1: \section{Analysis}
2: In this section, we explore the bulk properties of the \loneosI RR
3: Lyrae star sample that was described in the previous section. First,
4: we present the spatial distribution of the entire sample. Next, we
5: present the period-amplitude distribution of the sample. Finally, we
6: explore the spatial distributions of the different populations that
7: emerge from the period-amplitude distribution.
8:
9: \subsection{Radial Density Distribution}
10: As described in previous sections, we tie our absolute photometry to
11: synthetic SDSS-based magnitudes ($V_{SDSS}$), $M_{LONEOS \ V}$. This
12: allows for a straightforward calculation of distances from the
13: intensity-weighted mean magnitudes. Distances are calculated from the
14: following formula:
15: \begin{equation}
16: d = 10 ^{\frac{ <M_{LONEOS \ V}> - M_{V} - 10 }{5}} \ kpc,
17: \end{equation}
18: We will use RR Lyrae absolute magnitudes of $M_{V} = +0.71 \pm 0.12$
19: \citep{Layden96} derived from statistical parallax analysis for a
20: sample with $\langle[Fe/H]\rangle = -1.61$.
21: %, and $M_{V} = +0.55 \pm 0.1$ \citep{Demarque2000} derived from theoretical
22: % synthetic horizontal-branch models for $\langle[Fe/H]\rangle =
23: % -1.61$\footnote{$M_{V} = 0.21[Fe/H] + 0.89$}.
24: Assuming that the Sun is at $-7.8$ kpc \citep{Carney_Galactic_Center}
25: from the galactic center along the y-axis ($x_{gal},y_{gal},z_{gal} =
26: (0, -7.8, 0)$ kpc), we calculate the galactocentric 3-D coordinates of
27: each RR Lyrae star.
28:
29: We measure the spherically symmetric radial density distribution by
30: counting the number of RR Lyrae stars in galactocentric radial bins
31: and calculating the survey volume contained in each bin. We use 1 kpc
32: bins for all the density distributions. This provides us with a
33: measure of the radial density distribution averaged along the survey
34: lines-of-sight. Using the boundaries of the survey we numerically
35: integrate the volume contained in each radial bin, taking into account
36: detection efficiency at each position in the Monte Carlo integration.
37: Figure \ref{fig:volume} shows the survey effective volume
38: distribution.
39: %
40: \begin{figure}
41: \epsscale{1.0}
42: \centering
43: \ifsubmode
44: \plotone{f16.eps}
45: \fi
46: \caption[Raw and Detection Efficiency-Corrected Survey Volume]{This figure
47: shows the raw (circles) and the detection
48: efficiency-corrected (triangles) survey volumes. Note that the difference
49: between the two curves is not monotonically increasing
50: because there are lines-of-sight which are pointed towards the
51: Galactic Center and also the anti-Center which contribute at
52: different ranges of Galactocentric radius. If the volume calculation
53: assumes that the Sun is at the center of the Galaxy, the
54: difference between these two curves does indeed increase monotonically. }
55: \label{fig:volume}
56: \end{figure}
57:
58:
59: Assuming a spherically symmetric distribution, Figure
60: \ref{fig:radial_all} shows the radial profile for our sample with mean
61: magnitudes in the range [13,17.5] with $M_{V} = +0.71$. The one-sigma,
62: asymmetric Poisson error bars are assigned and calculated using the
63: method of \citet{Gehrels}. Using non-linear least squares, we fit
64: power laws to the radial density distributions:
65: \begin{equation}
66: \rho(R_{gal}) = (397 \pm 56) \times R_{gal}^{(-2.43 \pm 0.06)} \
67: (kpc^{-3}).
68: \end{equation}
69: with \Chisqr per degree of freedom of 0.91 . All fitted
70: parameter estimates are calculated using a Monte Carlo
71: bootstrap technique \citep{NR}.
72: Table \ref{table:other_surveys} summaries results from other surveys. This
73: result is within $1-\sigma$ of the QUEST-I (\cite{Vivas2006}, $M_{V} =
74: +0.55$, $R_{SUN} = 8.0$ kpc) and the SDSS (\cite{Ivezic00}, $M_{V} =
75: +0.7$, $R_{SUN} = 8.0$ kpc) results, assuming a spherical halo
76: distribution. The power law exponent is slightly sensitive to the
77: distance to the Galactic Center. If $R_{SUN} = 8.5$ kpc, then
78: exponent is reduced to $-2.50 \pm 0.05$. As shown in Table
79: \ref{table:other_surveys}, the measured distance to the Galactic
80: Center has been decreasing with time. This means that earlier
81: measurements will tend to have steeper profiles than our measurement.
82: %
83: \begin{figure}
84: \epsscale{1.0}
85: \centering
86: \ifsubmode
87: \plotone{f17.eps}
88: \fi
89: \caption[\loneosI RR Lyrae Radial Distribution ($M_{V} = +0.71$)]{The radial
90: density distribution of 838 {\it ab}-type
91: RR Lyrae with a mean magnitude range
92: from 13.0 to 17.5, using $M_{V} = +0.71$. We do not find any
93: statistically significant
94: deviations from a smooth stellar halo. Our survey volume does not
95: intersect known
96: locations of the tidal debris from the Sagittarius dwarf galaxy. }
97: \label{fig:radial_all}
98: \end{figure}
99: %
100: \tabletypesize{\tiny}
101: \begin{deluxetable*}{lcccccc}
102: %\tablecolumns{7}
103: %\tablewidth{0pc}
104: %\rotate
105: \setlength{\tabcolsep}{0.02in}
106: \tablecaption{Halo RR Lyrae Survey Comparisons
107: \label{table:other_surveys}}
108: \tablehead{
109: \colhead{Survey} & \colhead{Tracer} & \colhead{Exponent} &
110: \colhead{Flattening Ratio (q)} & \colhead{Objects} & \colhead{$R_{0}$ (kpc)} & \colhead{Comment}\\ }
111: \startdata
112: \loneosI (All) & RR Lyrae & $-2.43 \pm 0.06$ & $q = 1$ & 838 & 7.6 & $M_{V} = +0.71$ \\
113: \loneosI (All) & RR Lyrae & $-3.15 \pm 0.07$ & Variable q & 838 & 7.6 & $M_{V} = +0.71$ \\
114: \loneosI (Oo I) & RR Lyrae & $ -2.26 \pm 0.07 $ & $q = 1$ & 619 & 7.6 & $M_{V} = +0.71$ \\
115: \loneosI (Oo II) & RR Lyrae & $ -2.88 \pm 0.11 $ & $q = 1$ & 219 & 7.6 & $M_{V} = +0.71$ \\
116: QUEST-I (\cite{Vivas2006}) & RR Lyrae & $-2.5 \pm 0.1$ & $q = 1$ & 395 & 8.0 & $M_{V} = +0.55$ \\
117: QUEST-I (\cite{Vivas2006}) & RR Lyrae & $-3.1 \pm 0.1$ & Variable q & 395 & 8.0 & $M_{V} = +0.55$ \\
118: SDSS (\cite{Ivezic00}) & RR Lyrae & $-2.7 \pm 0.2$ & $q = 1$ & 148 & 8.0 & $M_{V} = +0.7$ \\
119: \cite{Chiba_Beers00} & Halo ($\langle[Fe/H]\rangle \leq -1.8$) & $-3.55 \pm 0.13$ & Variable q ($q_{0} \sim 0.65$) & 413 & \nodata & Kinematics \\
120: \cite{Chiba_Beers00} & Halo ($-1.6 < \langle[Fe/H]\rangle \leq -1.0$) & $-3.47 \pm 0.18$ & Variable q ($q_{0} \sim 0.55$) & 331 & \nodata & Kinematics \\
121: % \cite{Robin00} & Population Synthesis & $-2.44$ & $q = 0.76$ & & & \\
122: \cite{Wetterer96} & RR Lyrae & $-3.0 \pm 0.08$ & $q = 1$ & 42 & 7.6 & $M_{V} = +0.74$ \\
123: \cite{Wetterer96} & RR Lyrae & $-3.5 \pm 0.08$ & Variable q & 42 & 7.6 & $M_{V} = +0.74$ \\
124: \cite{Preston91} & RR Lyrae & $-3.2 \pm 0.1$ & Variable q & 47 & 8.0 & $M_{V} = +0.6$ \\
125: \cite{Saha85} & RR Lyrae & $-3.0 $ & $q = 1$ & 29 & 8.7 & P-L-A relation ($M_{B} = M_{bol} +0.34)$ \\
126: \cite{Zinn1985} & Globular Clusters & $-3.5$ & \nodata & 121 & \nodata & \nodata \\
127: \cite{Hawkins84} & RR Lyrae & $-3.1 \pm 0.2$ & $q < 0.9$ & 24 & 8.7 & $M_{B} = +0.9$ \\
128: \cite{Oort75} & RR Lyrae & $-3.0 $ & $q < 0.8$ & 1108 & 8.7 & $M_{B} = +0.9$ \\
129: \cite{Kinman66} & RR Lyrae & $-3.5 $ & q= 0.6 - 0.8 & 38 & 10.0 & $M_{V} = +0.42$ \\
130: \enddata
131: \tablecomments{
132: We compare the \loneosI RR Lyrae survey results to those of other surveys using RR Lyrae stars and other tracers.
133: }
134: \end{deluxetable*}
135:
136:
137: The shapes of the stellar and dark matter halos should have encoded in
138: them some information of their formation process. In general, the
139: isodensity contours for the RR Lyrae number density can be
140: described by an ellipsoid:
141: \begin{equation}
142: \frac{x^{2}}{a^{2}} + \frac{y^{2}}{b^{2}} + \frac{z^{2}}{c^{2}} = 1.
143: \end{equation}
144: Let us consider the case where $a=b$ (i.e., a spheroid) and define the
145: flattening ratio as, $q \equiv c/a$. Thus, the isodensity contours can
146: be described as:
147: \begin{equation}
148: x^{2} + y^{2} + \frac{z^{2}}{q^{2}} = a^{2},
149: \end{equation}
150: where $q < 1$ and $q > 1$ are oblate and prolate spheroids,
151: respectively. It is widely observed that the inner regions ($R_{gal}
152: = 5-10 kpc$) of the stellar halo are highly flattened ($q \sim 0.5$),
153: while the outer regions are more spherically symmetric
154: \citep[e.g.,][]{Kinman66,Preston91,Chiba_Beers00}. We fit
155: spheroids with various constant flattening to the \loneosI RR Lyrae.
156: In addition, we also use a model with a spatial variable flattening
157: ratio \citep{Preston91}, where q varies as function of semi-major
158: axis, a:
159: $$ q = \left\{
160: \begin{array}{cc}
161: q_{0} + [1 - q_{0}]\frac{a}{20} &\mbox{ if } a < 20 \ kpc \\ 1 &\mbox{
162: if } a > 20 \ kpc
163: \end{array}\right.
164: $$ where $q_{0} = 0.5$ and a is the semi-major axis. In this variable
165: q model, the flattening smoothly varies from 0.5 to 1.0 over the
166: course of 20 kpc.
167: Figure \ref{fig:spherical_and_variable_q_power_laws} shows
168: the radial profiles and power law fits
169: for both the spherical and the variable flattening models.
170: The power law exponents, \Chisqr, degrees of freedom, and
171: \Chisqr probabilities are listed in Table \ref{table:flattening}.
172: The degrees of freedom are not constant since the data are binned
173: and the value of $a$ depends on $q$.
174: All the flattening models have
175: larger \Chisqr probability than the spherically
176: symmetric model indicating that flattened models are favored,
177: but there is no clear minimum in the \Chisqr probabilities.
178: The variable q model yields a steeper density profile with a power law
179: exponent of $-3.15 \pm 0.07$ with a \Chisqr per degree of freedom of
180: 0.7. This agrees with most other RR Lyrae surveys (Table
181: \ref{table:other_surveys}).
182: %
183: \begin{figure}
184: \epsscale{1.0}
185: \centering
186: \ifsubmode
187: \plotone{f18.eps}
188: \fi
189: \caption[Radial density profiles for spherical and variable flattening models]{
190: Radial density profiles for spherical and variable flattening models
191: are shown in the left and right columns, respectively. The variable
192: flattening model is that of \citet{Preston91}, where the flattening
193: ratio variable from 0.5 to 1.0 with increasing distance from the
194: Galactic Center. The top row of plots are for the entire sample of RR
195: Lyrae stars. The second and third rows are for sub-samples of RR
196: Lyrae, classified as Oosterhoff I and II, respectively. The \loneosI
197: RR Lyrae data are fit by the spherical and variable flattening models
198: equally well. The fits are not statistically different.
199: }
200: \label{fig:spherical_and_variable_q_power_laws}
201: \end{figure}
202: %
203: \begin{deluxetable}{lcccccc}
204: \tablecolumns{7}
205: \tablewidth{0pc}
206: \tablecaption{Flattening Fits
207: \label{table:flattening}}
208: \tablehead{
209: \colhead{Sample} & \colhead{Flattening} & \colhead{Exponent} &
210: \colhead{\Chisqr} & \colhead{Degrees of} &
211: \colhead{Probability} \\
212: & \colhead{Ratio (q)} & & & \colhead{Freedom ($\nu$)} &
213: \colhead{(Q(\Chisqr,$\nu$))}
214: }
215: \startdata
216: All & 1.0 & -2.43 & 24.7 & 24 & 0.42 \\
217: All & Variable q & -3.15 & 18.3 & 23 & 0.74 \\
218: All & 0.9 & -2.48 & 16.1 & 24 & 0.88 \\
219: All & 0.8 & -2.53 & 17.7 & 24 & 0.82 \\
220: All & 0.7 & -2.57 & 14.4 & 23 & 0.91 \\
221: All & 0.6 & -2.58 & 19.8 & 23 & 0.65 \\
222: All & 0.5 & -2.57 & 13.4 & 22 & 0.92 \\
223: All & 0.4 & -2.57 & 15.4 & 22 & 0.84 \\
224: All & 1.1 & -2.35 & 17.5 & 25 & 0.86 \\
225: All & 1.2 & -2.29 & 19.1 & 25 & 0.79 \\
226:
227:
228: \\[-1.4ex] \hline \hline \\[-1.4ex]
229:
230: OoI & 1.0 & -2.26 & 18.9 & 24 & 0.76 \\
231: OoI & Variable q & -2.93 & 16.0 & 23 & 0.86 \\
232: OoI & 0.9 & -2.29 & 13.8 & 24 & 0.95 \\
233: OoI & 0.8 & -2.36 & 14.8 & 24 & 0.93 \\
234: OoI & 0.7 & -2.40 & 12.6 & 23 & 0.96 \\
235: OoI & 0.6 & -2.42 & 19.8 & 23 & 0.65 \\
236: OoI & 0.5 & -2.41 & 10.5 & 22 & 0.98 \\
237: OoI & 0.4 & -2.41 & 10.9 & 22 & 0.98 \\
238: OoI & 1.1 & -2.18 & 13.3 & 25 & 0.97 \\
239: OoI & 1.2 & -2.15 & 15.9 & 25 & 0.92 \\
240:
241: \\[-1.4ex] \hline \hline \\[-1.4ex]
242:
243: OoII & 1.0 & -2.88 & 7.2 & 24 & 0.9996 \\
244: OoII & Variable q & -3.70 & 5.9 & 23 & 0.9998 \\
245: OoII & 0.9 & -2.97 & 5.6 & 24 & 0.9999 \\
246: OoII & 0.8 & -3.02 & 5.7 & 24 & 0.9999 \\
247: OoII & 0.7 & -3.09 & 7.5 & 23 & 0.9990 \\
248: OoII & 0.6 & -3.07 & 5.8 & 23 & 0.9998 \\
249: OoII & 0.5 & -3.01 & 9.6 & 22 & 0.9895 \\
250: OoII & 0.4 & -3.04 & 13.0 & 22 & 0.9331 \\
251: OoII & 1.1 & -2.80 & 10.0 & 25 & 0.9966 \\
252: OoII & 1.2 & -2.71 & 11.8 & 25 & 0.9881 \\
253:
254: \enddata
255: \tablecomments{
256: Power law fits with varying flattening models.
257: }
258: \end{deluxetable}
259:
260:
261: The poor constraint on the stellar halo flattening from the \loneosI
262: RR Lyrae is likely a result of the fact that the Galactic latitude
263: distribution of these RR Lyrae are
264: highly concentrated at mid-range Galactic latitude
265: with no fields near the Galactic Poles.
266: The difference between the spherical and
267: flattened model is smallest at lower
268: Galactic latitudes.
269:
270:
271: All the recent RR Lyrae surveys are in disagreement with
272: \citet{Chiba_Beers00} who used $\sim 1200$ non-kinematically selected
273: metal poor stars. One reason for the difference could be the method
274: used. RR Lyrae surveys use 3D Galactocentric coordinates to measure
275: densities. In these surveys, the completeness and survey volume are
276: relatively well-defined. However, \citet{Chiba_Beers00} use data
277: derived from many different surveys and is not a volume-limited
278: sample. They use a maximum likelihood method which relies on
279: kinematic information and Jeans' theorem to measure the power
280: law. Since the completeness and survey volumes are not easily
281: understood, it is not possible to measure the density distribution
282: using a binning technique. Thus, it is difficult to understand if
283: there is a real difference between RR Lyrae stars and generic
284: metal-poor stars. In addition, RR Lyrae stars and the metal poor halo
285: stars from \citet{Chiba_Beers00} may possibly be tracing different
286: populations. RR Lyrae abundances are influenced by the horizontal
287: branch morphology of the RR Lyrae progenitor system. Thus, they may
288: not reveal the entire history of Galactic formation and could have
289: different spatial distributions than a sample of metal poor halo stars.
290:
291: \subsection{Period-Amplitude Distribution}
292: The period-amplitude distribution of RR Lyrae stars can be used to
293: explore the formation history of the Galactic halo
294: \citep[e.g.,][]{Catelan2006}. If some fraction or all of the stellar halo was
295: formed from the accretion of dwarf galaxies, which might have very
296: diverse chemical histories and ages (i.e., different horizontal branch
297: morphologies), the signatures of these accretion events should be
298: imprinted in the period-amplitude distribution. Thus, accretion
299: activity can be probed even after phase-space has been sufficiently
300: mixed that spatial coherence is lost.
301:
302:
303: Our large sample of RR Lyrae stars allows us to further investigate
304: this possibility and to explore differences in the two components. In
305: lower-right panel of Figure \ref{fig:AFTER_per_amp_cuts},
306: we plot the period-amplitude
307: distribution of 838 ab-type RR Lyrae stars. Since LONEOS uses an
308: unfiltered CCD system, whose sensitivity peaks in the red, amplitudes
309: for our RR Lyrae will be systematically smaller than those taken in
310: the V-band. The observed amplitude of RR Lyrae increases as you
311: observe them in bluer passbands.\footnote{Note that the transformation
312: to the $V_{SDSS}$ magnitude system does not affect the measured
313: \loneosI amplitudes. The amplitudes are strictly a function of the
314: passband used.} Using a subset of \loneosI RR Lyrae (12 stars) which
315: have previously measured V-band amplitudes taken from SIMBAD database,
316: we have empirically derived a scaling relation between the LONEOS
317: amplitudes ($A_{LONEOS}$) and V-band amplitudes ($A^{V}$) and scaled
318: the M3 RR Lyrae period-amplitude relation to the \loneosI amplitude
319: scale. The mean value of $A_{LONEOS}/A^{V}$ is $0.7$ with a standard
320: deviation of $0.1$. Figure \ref{fig:loneos_per_amp_M3}
321: shows that data from the lower-right panel in
322: Figure \ref{fig:AFTER_per_amp_cuts}
323: overlaid with the linear period-amplitude
324: relation for M3 (the quintessential OoI globular cluster)
325: \citep{Cacciari2005} corrected to the \loneosI amplitude scale.
326: Despite only having 12 stars to scale
327: LONEOS amplitudes to V-band amplitudes, the
328: main locus agrees well with the period-amplitude relation of
329: M3. Thus, we associate the main locus with the OoI component of the
330: stellar halo. In addition, Figure \ref{fig:loneos_per_amp_M3} shows a
331: striking concentration of RR Lyrae (26\% of the total sample) with
332: longer periods offset from the main locus of points.
333: %
334: \begin{figure}
335: \epsscale{1.0}
336: \centering
337: \ifsubmode
338: \plotone{f19.eps}
339: \fi
340: \caption[\loneosI period-amplitude distribution with M3 relation]{
341: The LONEOS RR Lyrae period-amplitude distribution (same as Figure \ref{fig:AFTER_per_amp_cuts})
342: with the linear log(period)-amplitude relation for globular cluster M3 (OoI)
343: superimposed (solid line). Note the amplitudes are based on the
344: unfiltered response of the \loneosI camera system. The M3 relation has
345: been converted to the \loneosI amplitude scale. We derived a
346: conversion between \loneosI amplitude and V-band amplitudes from
347: \citet{Cacciari2005}. The dashed line indicates where we divide our
348: sample into OoI and OoII components.}
349: \label{fig:loneos_per_amp_M3}
350: \end{figure}
351:
352: We fit a line to the primary locus ($P_{primary}(A)$) and measure the
353: period shift,
354: \begin{equation}
355: \Delta\ Log(Period) = P_{RR}(A) - P_{primary}(A),
356: \end{equation}
357: which refers to the Log(Period) distance (at fixed amplitude) of {\em
358: each} RR Lyra ($P_{RR}(A)$) from this best fit line ($A = -4.76 \times
359: Log(P_{primary}) - 0.51$). Figure \ref{fig:period_shift} shows the
360: distribution of the period shift at fixed amplitude from the main
361: locus. This figure clearly shows the existence of a secondary locus.
362: It is shifted from the main locus by $\Delta Log(P) \sim
363: 0.075$\footnote{Periods are measured in days.}, which is approximately
364: the shift between RR Lyrae in OoI to OoII globular clusters
365: \citep[e.g.,][]{Sandage1981,Clement1999}. Thus, we associate the
366: main locus of points with an OoI halo component, and the shifted locus
367: with an OoII halo component.
368: %
369: \begin{figure}
370: \epsscale{1.0}
371: \centering
372: \ifsubmode
373: \plotone{f20.eps}
374: \fi
375: \caption[\loneosI Period Shift]{Histogram of the period shift of
376: RR Lyrae relative to the primary locus (OoI). We fit a line to the
377: primary locus and measure $\Delta\ Log(Period)$ which refers to the
378: Log(Period) distance (at fixed amplitude) of {\em each} RR Lyrae star
379: from this best fit line. Thus, the primary locus is centered at
380: $\Delta\ Log(Period) = 0.0$. The location of the secondary locus is
381: located at approximately the location where we would expect the OoII
382: locus \citep{Sandage1981,Cacciari2005}. For the subsequent analysis,
383: we divide our sample at $\Delta\ Log(Period) = 0.045$, where period is
384: measured in days. }
385: \label{fig:period_shift}
386: \end{figure}
387:
388:
389: In Figure \ref{fig:quest_period_shift}, we plot the period shift distribution
390: of 395 RR Lyrae from the QUEST-I catalog \citep{QUEST04}
391: with mean magnitudes in the range [13.5, 19.7] as well as
392: the LONEOS-I period-shift distribution. This QUEST-I distribution
393: also contains a sizable fraction ($\sim 20\%$)
394: of longer period RR Lyrae than the main OoI locus.
395: \cite{Catelan2006} has pointed out that the Oosterhoff
396: classification the QUEST-I RR Lyrae may be unreliable due to their
397: sparse temporal sampling. This could cause a smearing of the period shift
398: distribution in the QUEST-I data.
399: %
400: \begin{figure}
401: \epsscale{1.0}
402: \centering
403: \ifsubmode
404: \plotone{f21.eps}
405: \fi
406: \caption[Period shifts of the QUEST-I RR Lyrae Survey]{Period shifts
407: (with respect to M3) of the QUEST-I RR Lyrae Survey is shown as a solid
408: line, while the LONEOS-I period shifts are shown as a dashed line.
409: In contrast to the LONEOS-I sample, there is no clearly displaced
410: secondary peak associated with an OoII component for the QUEST-I sample.
411: However, there is a long period tail.}
412: \label{fig:quest_period_shift}
413: \end{figure}
414:
415:
416:
417:
418: It has been well established that OoI and OoII globular clusters show
419: spatial and kinematic differences, which may signify different origins
420: \citep{Lee99}. It is natural to ask if the OoI and OoII components of
421: the stellar halo show similar differences. We divided our sample into
422: the two components described above. The division was positioned at the
423: minimum of the period shift distribution, $\Delta(P) = 0.045$. The
424: division is denoted by the dashed line in Figure
425: \ref{fig:loneos_per_amp_M3}.
426: %The distribution of Galactocentric radii of each component
427: %are statistically different based on a Kolmogorov-Smirnov (K-S) test
428: %(Figure \ref{fig:cumulative_r_gal_histograms}). We see that the OoI
429: %distribution peaks before the OoII component and there is a sudden
430: %break in slope at approximately 18 kpc, which corresponds to the
431: %turnaround of the survey volume distribution. One the other hand, the
432: %OoII component peaks after the OoI component and shows a significant
433: %drop in number at 15 kpc, but the volume is still increasing at this
434: %Galactocentric radius.
435: The radial number density profiles for each component are shown in
436: Figure \ref{fig:radial_above_below}. From these
437: profiles, we derive the following radial dependences:
438: \begin{equation}
439: \rho_{OoI} \sim R^{-2.26 \pm 0.07},
440: \end{equation}
441:
442: \begin{equation}
443: \rho_{OoII} \sim R^{-2.88 \pm 0.11}.
444: \end{equation}
445: The OoII component has a steeper profile than the OoI component. This
446: is consistent with \citet{Lee99} who, using a smaller RR Lyrae sample,
447: found that the OoI component dominated farther from the Galactic
448: plane, possibly indicating that the OoII component falls off more
449: rapidly than the OoI component. In the following sections, we will
450: examine whether this difference could be caused by some systematic
451: effect in the detection of the two populations.
452: %
453: \begin{figure}
454: \epsscale{1.0}
455: \centering
456: \ifsubmode
457: \plotone{f22.eps}
458: \fi
459: \caption[OoI and OoI Radial Distributions]{The spherically symmetric (q=1)
460: radial distribution for the OoI and OoII components.
461: The bins with zero density are included in the fit; the $1\sigma$
462: Poisson upper limits are calculated using the method of \citet{Gehrels}.
463: }
464: \label{fig:radial_above_below}
465: \end{figure}
466:
467: The possibility of an intermediate Oosterhoff halo component still
468: exists, but our current sample cannot address this. It is critical to
469: effectively identify and veto RR Lyrae which exhibit the Blazhko
470: effect (i.e. amplitude and phase variations). These stars will add
471: scatter to the period-amplitude relation \citep{Cacciari2005}. In
472: addition, we have found that the \loneosI measured RR Lyrae amplitudes
473: tend to be underestimated (Figure \ref{fig:amp_errors}). This effect
474: would tend to diminish the number of RR Lyrae with intermediate Oo
475: classification.
476: %\loneosII data will provide better temporal light
477: %curve coverage which should enable us to veto Blazhko stars and
478: %better estimate amplitudes.
479:
480: Additionally, microlensing surveys (e.g., MACHO, OGLE) have found that
481: as much as a third of RR Lyrae stars in the LMC and SMC exhibit the
482: Blazhko effect \citep[e.g.,][]{MACHO2000}. In order to address the
483: possibilities of an intermediate Oo component, we need better temporal
484: coverage \citep[e.g.,][]{Catelan2006}. A mixture of OoI and OoII could
485: masquerade as an intermediate.
486: %The data from the \loneosII camera
487: %will produce cleaner period-amplitude distributions. This will be
488: %enable us to search for additional ``fine structure'' in the
489: %period-amplitude distribution.
490:
491: Is the difference in radial distributions associated with the
492: Oosterhoff effect? Does the Oosterhoff effect manifest itself in such
493: a way? Previously, we have chosen to divide our full sample into two
494: components that resemble OoI and OoII populations; this maximizes the
495: role of the Oosterhoff effect on the sample. Instead, we now divide
496: our entire sample along a line in period-amplitude space perpendicular
497: to Oosterhoff dividing line, minimizing the role of the Oosterhoff
498: effect. The resulting power law exponents ($\alpha$) are consistent
499: with each other ($\alpha = -2.50 \pm 0.07$ and $\alpha = -2.39 \pm
500: 0.1$). This indicates that the Oosterhoff effect is likely associated
501: with differing radial distributions.
502: