle Physics and Astrophysics,
and Department of Physics,
Virginia Tech


Spangler and Reynolds (1990 ApJ 361, 116) compared radio wave scattering sizes of eight extragalactic sources seen along low-latitude lines of sight through the Galaxy with H-alpha intensities measured along those lines of sight using the Wisconsin Fabry-Perot interferometer. They found that some correlation exists between scattering and H-alpha intensity, as one might expect. However, the H-alpha observations were made with a 50 arcminute beam size; higher resolution observations might provide a more accurate measure of the intensity along the scattering line of sight.

As part of a sensitive, wide-field, CCD H-alpha survey, we made arcminute resolution images of fields containing the Spangler-Reynolds radio sources. The images were taken on the nights of 1995 August 2-3, 24-25, and 29-30. Total integration times were between 1 and 2 hours (consisting of 10-minute images, which could be combined later). We used our Spectral Line Imaging Camera (SLIC) to take the images. Calibration of the images was accomplished by imaging the planetary nebula NGC7027 (in the vicinity of the four observed fields). In our experience, planetary nebulae make excellent calibration sources for our images, providing calibrations with consistencies of 5-10%, and in similar agreement with calibrations based on the use of more extended H-alpha sources, such as the North America Nebula.

Images were taken of the circled regions shown in the map below. The images are numbered following a clockwise order starting from the top image in the map (the region of the largest declination). The resulting images (in GIF format) are image 1, image 2, image 3, and image 4.

[Map of Cygnus region]

Each image is 12 degrees in width. The small circles on each image represent, approximately, the 50-arcminute beams of the Wisconsin Fabry-Perot interferometer, centered on the locations of the radio sources. The Spangler-Reynolds radio sources were numbered 1-8 in their paper; in these images the source locations, in clockwise order through the images, are:

        8, 4, and 1 in the first image;
        2 in the second image;
        3 in the third image; and finally
        6, 5, and 7 in the last image.
This is also the ordering we use in the Table below.

The brightness at any point in an image is due to atmospheric emission, scattered ground light, geocoronal H-alpha emission from the near-Earth environment, scattered starlight, and other backgrounds, plus interstellar H-alpha emission. While a narrow-band interference filter centered on H-alpha reduces the contiuum backgrounds relative to interstellar H-alpha, it is difficult to determine a value for the absolute level of faint interstellar H-alpha emission at any given location, given the fact that the other backgrounds can be as strong or stronger. However, since the other backgrounds vary quite smoothly across an image, it is possible to determine the intensity of an H-alpha structure relative to any local background, or even relative to the local typical intensity level.

By comparing the typical intensity in a 50-arcminute box centered on a radio source location with the intensity within a box of small diameter (a few pixels at most) at the source location, we can determine if the Spangler-Reynolds 50-arcminute H-alpha intensities are in need of any additive adjustment, and by how much. For example, an otherwise smooth distribution of H-alpha emission with a small hole located at the position of a radio source (considerable less than 50 arcminutes in size), would result in an H-alpha intensity assigned by Spangler and Reynolds which would be inaccurately biased upward.

To determine the typical intensity in a 50-arcminute box we used the median intensity in the box (not the mean, which is biased to large values by the presence of any isolated stars in the box). We used a median value in a 3x3 box centered on the radio source as the "on-source" intensity. The difference indicates by how much the Spangler-Reynolds value should be adjusted (in rayleighs, R).

The Table shows the radio sources used by Spangler and Reynolds, their 50-arcminute H-alpha intensities, and our adjustments, based upon our arcminute-resolution images. Typically the adjustment is small enough (relative to the Spangler-Reynolds value) to suggest the original values are should not be changed. In a few cases the Spangler-Reynolds intensities should be lowered by a significant amount.

Table: Adjustments in H-alpha Intensities
Spangler-Reynolds   Spangler-Reynolds    Adjustment    Result     EM
  Source Number   H-alpha Intensity (R)      (R)        (R)   (pc cm**-6)
        8                18                 -5.7       12.3        28
        4                 6.7                ---        6.7        15
        1                 6.8                ---        6.8        15
        2               130                 +7.1      137         308
        3                18                  ---       18          41
        6                35                -12.3       22.7        51
        5                26                  ---       26          59
        7                7.2                -3.6        3.6         8


After adjustment, the table shows the new emission measures (EM(rm pc**-6), approximately equal to 2.25 times the H-alpha intensity, in R). The original figure of Spangler and Reynolds is shown below (displaying the 1-GHz scattering size, in milliarcseconds, versus EM). The red circles and arrows indicate our adjusted EMs. The new EM values are, perhaps, more correlated with the radio source scattering sizes.

[Figure of scattering vs. EM]

This work is a first step towards our using the survey for a more extensive study of the relationship between the warm ionized medium and the scattering of radio waves.

This work was supported by National Science Foundation grant AST-9319670, and a grant from the Horton Foundation to Virginia Tech.