1: \documentclass[preprint]{aastex}
2: %\documentclass[12pt,preprint]{aastex}
3:
4: \usepackage{graphicx}
5: %\usepackage{epstopdf}
6:
7: \newcommand{\feh}{[{\rm Fe/H}]}
8: \newcommand{\znh}{[{\rm Zn/H}]}
9: \newcommand{\nstrs}{10~}
10: \newcommand{\msol}{M_\odot}
11: \newcommand{\pkts}{P_{KS}}
12: \newcommand{\etal}{et al.\ }
13: \newcommand{\delv}{\Delta v}
14: \newcommand{\Lya}{Ly$\alpha$ }
15: \newcommand{\lya}{Ly$\alpha$ }
16: \newcommand{\ndmp}{31 }
17: \newcommand{\nnew}{14 }
18: \newcommand{\nrun}{8,500}
19: \newcommand{\kms}{km~s$^{-1}$ }
20: \newcommand{\cm}[1]{\, {\rm cm^{#1}}}
21: \newcommand{\N}[1]{{N({\rm #1})}}
22: \newcommand{\e}[1]{{\epsilon({\rm #1})}}
23: \newcommand{\f}[1]{{f_{\rm #1}}}
24: \newcommand{\rAA}{{\AA \enskip}}
25: \newcommand{\sci}[1]{{\rm \; \times \; 10^{#1}}}
26: \newcommand{\ltk}{\left [ \,}
27: \newcommand{\ltp}{\left ( \,}
28: \newcommand{\ltb}{\left \{ \,}
29: \newcommand{\rtk}{\, \right ] }
30: \newcommand{\rtp}{\, \right ) }
31: \newcommand{\rtb}{\, \right \} }
32: \newcommand{\ohf}{{1 \over 2}}
33: \newcommand{\nohf}{{-1 \over 2}}
34: \newcommand{\rhf}{{3 \over 2}}
35: \newcommand{\smm}{\sum\limits}
36: \newcommand{\perd}{\;\;\; .}
37: \newcommand{\cmma}{\;\;\; ,}
38: \newcommand{\intl}{\int\limits}
39: \newcommand{\mkms}{{\rm \; km\;s^{-1}}}
40: \newcommand{\ew}{W_\lambda}
41:
42: \newcommand{\vdag}{(v)^\dagger}
43: \newcommand{\myemail}{jet@lco.cl}
44:
45: %\special{papersize=8.5in,11in}
46:
47: \slugcomment{Accepted for publication in PASP: May 2007}
48:
49: \shorttitle{Water Vapor at Las Campanas Observatory}
50: \shortauthors{Thomas-Osip et al.}
51:
52: \begin{document}
53:
54: \title{Calibration of the Relationship between Precipitable Water Vapor and
55: 225~GHz Atmospheric Opacity via Optical Echelle Spectroscopy at Las Campanas
56: Observatory}
57:
58: \author{Joanna Thomas-Osip,\altaffilmark{1} Andrew McWilliam, M.~M. Phillips\altaffilmark{1}, N. Morrell\altaffilmark{1} and I. Thompson}
59: \affil{The Observatories of the Carnegie Institute of Washington\\
60: 813 Santa Barbara St, Pasadena, CA 91101}
61: \email{ jet@lco.cl, andy@ociw.edu, mmp@lco.cl, nmorrell@lco.cl, ian@ociw.edu}
62: %\email{andy@ociw.edu}
63: %\email{mmp@lco.cl}
64: %\email{nmorrell@lco.cl}
65: %\email{ian@ociw.edu}
66:
67: \author{T. Folkers}
68: \affil{Arizona Radio Observatory, University of Arizona,Tucson AZ 85721, USA}
69: \email{tfolkers@hamms.as.arizona.edu}
70:
71: \author{T. Folkers}
72: \affil{Physics Department and Astronomy Department\\
73: University of Michigan, Ann Arbor, MI 48109}
74: \email{fca@umich.edu}
75:
76: \and
77:
78: \author{M. Lopez-Morales}
79: \affil{Carnegie Institution of Washington, Department of Terrestrial Magnetism\\
80: 5241 Broad Branch Road, NW, Washington, DC 20015}
81: \email{mercedes@dtm.ciw.edu}
82:
83:
84: \altaffiltext{1}{Las Campanas Observatory, Colina El Pino, Casilla 601, La Serena, Chile}
85:
86:
87: \begin{abstract}
88:
89: We report precipitable water vapor (PWV) measurements made at Las Campanas
90: Observatory using optical spectra of H$_2$O lines obtained with the Magellan
91: echelle spectrograph, and calculated using a robust technique that is accurate to
92: 5-10\%.
93: Calibration of the relationship between our PWV measurements and opacity values at
94: 225~GHz was made possible by simultaneous tipping radiometer observations. Based on this
95: calibration, we present Las Campanas Observatory winter-time precipitable water vapor
96: statistics, measured using the tipping radiometer, during a two month campaign. The
97: median value of 2.8$\pm$0.3 mm is consistent with that measured at the nearby La Silla
98: Observatory during the VLT site survey.
99: We conclude that, in the Southern hemisphere winter months, we can expect good conditions
100: for infrared observing ($\lesssim$1.5 mm) approximately 10\% of the time at Las Campanas
101: Observatory.
102: \end{abstract}
103:
104: \keywords{radio lines: general --- site testing --- techniques: spectroscopic }
105:
106: \section{Introduction}
107:
108: As part of the Giant Magellan Telescope site testing program at Las Campanas Observatory,
109: we are interested in characterizing precipitable water vapor due to its impact on mid-IR
110: observations through decreased transparency, increased thermal-IR background, and the
111: introduction of extraneous spectral features. In this study, we report on our efforts to
112: absolutely calibrate the relationship between 225~GHz opacity and precipitable water vapor (PWV).
113:
114: The opacity at 225~GHz has been shown to be directly related to the column of precipitable
115: water vapor in the atmosphere \citep[c.f.][]{pla78,cb95,dav97}. The relationship, however,
116: is also a function of pressure (and therefore altitude) and temperature due to the fact that the
117: 225~GHz opacity derives from the wing of a strong water vapor absorption line at 183~GHz
118: arising from the 3-2 rotational transition \citep{wat76} where the wings are mostly due to
119: pressure broadening. Efforts to calibrate this relationship have taken several different
120: forms including deducing the column of PWV from a model of the atmospheric transmission at
121: mm wavelengths and measuring the PWV using radiosonde type devices. Comparison between
122: methods can be complicated by the effects of other atmospheric constituents (e.g. $O_3$,
123: $O_2$, and $N_2$) on the dry air opacity.
124:
125: \citet{swi90} combined high resolution echelle spectroscopy of a water vapor
126: line at 6943.79~\AA\ with a technique devised by \cite{b75} (henceforth B75)
127: to calibrate mid-infrared sky radiance measurements obtained at the La Silla
128: Observatory during the VLT site survey. Through the use of modern partition
129: functions and the addition of 13 new optical lines for PWV measurement, we have
130: improved substantially on the B75 method when used in conjunction with high
131: resolution echelle spectroscopy. Section~\ref{sec-opt} describes the method
132: in detail and our results using it. Our campaign to measure the 225~GHz
133: opacity using a tipping radiometer, the resulting calibration and PWV
134: statistics for Las Campanas Observatory are presented in \S~\ref{sec-tip}.
135:
136: \section{Optical Echelle Spectroscopy}\label{sec-opt}
137:
138: The Magellan echelle spectrograph (MIKE) was used to acquire spectra of
139: rapidly rotating A and B stars in the visual magnitude range 4.0 to 6.0.
140: Due to the high continuous opacity these hot stars show very few absorption lines,
141: particularly in the red optical region. Furthermore, their
142: high rotational velocities ensure that any extant stellar lines are rotationally
143: broadened, to widths much larger than lines arising from absorption through the
144: earth's atmosphere.
145:
146: The Las Campanas Observatory site monitoring program includes a list of 16 telluric
147: standard stars\footnote{http://www.lco.cl/lco/operations-inf/gmt-site-testing-1/stars-for-measuring-pwv-with-mike/stars-for-measuring-pwv}
148: that pass close to the zenith, such that at any time one, or more, stars are within
149: 2 hours of the meridian. Observers were instructed not to submit echelle spectra taken
150: through cirrus clouds as a precaution against spurious PWV measurements arising from
151: clumpy clouds.
152:
153: Typically, spectra were acquired when the hot stars were close to the zenith.
154: The spectral resolving power was normally $\sim$40,000 with S/N in excess of 100:1.
155: The echelle spectral format offers the advantage that light of a given wavelength
156: occurs in two consecutive orders. This permits independent measurement of each line
157: in two orders, providing that the lines fall on the echelle CCD array and there
158: is sufficient signal.
159:
160: Reductions of the raw spectra were performed using the MIKE Pipeline reduction
161: software, written by Dan Kelson, using algorithms outlined in Kelson et al. (2000),
162: Kelson (2003) and Kelson (2006).
163:
164: While the site survey program, including the measurement of PWV, is on-going,
165: this paper is based on measurements of 15 spectra, taken on 15 nights
166: from 21 July to 21 November 2005. The Tipper was not operational on every
167: night that echelle spectra were taken; thus, our Tipper calibration is based
168: on 11 of the 15 spectra.
169:
170: Figures~\ref{fig-bluered} and \ref{fig-drywet} show the spectrum near the H$_2$O line
171: used by B75 at 6943.79~\AA, in consecutive orders on our driest and wettest nights.
172: The equivalent widths of the line in these spectra ranges from 5 to 52~m\AA .
173:
174: \subsection{Analysis of the Spectra}\label{sec-ana}
175:
176: Following B75, we analyze our spectra using a robust and computationally simple method,
177: but we employ improved partition functions that result in a slightly different choice of
178: line excitation potentials than B75. We have also added five new weak optical lines that
179: may be used for reliable PWV estimates.
180:
181: B75 pointed out that for a given temperature there is a certain, unique, energy level
182: which corresponds to a temperature-insensitive absorption coefficient.
183: This follows from the Boltzmann equation (Equation~1)
184: describing the population of the energy levels of atoms and molecules:
185: at any given temperature, T, the population of levels with a unique excitation potential
186: matches the sensitivity of the partition function to temperature; in this case the
187: fractional population of the level (n$_i$/N$_{\rm tot}$) is approximately constant with T,
188:
189: \begin{equation}
190: %\large
191: \frac{n_i}{N_{Tot}} = \frac{g_i e^{-EP_i/kT}}{Q(T)}
192: \end{equation}
193:
194: \noindent where the partition function, Q(T), is given by
195:
196: \begin{displaymath}
197: %\large
198: Q(T) = \sum g_j e^{-EP_j/kT}
199: \end{displaymath}
200:
201: In Figure~\ref{fig-bfac} we show a plot of (n$_i$/N$_{\rm tot}$) versus temperature for
202: three different
203: energy levels, over a range of temperatures expected in the earth's atmosphere; the
204: plots were calculated using Equation~1, and are normalized to a temperature of T=270~K.
205: Figure~\ref{fig-bfac} is similar to Figure~1 in B75, but here we have used the
206: partition function taken from the
207: Kurucz\footnote{http://kurucz.harvard.edu/} website that we believe is more appropriate
208: than the simple approximation for the H$_2$O partition function employed by B75.
209: Figure~\ref{fig-bfac} indicates
210: that for levels with excitation energy in the range 225--300 cm$^{-1}$ the
211: ratio (n$_i$/N$_{\rm tot}$) varies by less than $\sim$5\% over the temperature range
212: from 220 to 300~K. The excitation energy range
213: least sensitive to atmospheric temperature, around 270--280 cm$^{-1}$, is slightly
214: higher than the value of 225 cm$^{-1}$ indicated by B75, and probably results from our
215: use of modern partition functions. We note that even the $\sim$5\% uncertainty could
216: be reduced by use of a weighted average of the results from lines
217: spanning the critical range of excitation energies.
218:
219: Figure~\ref{fig-atmos} shows a
220: MSIS-E-90\footnote{http://modelweb.gsfc.nasa.gov/models/msis.html} model atmosphere
221: for Las Campanas Observatory on one of the nights for which we have a PWV
222: spectrum; it can be seen that the temperature range of the model is covered by the
223: range in Figure~\ref{fig-bfac}. The largest temperature deviation from 270~K, in the model,
224: is 210~K and occurs at the tropopause
225: between 10 and 20~km height; it encompasses approximately 20\% of the mass of
226: the atmosphere. We note that the water vapor pressure above ice, or water, at
227: 220~K is approximately 0.6\% of the vapor pressure at 270~K (e.g. Mason 1971). Thus, if the
228: H$_2$O is uniformly distributed through the atmosphere we estimate
229: that approximately 0.1\% of the mass of the atmospheric water vapor
230: lies in the 10--20~km region, where the temperature differs most from 270~K.
231: In other words the coolest region of the atmosphere contains a negligible fraction
232: of the total water vapor, and thus even the small systematic deviation of
233: n$_i$/N$_{\rm tot}$ from this region will be reduced to insignificance compared to the
234: whole atmosphere. We conclude that lines with excitation levels near 270--280 cm$^{-1}$
235: can provide very robust estimates of the atmospheric PWV.
236:
237: Under the assumption of pure absorption radiative transfer, and a constant
238: line profile function though the earth's atmosphere, B75 pointed out that the
239: PWV, in cm, is given by:
240:
241: \begin{equation}
242: %\large
243: PWV = L /SX
244: \end{equation}
245:
246: \noindent where L is the log flux,
247: S is the line strength parameter, and X is the airmass. B75 referred to
248: L as the ``log equivalent width'', and defined it as the
249: the integral of the natural log of the flux removed from the continuum
250: by the line:
251:
252: \begin{displaymath}
253: %\large
254: L = \int^{\infty}_{-\infty} -log_e \left(\frac{I(\nu)}{I_c}\right)d\nu
255: \end{displaymath}
256:
257: \noindent The line strength parameter S depends on atomic parameters as
258: follows:
259:
260: \begin{equation}
261: %\large
262: S = gf \left(\frac{\pi e^2}{mc}\right) \frac{e^{-EP/kT}}{Q(T)} \left(\frac{N_A}{A_{H_2O}}\right)
263: \end{equation}
264:
265: \noindent Note that the $gf$ value can be calculated from the Einstein
266: spontaneous emission coefficient, $A_{ul}$, by the following relation:
267:
268: \begin{displaymath}
269: %\large
270: gf = 1.4992 \lambda^2 g_{up} A_{ul}
271: \end{displaymath}
272:
273: \noindent where $\lambda$ is in cm and $A_{ul}$ is in Hertz.
274:
275: In Equation~3 $N_A$ is Avogadro's number, 6.022$\times$10$^{23}$, and
276: $A_{H_2O}$ is the atomic mass of H$_2$O, at 18.0. Note that use of
277: Equation~3 will give PWV, in cm, if the log flux integration, L, is performed
278: over the frequency spectrum of the line in Hertz units. For an assumed
279: temperature of 270~K, which is our single model atmosphere parameter,
280: Equation~3 reduces to:
281:
282: \begin{equation}
283: %\large
284: S_{270} = \frac{2.917335\times10^8}{\sigma_{vac}^2} A_{ul} g_{up} e^{-EP/187.6524}
285: \end{equation}
286:
287: \noindent In Equation~4 we have applied the additional factor
288: $-c\sigma_{vac}^2$
289: (where $\sigma_{vac}$ is the vacuum wavenumber) in the calculation of
290: the strength factor, S, to account for the line flux measured in Kaysers
291: (cm$^{-1}$), consistent with the S factors given in B75.
292:
293: In practice the line profile depends most strongly on the atmospheric pressure,
294: which is a function of height, so a constant line profile condition, mentioned by B75,
295: does not apply. However, in the case of unsaturated lines, the line profile does not
296: affect the total absorbed flux, and Equation~2 becomes valid, even with large
297: variations in the line profile through the atmosphere. In this study we
298: computed PWV values from our measured H$_2$O line log fluxes, with wavelengths
299: in cm$^{-1}$ units, using Equations~2 and~4; this is correct for unsaturated
300: lines. For saturated lines the variability of the line profile with height
301: signals the break-down of Equation~2. In this case, to properly measure PWV
302: would require a model atmosphere that accounts for the relation between line
303: profile and column mass of H$_2$O.
304:
305: B75 measured the line strength parameter, S, using the McMath Solar Telescope
306: for numerous lines with excitation near 225 cm$^{-1}$, mostly for lines in the
307: near-infrared region of the spectrum. However, B75's list did include one
308: line in the optical, at 6943.79~\AA. As a check of the B75 value of S for
309: this line, we computed S from the Kurucz\footnote{http://kurucz.harvard.edu/}
310: $gf$ value using Equation~4. The Kurucz $gf$ value, which was taken from
311: theoretical predictions of Partridge \& Schwenke (1997), leads to a value of
312: S which is $\sim$60\% higher than the B75 measurement, although there is
313: reason to suspect the veracity of this result since the relative Kurucz
314: $gf$ values appear to be inconsistent with the appearance of lines in our
315: telluric spectra.
316:
317: Given this disagreement, we decided to check the result from the
318: 6943.79~\AA\ line by identifying additional optical H$_2$O lines with
319: excitation energies near 270--280 cm$^{-1}$ and reliable, measured, $gf$
320: values. Our criteria for selecting the additional lines also included the
321: requirement that they have a strength that could be measured in our spectra,
322: but not so strong that they might be saturated, and that the lines be clear
323: from blends, including H$_2$O lines and other lines present in our telluric
324: spectra (e.g. from the O$_2$ molecule). We searched the HITRAN
325: \footnote{http://cfa-www.harvard.edu/HITRAN/} database of H$_2$O line
326: parameters for suitable lines and visually inspected our telluric spectra to
327: check for non H$_2$O blends.
328:
329: In Table~\ref{tab-params} we show line parameters for the 6943.79~\AA\ line of B75 and 5
330: new weak lines identified in this work. The HITRAN database included optical H$_2$O line
331: $gf$ values from a variety of sources. The $gf$ values for our lines in
332: Table~\ref{tab-params} came from Coheur et al. (2002), measured using a high resolution
333: spectrograph and a controlled, fixed, water vapor path. For the 8 strong
334: H$_2$O lines listed in Table~\ref{tab-strong} that were used for the saturation
335: investigation (see \S~\ref{sec-sat}), the $gf$ values were taken from
336: Coheur et al. (2002) and similar laboratory measurements of Brown, Toth, \& Dulick (2002).
337: For the 6943.79~\AA\ line the empirical Coheur et al. (2002) $gf$ value indicated a line
338: strength parameter, S, that agreed with the B75 value to within 1\%. Given the good
339: agreement between the experimentally determined S factors, and the
340: difficulty in making theoretical $gf$ predictions for weak lines (due to small overlap of
341: the level wavefunctions), we chose not to use the Kurucz H$_2$O $gf$ values.
342:
343: \subsection{Results of Optical PWV Measurements}\label{sec-optres}
344:
345: The range of line equivalent widths, in m\AA , for our list of temperature insensitive
346: H$_2$O lines is indicated in Table~\ref{tab-params}. Table~\ref{tab-pwv} shows the log
347: line fluxes and the resultant PWV values computed using Equations~2 and~4 (multiplied by
348: 10 to get PWV in mm). The PWV values range from 1 to 7mm for the 15 nights, and there is
349: good agreement, typically $\sim$10\%, between the results for each line. This dispersion
350: is consistent with the accuracy of the measured line $gf$ values: good $gf$ values are
351: usually accurate to 5 to 10\%.
352: The weakest line in Table~\ref{tab-params} is at 5954.9~\AA , and is most reliable on nights
353: with more than $\sim$5mm PWV. The strongest line in Table~\ref{tab-params}, at
354: 7287.4~\AA, is useful for measuring PWV on very dry nights; on wet nights it will be the
355: first line to become saturated.
356:
357: In Figure~\ref{fig-pwv_comp} we compare the individual PWV results for each line with
358: the average
359: for each spectrum. The figure shows no clear systematic differences between the PWV
360: values for the individual lines. The rms scatter of the individual line PWV measurements
361: about the mean for each spectrum is 0.3mm H$_2$O on average, corresponding to $\sim$10\%.
362:
363: In Figure~\ref{fig-pwv_comp} the 7287.4~\AA\ point for the wettest night appears below the
364: main correlation, slightly more than might be expected from the general scatter of the
365: points. In addition to this for the next two most wettest nights the 7287.4~\AA\ PWV points
366: are also the lowest of the 6 lines used. Although these differences are extremely
367: subtle, and could easily be random, they may indicate the beginning of saturation in
368: the 7287.4~\AA\ line; indeed, the expected effect of saturation is to reduce the computed
369: PWV from Equations~2 and~4. If that is the case then the strength of our 7287.4~\AA\ line on
370: the wettest night, at 134 m\AA , represents the useful upper limit for PWV measured using
371: unsaturated H$_2$O lines under the assumption of no saturation.
372:
373: The method used to measure PWV here could be applied to sites at higher altitude,
374: e.g. Mauna Kea. The lines will be narrower at higher altitude, due to the lower pressure
375: broadening, and thus, saturation will occur at smaller equivalent widths.
376:
377:
378: \subsection{Saturation Investigation}\label{sec-sat}
379:
380: As mentioned earlier, the Brault method used to compute PWV in this paper is quite robust,
381: providing that the lines are not saturated. To use saturated lines for reliable PWV
382: measurement would require a detailed model atmosphere treatment where the run of temperature
383: with PWV, or fractional PWV, is specified through the atmosphere. Given the small scatter in
384: our results it seems likely that all of the lines measured in Table~\ref{tab-params}
385: are unsaturated, although the line at 7287.36~\AA\ may be low on our wettest night,
386: indicating the onset of saturation. The use of saturated lines will result in lower computed
387: PWV values, with the error becoming more pronounced with increasing line strength.
388:
389: It would be useful to know the approximate saturation limit in terms of line strength,
390: so that we do not use overly strong lines in the PWV calculations; this is particularly
391: important for the wettest nights. In order to do this we have made a search for strong
392: H$_2$O lines in the HITRAN database, with lower level excitation energies in the range
393: 225--300 cm$^{-1}$ (the range that gives PWV results insensitive to atmosphere
394: temperature). We have calculated the parameters for this list of 8 strong H$_2$O lines,
395: as shown in Table~\ref{tab-strong}, along with the equivalent widths of each line
396: on our wettest night (2 August 2005). It is interesting that the stronger H$_2$O lines
397: tend to lie at redder wavelengths than those in Table~\ref{tab-params}.
398:
399: In Figure~\ref{fig-pwv_sat} we show a plot of reduced equivalent width
400: (REW = log$_{10}$ EW/$\lambda$ ) for all 14 H$_2$O lines (from Tables~\ref{tab-params} and
401: \ref{tab-strong}) versus PWV, for the nights of 2 and 20 August 2005, calculated using
402: Equation~2. It is clear that the stronger lines on 2 August give systematically lower PWV
403: values, decreasing roughly linearly with increasing strength; for lines weaker than
404: REW$\sim$$-$4.6 to $-$4.7 the PWV values are approximately constant. This saturation limit
405: corresponds to
406: $\sim$160~m\AA\ at 7300~\AA ; this is larger than the 134~m\AA\ for the line at 7287.36~\AA\
407: on the wettest night, which we had previously suspected could be affected by saturation.
408: Therefore, future PWV measurements that use the Brault method should disregard lines above
409: REW$\sim$$-$4.65. The 20 August points in Figure~\ref{fig-pwv_sat} show that when the
410: lines are below the saturation limit they give results in good agreement; this eliminates
411: the possibility that the decreasing PWV values for strongest lines on the wettest night are
412: due to a systematic error in the line $gf$ values. The unsaturated 20 August points have a
413: standard deviation of 11\% in PWV, giving an error on the mean PWV of only 3.3\%.
414:
415: Because line saturation in red giant stars begins around REW$\sim$$-$5.1 it is clear that
416: the telluric H$_2$O lines must be more strongly broadened than lines in red giants.
417: Atmospheres of red giant stars are much hotter and have much larger turbulent velocities
418: than the earth's atmosphere; therefore, the main source of broadening for the telluric
419: H$_2$O lines cannot be due to temperature or turbulence. The only possibility is that the
420: larger broadening of the telluric lines is due to the much denser earth's atmosphere
421: (i.e. pressure broadening), for which one would expect the collision frequency and
422: Van der Waals forces to be greater than for lines formed in the low density atmospheres of
423: stars.
424:
425: Given the larger saturation point for telluric lines, than for stellar lines, the pressure
426: broadening must be overwhelmingly dominant. Thus, it may be that the saturation point
427: changes with atmospheric pressure, for example due to the change in the collision frequency
428: experienced by the H$_2$O molecules. It is also possible that the saturation point is
429: affected by humidity, both because of the presumably larger H$_2$O collision cross-section,
430: $\sigma$v, and also resulting from different Van der Waals forces for H$_2$O compared to
431: other atmospheric gases.
432:
433: Future PWV measurements can utilize the stronger H$_2$O lines listed in Table~\ref{tab-strong},
434: provided that the saturation limit is not exceeded.
435:
436:
437: \subsection{Temperature Diagnostic}
438:
439: We have attempted to estimate the mean temperature for the wettest night, 2 August 2005, using
440: H$_2$O lines covering a range of excitation energies. Iterative calculations were performed
441: to find the temperature where the line PWVs matched the temperature-independent PWV value. This
442: mean temperature is actually weighted by the water vapor mass, so it should reflect the
443: temperature closer to the ground, where most of the water vapor lies.
444:
445: We identified and measured nine lines from the HIRTRAN database with energies ranging from 0 to
446: 756 cm$^{-1}$. The derived temperatures ranged from 268~K to 290~K, with a mean of 279~K and an
447: rms scatter of 9.5~K. Some of the scatter may have resulted from blending of weak high
448: excitation lines. We note that the mean temperature verifies our use of a temperature of 270~K
449: for the PWV calculations.
450:
451: Apart from verifying our initial assumption it is not clear how useful the mean temperature
452: diagnostic would be; perhaps it may be used to compare with mass-weighted temperatures
453: calculated from tailored atmosphere models. A better way to compare with theoretical models
454: would be to use profiles of strongly saturated lines to derive the run of temperature with
455: H$_2$O opacity; but this would require accurate estimates of the line broadening in the
456: theoretical models.
457:
458: \section{225~GHz Tipping Radiometer Opacity}\label{sec-tip}
459:
460: Opacity at 225~GHz was studied through the use of the University of Arizona Radio Observatory
461: (ARO) 225~GHz Tipping Radiometer (Tipper) on loan to Las Campanas Observatory for the 2005 Southern hemisphere winter season. This radiometer is one of four originally constructed for Millimeter Array site testing purposes by the NRAO and has been in regular use at ARO for sky quality assesment since 1995. These systems have been used to characterize many other sites including Mauna Kea (one of the units has been in operation since 1989 at the Caltech Submillimeter Observatory), Antarctica, Cerro Chanjnantor in Chile, South Baldy and the VLA site in New Mexico \citep{oh89,sch90,cha04}.
462:
463: The characteristics of this radiometer have been extensively described elsewhere
464: \citep{liu87,mc87,cb94} therefore only a short description is provided here. The radiometer
465: is illuminated via a mirror (with a beamwidth of 3.4 degrees) that rotates through 11 elevation
466: angles between an airmass of one and three thus providing a "tip" (measurement of the sky
467: brightness as a function of zenith angle). The beam enters a temperature controlled box in
468: which a chopper wheel allows the receiver to alternately view in repeatable sequences the sky,
469: a cold (45~C) load, the sky, and finally a hot (65~C) load. Use of two different
470: temperature reference loads allows the gain to be measured. The opacity and an estimate of its
471: uncertainty are then derived according to a slab model of the atmosphere from a linear fit to
472: the natural logarithm of the sky brightness temperature as a function of airmass.
473:
474: Opacities were obtained approximately every 5 minutes continuously between
475: between July 17, 2005 and August 30, 2005 with some exceptions due to storm
476: activity. Of these data, 2999 measurements were selected from the 29 full
477: and partial nights when the conditions were judged to be photometric.
478: The requirement of photometric conditions was imposed not only because of
479: concerns that non-photometric conditions would lead to incorrect fluxes and
480: opacity measurements from the Tipping Radiometer, but also due to the desire
481: to characterize the PWV on nights suitable for mid-infrared astronomy.
482:
483: \subsection{Relationship of 225~GHz opacity to precipitable water vapor}\label{sec-cal}
484:
485: The relationship between 225~GHz opacity, $\tau$, and the column of precipitable water vapor,
486: PWV in mm, can be represented as follows:
487: \begin{equation}
488: %\large
489: \tau=\tau_{dry-air}+ B \cdot PWV
490: \end{equation}
491: where the opacities and the coefficient B are reported in nepers/airmass and nepers/airmass/mm
492: $H_2O$, respectively. As used here, a neper is a measurement of attenuation similar to a decibel but based on the natural logarithm, as opposed to base 10.
493:
494: \citet{pla78} reports the first estimate of the relationship between $\tau$ and PWV. The
495: opacity measurements made at UC Hat Creek Observatory (with an altitude of 1050-m) were
496: correlated with PWV as measured by a radiosonde in Medford, OR (210 km to the NW). Although
497: no values below 4 mm PWV were measured, the coefficient B was found to be 0.06. In the
498: development of the NRAO tipping radiometers, one of which was used in our study, \citet{mc87}
499: used a similar relationship but included a constant opacity offset of 0.005 nepers/airmass due
500: to absorption by oxygen. They also mention that the relationship is weakly dependent on elevation.
501:
502: \citet{cb95} report on their calibration efforts in Antarctica using another NRAO tipping
503: radiometer to measure 225~GHz opacities and Radiosonde upper air soundings to measure PWV.
504: Their coefficients can be found in Table \ref{tab-cal}. A re-analysis of this data is
505: presented by \citet{cha04} but does not change the finding of a significant non-zero dry air opacity.
506: The cause of the dry air opacity is not well understood as many atmospheric transmission models
507: \citep[e.g][]{par01, gro89, lie89} have failed to reproduce it. A new model that accounts theoretically for the collision-induced absorption by $N_2$-$N_2$ molecular partners may be able to resolve this discprepancy \citep{pai04}. Further work in understanding and characterizing this parameter is especially important to characterizing the low-precipitable-water-vapor sites at mm wavelengths.
508:
509: At Mauna Kea this relationship was established in a different way \citep{dav97} and the
510: coefficients can also be found in Table \ref{tab-cal}. The 225~GHz opacities were measured
511: with the Caltech Submillimeter Observatory (CSO) tipping radiometer, also one of the original
512: NRAO tipping radiometers. The PWV values were obtained through the use of an atmospheric
513: spectral synthesis model called FASCOD2. As the PWV has not been independently measured, it
514: is unclear whether this model might also underestimate the dry air opacity. Given that this
515: calibration is now canonically used to produce PWV values from 225~GHz opacities for Mauna Kea
516: as well as to draw comparisons to other sites, it is important to measure the dry air opacity
517: there.
518:
519: At Las Campanas Observatory, the precipitable water vapor values determined with the MIKE
520: echelle spectrograph in Section \ref{sec-optres} and shown in Table \ref{tab-pwv} were
521: correlated with the 225~GHz opacity measurements as shown in Figure \ref{fig-pwv_tau}. The
522: opacities at the nearest times to those in Table \ref{tab-pwv} were selected. There was
523: never more than 5 minutes difference and a total of 11 points with times in common were found.
524: A linear relationship was determined using a linear least squares regression with errors in
525: both the independent and dependent variables. Our coefficients are also listed in
526: Table \ref{tab-cal}. More points would be needed to better constrain the dry air opacity.
527:
528: A meaningful comparison of the available calibration coefficients at different sites is made
529: difficult by the lack of uncertainties reported for the calibrations cited in
530: Table \ref{tab-cal}. \citet{cb95} reported very small uncertainties for their calibration
531: efforts at the South Pole. Their results, however, were superceded by a re-analysis of the
532: tipping radiometer measurements in which the opacities were adjusted systematically upwards
533: by 15\% \citep{cha04}. No mention of the relative uncertainties was made in \citet{cha04}
534: and since adjustments were systematic in nature we will quote the original uncertainty values from \citet{cb95}
535: in Table \ref{tab-cal}. Regardless of the lack of uncertainties in the reported
536: calibrations, it appears that the dry air opacity decreases with altitude and as expected
537: due to the pressure broadening seen in the wing of the 183~GHz line, the coefficient B,
538: also decreases with altitude. In other words, at lower altitudes (higher pressure) less
539: PWV is required to produce the same opacity at 225 GHz as seen at higher altitudes.
540:
541: \subsection{Calibrated precipitable water vapor at Las Campanas Observatory}\label{sec-pwv}
542:
543: Given the calibration determined in the previous section, millimeters of precipitable water
544: vapor and its uncertainty estimate are calculated from the tipper opacities and errors in
545: the following manner:
546:
547: \begin{eqnarray}
548: %\large
549: PWV=\frac{\tau - \tau_{dry-air}}{B}\\
550: \sigma_{pwv}=\sqrt{\sigma_{\tau_{dry-air}}^2+(\tau \cdot \sigma_B)^2+B \cdot \sigma_{\tau}^2}\label{eq:pwvcal}
551: \end{eqnarray}
552:
553: Figure \ref{fig-tippwv} shows the resulting precipitable water vapor values at night under
554: photometric conditions as a function of time during our campaign. The fraction of the
555: measurements for which PWV was found to be below a given value is displayed in
556: Figure \ref{fig-tippwvhist}. Uncertainties in our PWV estimates are on the order of 10\%.
557:
558: Over the 1.5 month period covered by our measurements, we find a median
559: PWV of 2.8$\pm$0.3 mm. Assuming that this time period is representative of
560: the winter PWV characteristics at Las Campanas Observatory, then we can
561: expect good
562: conditions for mid-infrared observing ($<$1.5 mm PWV) 10\% of the time in
563: winter \citep[c.f.][]{gio01}. These results
564: agree well with measurements made at the nearby La Silla Observatory during
565: the VLT site survey \citep{swi90}. La Silla is located 24~km south of Las
566: Campanas, and is nearly exactly the same altitude. The VLT site survey
567: measurements were obtained using a mid-infrared sky radiance monitor
568: \citep{mg82}. The absolute scale of these data was calibrated using the
569: B75 method on coud\'{e} spectra of the 6943.79~\AA\ line obtained
570: simultaneously with mid-infrared sky radiance measurements on several
571: nights.
572:
573: \section{Conclusions}
574:
575: We report on the calibration of the relationship between the column of precipitable water
576: vapor, PWV, and opacity at 225~GHz at Las Campanas Observatory as measured by an NRAO
577: Tipping Radiometer on loan from the University of Arizona Radio Observatory and the high
578: resolution echelle spectrograph, MIKE, on the Magellan Clay Telescope. We have expanded
579: the absolutely calibrated method for measuring PWV using temperature insensitive lines in
580: high resolutions stellar spectra presented by \citet{b75} with improved partition functions
581: and 13 additional lines. We found that optical H$_2$O lines with reduced equivalent
582: widths weaker than $\sim$$-$4.7 to $-$4.6 are unsaturated and suitable for measuring PWV
583: using the Brault method. This saturation level indicates that the lines are strongly
584: pressure broadened. A calculation of the mass-weighted mean temperature validates the
585: use of a single slab model atmosphere with T=270~K to compute PWV values. The linear
586: relationship we found between the MIKE PWV and the 225~GHz opacities is consistent with
587: opacities formed from a pressure broadened wing of the strong water vapor line at 183~GHz
588: and including a dry-air opacity component. The effect of altitude (due to atmospheric
589: pressure changes) is demonstrated by comparing our calibration at Las Campanas Observatory
590: with those few available in the literature. We note the paucity of such absolute PWV
591: calibration efforts and encourage further work especially at sites like Mauna Kea and
592: Paranal where the capability for high resolution echelle spectra already exists.
593:
594:
595: Based on the relationship determined between the measured PWV and 225~GHz
596: opacities, we find a median PWV of 2.8$\pm$0.3 mm and a range between 0.5~mm
597: and 7.5~mm for our two month Southern hemisphere winter time campaign. This
598: is consistent with that measured at the nearby La Silla Observatory during
599: the VLT site survey \citet{swi90}. Finally, if our campaign is representative
600: of the Southern hemisphere winter months, we can expect good conditions for
601: infrared observing ($\lesssim$1.5 mm) at the tenth percentile level.
602:
603: \acknowledgments
604:
605: We are grateful to Lucy Ziurys and the Arizona Radio Observatories for the loan of the Tipping Radiometer. We also appreciate Mike Meyer's help in arranging for the loan. A.M. gratefully acknowledges support from NSF grants AST-96-18623 and AST-00-98612 and I.T. from NSF grant AST-0507325. J.T-O. thanks John Bally for a particularly fruitful conversation regarding PWV calibration efforts and John Grula for his superb efforts in tracking down obscure references. Finally, we are indebted to Magellan/MIKE observers, Andrea Dupree, Varsha Kulkarni, James Lauroesch, and George Preston, for obtaining spectra.
606:
607: % I.T. acknowledges support from NSF grant XXX.
608:
609: {\it Facilities:} \facility{Magellan (MIKE)}, \facility{ARO (Tipper)}.
610:
611: \begin{thebibliography}{}
612:
613: \bibitem[Brault, Hall, \& Fender (1975)]{b75}
614: Brault, J.W., Fender, J.S., \& Hall, D.N.B. 1975, \jqsrt, 15, 549.
615:
616: \bibitem[Brown et al. (2002)]{b02}
617: Brown, L.R., Toth, R.A., \& Dulick, M. 2002, J. Mol. Spectrosc., 212, 57
618:
619: \bibitem[Coheur et al.(2000)]{c00}
620: Coheur, P.-F., Fally, S., Carleer, M., Clerbaux, C., Colin, R. et al.
621: 2000, \jqsrt, 74, 493.
622:
623: \bibitem[Chamberlin(2004)]{cha04}
624: Chamberlin, R.A. 2004, \pasa, 21, 264.
625:
626: \bibitem[Chamberlin \& Bally(1994)]{cb94}
627: Chamberlin, R.A. \& Bally J. 1994, \ao, 33,1095.
628:
629: \bibitem[Chamberlin \& Bally(1995)]{cb95}
630: Chamberlin, R.A. \& Bally J. 1995, Int. J. Infrared Millimeter Waves, 16, 907.
631:
632: \bibitem[Davis et al.(1997)]{dav97}
633: Davis, G. R., Naylor, D. A., Griffin, M. J., Clark, T. A., \& Holland, W. S. 1997, \icarus, 130, 387.
634:
635: \bibitem[Delgado et al.(1999)]{del99}
636: Delgado, G., A. Otarola, V. Belitsky, D. Urbain 1999, ALMA memo 271.1.
637:
638: \bibitem[Giovanelli et al.(2001)]{gio01}
639: Giovanelli, R., J. Darling, C. Henderson, W. Hoffman, D. Barry, J. Cordes, S. Eikenberry, G. Gull, L. Keller, J. D. Smith, \& G. Stacey 200a, \pasp, 113, 803.
640:
641: \bibitem[Grossman(1989)]{gro89}
642: Grossman, E. 1989, A T Program, Ver 1.5, Air Head Software, Boulder CO.
643:
644: \bibitem[Kelson et al.(2000)]{kelson00}
645: Kelson D.D., Illingworth, G.D., van Dokkum, P.G., \& Franx, M. 2000, \apj, 531, 159.
646:
647: \bibitem[Kelson(2003)]{kelson03}
648: Kelson D.D. 2003, \pasp, 115, 688.
649:
650: \bibitem[Kelson(2006)]{kelson06}
651: Kelson D.D. 2006, \pasp, {\em submitted}.
652:
653: \bibitem[Liebe(1989)]{lie89}
654: Liebe, H.J., 1989, Int. J. Infrared Millimeter Waves, 10, 631.
655:
656: \bibitem[Liu(1987)]{liu87}
657: Liu, Z-Y. 1987, NRAO Millimeter Array Memo 41.
658:
659: \bibitem[Mason (1971)]{mas71}
660: Mason, B.J. 1971, ``The Physics of Clouds'' (Clarendon Press)
661:
662: \bibitem[McKinnon(1987)]{mc87}
663: McKinnon, M. 1987, NRAO Millimeter Array Memo 40.
664:
665: \bibitem[Morse \& Gillette(1982)]{mg82}
666: Morse, D. \& Gillett, F. 1982, AURA Eng. Report 73.
667:
668: \bibitem[Owen \& Hogg(1989)]{oh89}
669: Owen, F.N., \& Hogg, D.E. 1989, Proc. of URSI/IAU Symposium on Radio Astronomical Seeing, J.E. Baldwin \& W. Shouguan, Oxford: Pergamon Press, 42.
670:
671: \bibitem[Pardo, Cernicharo \& Serabyn(2001)]{par01}
672: Partridge, H. \& Schwenke, D.W. 1997, J. Chem. Phys. 106, 4618.
673:
674: \bibitem[Paine(2004)]{pai04}
675: Paine, S. 2004, SMA Tech. Memo 152.
676:
677: \bibitem[Partridge \& Schwenke(1997)]{ps97}
678: Pardo,J. R., Cernicharo, J. \& Serabyn, E. 2001, ITAP, 49/12, 1683.
679:
680: \bibitem[Plambeck(1978)]{pla78}
681: Plambeck, R. L. 1978, IEEE Trans Antennas Propag, AP-26, 737.
682:
683: \bibitem[Schwab et al.(1990)]{sch90}
684: Schwab, F.R., Hogg, D.E., \& Owen F.N. 1990, Proc. of URSI/IAU Symposium on Radio Astronomical Seeing, J.E. Baldwin \& W. Shouguan, Oxford: Pergamon Press, 116.
685:
686: \bibitem[Swings et al.(1990)]{swi90}
687: Swings, J.-P., Appenzeller, I., Ardeberg, A., Charvin, P.,
688: Lelievre, G., Perrier, C., Sarazin, M., Schuster, H.-E. \& Shaver, P.
689: 1990, VLT Site Selection Working Group Final Report, European Southern
690: Observatory VLT Report No. 62, ed. M. Sarazin
691:
692: \bibitem[Waters(1976)]{wat76}
693: Waters, J.W. 1976, "Absorption and Emission by Atmospheric Gases" in Methods of Experimental Physics, 12-B, Meeks, M.L. New York: Academic Press, 142.
694:
695: \end{thebibliography}
696:
697: \clearpage
698:
699: \begin{deluxetable}{rrrcccl}
700: \tablenum{1}
701: \tablewidth{0pt}
702: \tablecaption{Log of Observations}
703: \tablehead{\colhead{Obs. \#} & \colhead{HR \#} & \colhead{UT Date} & \colhead{UT Start} & \colhead{Airmass} & \colhead{Exp. (sec)} &
704: \colhead{Observer\tablenotemark{*} } }
705: \startdata
706: 1 & 5987 & 21/Jul/2005 & 22:49 & 1.116 & 24 & ALM \cr
707: 2 & 5987 & 22/Jul/2005 & 22:46 & 1.115 & 32 & ALM \cr
708: 3 & 806 & 24/Jul/2005 & 11:01 & 1.292 & 30 & ALM \cr
709: 4 & 5987 & 25/Jul/2005 & 01:37 & 1.033 & \hskip6pt 5 & AD \cr
710: 5 & 845 & 30/Jul/2005 & 10:42 & 1.002 & 15 & IT \cr
711: 6 & 5517 & 30/Jul/2005 & 22:44 & 1.002 & 15 & IT \cr
712: 7 & 5517 & 1/Aug/2005 & 22:48 & 1.001 & 15 & IT \cr
713: 8 & 5517 & 2/Aug/2005 & 23:08 & 1.005 & 15 & IT \cr
714: 9 & 6930 & 20/Aug/2005 & 01:52 & 1.043 & 15 & GWP \cr
715: 10 & 5987 & 20/Aug/2005 & 23:10 & 1.011 & 32 & GWP \cr
716: 11 & 5987 & 23/Aug/2005 & 23:08 & 1.014 & 62 & GWP \cr
717: 12 & 5987 & 24/Aug/2005 & 22:52 & 1.011 & 30 & GWP \cr
718: 13 & 6930 & 15/Sep/2005 & 01:00 & 1.098 & 10 & KL \cr
719: 14 & 8998 & 21/Nov/2005 & 23:45 & 1.030 & 49 & AMcW \cr
720: 15 & 3034 & 22/Nov/2005 & 08:21 & 1.002 & \hskip6pt 3 & NM \cr
721:
722: \enddata
723: \tablenotetext{*}{Observer key: AD (Andrea Dupree), IT (Ian Thompson), GWP (George Preston),
724: KL (Kulkarni \& Lauroesch), AMcW (Andrew McWilliam), NM (Nidia Morel),
725: ALM (Adams \& Lopez Moralez)}
726: \label{tab-log}
727: \end{deluxetable}
728:
729: \clearpage
730:
731: \begin{deluxetable}{cccllcc}
732: \tablenum{2}
733: \tablewidth{0pt}
734: \tablecaption{Parameters of Temperature Insensitive H$_2$O Lines\tablenotemark{*}}
735: %\tablehead{\colhead{$\sigma_{\rm vac}$ (cm$^{-1}$)} & \colhead{$\lambda_{\rm air}$ (\AA )}
736: % & \colhead{ EP (cm$^{-1}$)} & \colhead{gf}
737: % & \colhead{S$_{270}$ (cm$^{-2}$)} & EW$_{dry}$ (m\AA ) & EW$_{wet}$ (m\AA ) \cr
738: \tablehead{\colhead{$\sigma_{\rm vac}$} & \colhead{$\lambda_{\rm air}$}
739: & \colhead{EP} & \colhead{gf}
740: & \colhead{S$_{270}$} & \colhead{EW$_{dry}$} & \colhead{EW$_{wet}$} \cr
741:
742: \colhead{(cm$^{-1}$)} & \colhead{(\AA )}
743: & \colhead{(cm$^{-1}$)} &
744: & \colhead{(cm$^{-2}$)} & \colhead{(m\AA )} & \colhead{(m\AA )}
745: }
746: \startdata
747: 13717.1744 & 7288.10 & 275.50 & 5.323$\times$10$^{-9}$ & 0.2387 & 6 & 88 \cr
748: 13718.5755 & 7287.36 & 300.36 & 1.251$\times$10$^{-8}$ & 0.4912 & \llap{2}1 & \llap{1}34 \cr
749: 13823.1810 & 7232.21 & 300.36 & 5.602$\times$10$^{-9}$ & 0.2200 & 9 & 75 \cr
750: 13894.6353 & 7195.02 & 285.42 & 6.159$\times$10$^{-9}$ & 0.2619 & \llap{1}2 & 84 \cr
751: 14397.3641 & 6943.79 & 224.84 & 3.059$\times$10$^{-9}$ & 0.1795\rlap{\tablenotemark{**}} & 5\rlap{:} & 52 \cr
752: 16788.1104 & 5954.94 & 300.36 & 8.942$\times$10$^{-10}$ & 0.03511 & ... & \hskip6pt 6 \cr
753: \enddata
754: \tablenotetext{*}{S and gf values calculated from Coheur et al. (2002) A values.}
755: \tablenotetext{**}{Compares well with Brault et al. (1975) S value of 0.180.}
756: \label{tab-params}
757: \end{deluxetable}
758:
759: \clearpage
760:
761: \begin{deluxetable}{r|cc|cc|cc|cc|cc|cc|cc}
762: \tabletypesize{\tiny}
763: \rotate
764: \tablenum{3}
765: \tablecolumns{15}
766: \tablewidth{0pt}
767: \tablecaption{Line Fluxes\tablenotemark{*} and PWV Values}
768: \tablehead{
769: \colhead{Obs.} &
770: \multicolumn{2}{c}{7288.10\AA} & \multicolumn{2}{c}{7287.36\AA} &
771: \multicolumn{2}{c}{7232.21\AA} & \multicolumn{2}{c}{7195.02\AA} & \multicolumn{2}{c}{6943.78\AA} &
772: \multicolumn{2}{c}{5954.94\AA\tablenotemark{\ddag}} & \colhead{$\overline{\rm PWV}$\tablenotemark{\dag} (mm)} & \colhead{$\sigma$} \cr
773: & \colhead{$-$ln F} & \colhead{PWV} & \colhead{$-$ln F} & \colhead{PWV} & \colhead{$-$ln F} & \colhead{PWV} &
774: \colhead{$-$ln F} & \colhead{PWV} & \colhead{$-$ln F} & \colhead{PWV} & \colhead{$-$ln F} & \colhead{PWV} & & }
775: \startdata
776: 1 & ... & ... & ... & ... & 0.036\rlap{:} & 1.47 & 0.047\rlap{:} & 1.61 & 0.041 & 2.05 & ... & ... & 1.71 & 0.30 \cr
777: 2 & ... & ... & ... & ... & 0.032\rlap{:} & 1.31 & 0.031\rlap{:} & 1.06 & 0.028 & 1.40 & ... & ... & 1.26 & 0.17 \cr
778: 3 & ... & ... & ... & ... & ... & ... & 0.039\rlap{:} & 1.15 & 0.019\rlap{:} & 0.82 & ... & ... & 0.99 & 0.24 \cr
779: 4 & 0.013\rlap{:} & 0.53 & 0.044 & 0.87 & 0.023\rlap{:} & 1.01 & 0.027\rlap{7} & 1.00 & 0.027 & 1.19 & ... & ... & 0.97 & 0.30 \cr
780: & & & & & & & & & 0.017\rlap{:} & & & & & \cr
781: 5 & 0.122 & 5.10 & 0.216 & 4.39 & 0.116 & 4.92 & 0.126 & 4.80 & 0.095 & 4.92 & 0.022\rlap{:} & 5.26 & 4.85 & 0.36 \cr
782: & & & & & 0.101 & & & & 0.082 & & 0.015 & & & \cr
783: 6 & 0.110 & 4.60 & 0.196 & 3.98 & 0.098 & 4.12 & 0.127 & 4.84 & 0.079 & 4.42 & 0.011\rlap{:} & 3.13 & 4.36 & 0.32 \cr
784: & & & & & 0.087 & & & & 0.080 & & 0.011 & & & \cr
785: 7 & 0.093 & 3.89 & 0.184 & 3.74 & 0.085 & 3.79 & 0.100 & 3.81 & 0.060 & 3.68 & 0.013 & 3.70 & 3.77 & 0.21 \cr
786: & & & & & 0.082 & & & & 0.072 & & & & & \cr
787: 8 & 0.184 & 7.67 & 0.293 & 5.94 & 0.166 & 7.17 & 0.190 & 7.22 & 0.127 & 6.90 & 0.295\rlap{:} & 6.84 & 7.00 & 0.57 \cr
788: & & & & & 0.151 & & & & 0.122 & & 0.188 & & & \cr
789: 9 & 0.064 & 2.57 & 0.123 & 2.40 & 0.063 & 2.94 & 0.063 & 2.31 & 0.048 & 2.30 & 0.008\rlap{::} & 2.18 & 2.54 & 0.35 \cr
790: & & & & & 0.072 & & & & 0.038 & & & & & \cr
791: 10 & 0.063 & 2.61 & 0.121 & 2.44 & 0.065 & 2.95 & 0.067 & 2.53 & 0.049 & 2.45 & ... & ... & 2.63 & 0.27 \cr
792: & & & & & 0.066 & & & & 0.040 & & & & & \cr
793: 11 & 0.101 & 4.17 & 0.190 & 3.81 & 0.095 & 4.35 & 0.101 & 3.80 & 0.073 & 3.87 & ... & ... & 4.03 & 0.27 \cr
794: & & & & & 0.099 & & & & 0.068 & & & & & \cr
795: 12 & 0.085 & 3.52 & 0.158 & 3.18 & 0.075 & 3.64 & 0.083 & 3.14 & 0.061 & 3.20 & ... & ... & 3.36 & 0.30 \cr
796: & & & & & 0.087 & & & & 0.055 & & & & & \cr
797: 13 & 0.040 & 1.53 & 0.076 & 1.30 & 0.036 & 1.49 & 0.040 & 1.39 & 0.033 & 1.60 & ... & ... & 1.49 & 0.12 \cr
798: & & & & & 0.036 & & & & 0.030 & & & & & \cr
799: 14 & ... & ... & 0.122\rlap{:} & 2.44 & 0.034\rlap{:} & 1.63 & ... & ... & 0.037 & 2.40 & ... & ... & 2.10 & 0.54 \cr
800: & & & & & 0.038 & & & & 0.052\rlap{:} & & & & & \cr
801: 15 & 0.050 & 2.03 & 0.094 & 1.78 & 0.052 & 2.19 & 0.058 & 2.15 & 0.036 & 2.00 & ... & ... & 2.05 & 0.17 \cr
802: & & & & & 0.047 & & & & 0.036 & & & & & \cr
803: \enddata
804: \label{tab-pwv}
805: \tablenotetext{*}{\ Flux measurements for lines appearing in a second order are indicated below the first entry}
806: \tablenotetext{\dag}{\ Lines appearing in two orders were counted twice for the calculation of the average PWV}
807: \tablenotetext{\ddag}{\ The 5954.94\AA\ line was not used for the average PWV}
808: \end{deluxetable}
809:
810: \clearpage
811:
812: \begin{deluxetable}{cccllr}
813: \tablenum{4}
814: \tablewidth{0pt}
815: \tablecaption{Parameters for Strong H$_2$O Lines}
816: %\tablehead{\colhead{$\sigma_{\rm vac}$ (cm$^{-1}$)} & \colhead{$\lambda_{\rm air}$ (\AA )}
817: % & \colhead{ EP (cm$^{-1}$)} & \colhead{gf}
818: % & \colhead{S$_{270}$ (cm$^{-2}$)} & EW$_{dry}$ (m\AA ) & EW$_{wet}$ (m\AA ) \cr
819: \tablehead{\colhead{$\sigma_{\rm vac}$} & \colhead{$\lambda_{\rm air}$}
820: & \colhead{EP} & \colhead{gf}
821: & \colhead{S$_{270}$} & \colhead{EW$_{wet}$} \cr
822:
823: \colhead{(cm$^{-1}$)} & \colhead{(\AA )}
824: & \colhead{(cm$^{-1}$)} &
825: & \colhead{(cm$^{-2}$)} & \colhead{(m\AA )}
826: }
827: \startdata
828: 10704.4205 & 9339.33 & 300.36 & 2.730$\times$10$^{-7}$ & 10.7240\tablenotemark{2} & 1047. \cr
829: 11002.2178 & 9086.55 & 285.22 & 1.245$\times$10$^{-8}$ & 0.5303\tablenotemark{2} & 186. \cr
830: 11124.6353 & 8986.56 & 300.36 & 3.017$\times$10$^{-8}$ & 1.1852\tablenotemark{2} & 343. \cr
831: 12254.5139 & 8157.99 & 300.36 & 1.676$\times$10$^{-8}$ & 0.6584\tablenotemark{1} & 208. \cr
832: 11096.9124 & 9009.01 & 285.22 & 4.443$\times$10$^{-9}$ & 0.1892\tablenotemark{2} & 111. \cr
833: 12238.3078 & 8168.79 & 222.05 & 7.226$\times$10$^{-9}$ & 0.4309\tablenotemark{1} & 180. \cr
834: 12062.4127 & 8287.91 & 224.84 & 1.770$\times$10$^{-8}$ & 1.0397\tablenotemark{1} & 313. \cr
835: 12037.5135 & 8305.05 & 325.35 & 2.243$\times$10$^{-8}$ & 0.7714\tablenotemark{1} & 241. \cr
836: \enddata
837: \tablenotetext{1}{S and gf values calculated from Coheur et al. (2002) A values.}
838: \tablenotetext{2}{S and gf values calculated from Brown et al. (2002) A values.}
839: \label{tab-strong}
840: \end{deluxetable}
841:
842: \clearpage
843:
844: \begin{deluxetable}{lrrrl}
845: \tablenum{5}
846: \tablewidth{0pt}
847: \tablecaption{Comparison of calibration at LCO to that at other sites}
848: \tablehead{\colhead{Site} & \colhead{Altitude (km)} & \colhead{$\tau_{dry-air}$} & \colhead{B} & \colhead{Reference}}
849: \startdata
850: LCO Alcaino & 2410 & 0.015$\pm$0.013 & 0.076$\pm$0.005 & this work \cr
851: S. Pole & 2835 & 0.026$\pm$0.001\tablenotemark{*} & 0.083$\pm$0.002\tablenotemark{*} & \citet{cha04} \cr
852: Mauna Kea & 4100 & 0.016 & 0.05 & \citet{dav97} \cr
853: Llanno de Chanjantor & 5104 & 0.0068 & 0.0407 & \citet{del99} \cr
854: \enddata
855: \tablenotetext{*}{uncertainties from \citet{cb95}}
856: \label{tab-cal}
857: \end{deluxetable}
858:
859: \clearpage
860:
861: \begin{figure}
862: \plotone{f1.eps}
863: \caption{Two consecutive orders containing the H{$_2$}O line at 6943.78~\AA ,
864: for a night with 7mm PWV. Boxes indicate the regions shown in detail in the next figure.}
865: \label{fig-bluered}
866: \end{figure}
867:
868: \clearpage
869:
870: \begin{figure}
871: \plotone{f2.eps}
872: \caption{A comparison of the 6943.78~\AA\ H{$_2$}O line in the two consecutive orders for
873: a dry night, with PWV {$\sim$}1.0mm, and a wet night with PWV {$\sim$}7.0mm.}
874: \label{fig-drywet}
875: \end{figure}
876:
877: \clearpage
878:
879: \begin{figure}
880: \plotone{f3.eps}
881: \caption{A plot showing the fractional population of levels (n$_i$/N$_{\rm tot}$)
882: in H$_2$O as a function of temperature, in the range 220--300~K, for level energies
883: of 225, 275 and 300 cm$^{-1}$. Over the range of temperatures seen in the earth's
884: atmosphere the fractional population of levels in this energy level range changes
885: very little, and allows for robust measurement of PWV, without a detailed model atmosphere.}
886: \label{fig-bfac}
887: \end{figure}
888:
889: \clearpage
890:
891: \begin{figure}
892: \plotone{f4.eps}
893: \caption{MSIS-E-90 model atmosphere}
894: \label{fig-atmos}
895: \end{figure}
896:
897: \clearpage
898:
899: \begin{figure}
900: \plotone{f5.eps}
901: \caption{A plot showing the individual PWV results from all lines, compared to
902: the average PWV of each spectrum. The straight line corresponds to a 1:1 relation.
903: Symbols for different line wavelengths are indicated top left.
904: There are no unambiguous systematic trends of individual lines deviating from the mean.}
905: \label{fig-pwv_comp}
906: \end{figure}
907:
908: \clearpage
909:
910: \begin{figure}
911: \plotone{f6.eps}
912: \caption{A plot showing the results from PWV calculations for strong and weak lines
913: on the wettest night (2 August 2005, filled circles) and a moderately dry night
914: (20 August 2005, open boxes). For the wettest night there is a clear down-turn in the PWV
915: estimates above a reduced equivalent width around $-$4.6 to $-$4.7, indicating that lines
916: above that width suffer from saturation. On the drier night only one line was saturated;
917: the excellent agreement between the unsaturated lines shows that the down-turn of the PWV
918: points for the wettest night was not due to systematic error in line $gf$ values.}
919: \label{fig-pwv_sat}
920: \end{figure}
921:
922: \clearpage
923:
924: \begin{figure}
925: \plotone{f7.eps}
926: \caption{The correlation between MIKE echelle PWV Measurements and 225~GHz tipper opacities
927: at LCO during the period between mid-July and mid-Sept 2005. Where no error bars are
928: visible the errors are within the size of the symbol. The line represents a linear least
929: squares fit with characteristics denoted in the legend.}
930: \label{fig-pwv_tau}
931: \end{figure}
932:
933: \clearpage
934:
935: \begin{figure}
936: \plotone{f8.eps}
937: \caption{Clear night-time PWV as a function of time during our campaign. Uncertainties are
938: indicated in gray.}
939: \label{fig-tippwv}
940: \end{figure}
941:
942: \clearpage
943:
944: \begin{figure}
945: \plotone{f9.eps}
946: \caption{Solid line represents the fraction of PWV measurements below a given value. Dotted
947: lines represent the uncertainties due to both the measurements and the calibration. The 10,
948: 25, 50, 75, and 90 percentiles are 1.5, 2.1, 2.8, 3.6, and 4.6 mm, respectively.}
949: \label{fig-tippwvhist}
950: \end{figure}
951:
952: \end{document}
953: