1. Introduction
The
University of California and the California Institute of Technology have
initiated and are undertaking the California Extremely Large Telescope (CELT)
project which aims to build a 30-m diameter ground-based telescope for
astronomy at visible and infrared wavelengths (Nelson and Mast, 1999; Nelson,
2000). In support of this project, a survey of the southwestern USA and
northern Mexico has been carried out to identify possible telescope sites and
to assess their quality in terms of critical astronomical observing quality
indicators. In a study conducted for Cerro Tololo Inter-American Observatory
(CTIO) and the University of Tokyo to evaluate sites in northern Chile, cloud
cover and water vapor parameters derived from meteorological satellite data
were used to quantify observing conditions (Erasmus and van Staden, 2001).
A
similar approach has been taken in this study of the southwestern U.S.A and
northern Mexico. Firstly, the aerial distribution of cloud cover and water
vapor over the area of interest was mapped quantitatively so that preferred
areas could be identified. Based on this analysis, fifteen sites were selected
for further analysis and comparison. The 15 sites include the best potential
telescope sites and a number of existing observatory sites. Some existing
observatories have independent records and measurements of site quality. These
have been related to the satellite observations in order to validate the
evaluation and ranking of site quality based on the satellite data analysis.
The study area
covers 18oN to 40oN and 96oW to 124oW
(Figure 1). It has numerous mountain peaks with altitude, shape and orientation
potentially suitable for locating a large telescope. The area includes several
developed telescope sites such as Mt. Graham, Kitt Peak, Palomar Mountain and
San Pedro Martir.
Given
the size of the area and the large number of possible sites to be considered, a
survey method that can map cloud cover and water vapor quantitatively over the
entire area leading to the identification of a limited number of good sites is
needed. Additionally, the individual sites thus identified need to be compared
and ranked objectively in terms of their quality. The data set must also be
long enough to describe the climatology of the area (and the sites) and allow
for the assessment of diurnal, seasonal and inter-annual variability.
Meteorological
satellites monitor cloud cover and water vapor over large areas with a
temporal, spatial and radiometric resolution that is suitable for such a survey
(Figure 2). An important feature of satellite data is that these provide a
consistent measurement over a field of view so that a reliable comparison of
sub-regions can be made. The spatial (~10km) and temporal (3 hours) resolution
of the observations also ensures that measurements are representative of
conditions at particular sites and that the diurnal moisture and cloud cover cycle
is resolved. Satellite data archives now consist of directly comparable
measurements (consistent from one satellite to another) over extended periods
(more than five years). This allows for an evaluation of observing conditions
over a reasonably long time base-line. By referencing the period of satellite
data against other climate data with a longer time base-line, an assessment can
be made of the climatological representativeness of the study period.
Figure 2. Full earth-disk view of the GOES-8 satellite
at 11:45UT on October 25, 2000. Left, infra-red window channel (10.7µm) and right, water vapor channel (6.7µm).
This
report presents the results of the survey described above. An overview of the
meteorology and climatology of the study area and data period is provided in
section 2. Using basic topographic and climatological criteria, a preliminary
list of sites was compiled. Section 3 contains this information. Details on the
satellite and other data used in the study are provided in section 4. The
methodology used to derive cloud cover and water vapor parameters from the
satellite data is outlined in section 5. The analysis performed and results
obtained are presented in section 6. This includes an assessment of how cloud
cover and water vapor conditions in the study period compare with long-term
means, trends and variability (sections 6.1 and 6.2), quantitative mapping of
cloud cover and water vapor conditions over the study area (section 6.3), site
selection (section 6.4) and the objective comparison of cloud cover and water
vapor at the 15 selected sites (sections 6.5 and 6.6).
2. The
Meteorology and Climatology of the Study Area
The
southwestern U.S.A. and northern Mexico are located in the latitude region
where air, lifted into the upper troposphere along the inter-tropical
convergence zone and transported polewards, descends again to the surface. This
descending air produces a semi-permanent region of high pressure between about
25oN and 35oN. The descending air is dry since the
moisture it carried aloft over the equator has precipitated. As it descends, it
is also compressed and warmed adiabatically. Consequently, the sub-tropical
high pressure area is generally cloud free. At some height above the surface
the descending air encounters air near the surface which is cooler. At this
boundary, the normally observed decrease in temperature with height in the
atmosphere is reversed. The temperature inversion layer, due to its thermal
stability characteristics, prevents air near the surface moving upwards beyond
the inversion base. Moisture and dust are therefore trapped below the inversion
layer. The height of the inversion layer depends largely on the temperature of
the surface. A warm surface promotes the formation of convective eddies above the
ground, particularly in the day time, and this lifts the inversion height. At
inland locations, therefore, the inversion height tends to be elevated in
summer. At coastal locations the cool ocean surface has a moderating effect in
summer with the result that the inversion height exhibits small diurnal and
seasonal fluctuations.
The
region of high pressure in the sub-tropics moves north in the summer and south
in the winter. In winter, this southward movement allows mid-latitude cyclones
moving from west to east over the Pacific Ocean to take a more southerly track.
This causes the northern part of the study area to experience moist and cloudy
conditions in the winter. In summer, as the center of the subtropical high
pressure migrates northwards, the southern part of the study area comes under
tropical influence. In the extreme south this would include Tropical storms and
cyclones. Areas east of the continental divide are influenced by moisture that
is circulated by the Bermuda High off the Gulf of Mexico. Moisture from this
source is the fuel for much of the summer thunderstorm activity in these areas.
Locations further west come under the influence of the southwest Monsoon which
transports moisture from the Pacific over the southwestern U.S.A. and parts of
Mexico.
Inter-annual
variations in moisture and cloud cover are also observed over the region. This
variability is controlled to a significant degree by El Niño - Southern
Oscillation (ENSO) events. ENSO is a non-periodic inter-annual oscillation in
the atmospheric and oceanic conditions that occurs over the tropical Pacific
Ocean. Under normal or typical conditions, the weather of the southeastern
Pacific is dominated by the strong, semi-permanent south Pacific Anticyclone
(high pressure) which gives rise to persistent easterly winds along the Equator
between the Americas and the dateline. Under these conditions, strong upwelling
is encouraged in the ocean off the west coast of South America producing cold
sea surface temperatures. The cold water is transported into the tropics by the
combined effect of ocean currents and surface drag exerted by the trade winds.
This cold water reduces evaporation rates and also stabilizes the lowest layers
of the atmosphere in the Tropical eastern Pacific Ocean. In the north Pacific,
the Polar Jet Stream maintains its integrity thus steering winter-time cyclonic
storms along a favored track centered between northern California and the
Canadian border. The cold phase of ENSO known as La Niña is simply an
intensification of these “normal” conditions.
Every
two to seven years, with irregular periodicity, an anomalous warming of the sea
surface occurs in the Tropical eastern Pacific Ocean. This warming coincides with
a decrease in the strength of the south Pacific high and consequently a
decrease in the strength of trade winds and also in the amount of upwelling.
This anomaly, which usually lasts from 12 to 18 months has become known as El
Niño (the warm phase of ENSO) and directly affects the whole of the tropical
and equatorial south Pacific Ocean. Indirect effects are felt over the entire
Pacific basin including the southwestern U.S.A and northern Mexico. In the
northern Pacific the Polar jet stream weakens and takes a more northerly track
while the Subtropical jet strengthens. This tends to steer winter storms along
two tracks - one northerly track into Canada and one southerly track into
central and southern California. This enhances winter-time moist and cloudy
conditions in southern California, northern Mexico, Arizona, New Mexico and, to
a lesser extent, Nevada, Colorado and Utah. Conversely, more northerly
locations (Oregon, Washington, Idaho and Wyoming) experience a drier winter
(Figure 3).
The effect of
El Niño on Summer-time moisture and cloud is less clear. The strong El Niños of
1982 and 1991 produced more humid summers in Arizona. During some El Niño years
the onset of the Summer Monsoon was delayed and summers were drier than normal.
According to Reynolds et al (1999),
in the Great Basin area of the western United States, above normal precipitation
was recorded during El Niño years in 81% of the cases for the period April to
October. Significant droughts in the Great Basin such as that of 1989 coincided
with a La Niña event.
Given
the above information on climate variability within the study area, it is
important to characterize the long-term representativeness of the study period.
An analysis of the climate during the study period is presented in section 6.2.
A summary of ENSO events between 1993 and 1999, the period of satellite data
coverage, is presented here. Table 1 and Figure 4 show that ENSO events ranged
from strong El Niño to strong La Niña conditions during the study period. The
average SOI for the study period is -0.6 which is near normal but slightly
biased towards the warm phase of ENSO. Based on the discussion above, this
consideration of ENSO conditions suggests that moisture conditions over the
study area as a whole would be near normal or slightly more moist than normal
during the study period. A more detailed assessment may be found in section
6.2.
Table
1. Summary of ENSO conditions for the study period, July 1993 to September
1999.
|
PERIOD |
ENSO
CONDITIONS |
|
January 1993
– November 1993 |
Weak El Niño |
|
December
1993 – February 1994 |
Normal |
|
March 1994 –
November 1994 |
Weak El Niño |
|
December
1994 – November 1995 |
Normal |
|
December
1995 – February 1997 |
La Niña |
|
March 1997 –
April 1998 |
Strong El
Niño |
|
May 1998 –
April 1999 |
Strong La
Niña |
|
May 1999 –
September 1999 |
Normal |
Figure 4.
Southern Oscillation Index (SOI), pressure anomaly at Tahiti minus pressure
anomaly at Darwin, normalized and Niño region 1+2 (0oS – 10oS,
80oW – 90oW) sea surface temperature anomaly (oC)
for the period 1980-2000. Positive SOI values indicate a La Niña and negative
values an El Niño. Source: Climate Prediction Center, NOAA, (http://www.
elnino.noaa.gov/).
3. Preliminary Site
Selection
As a starting
point for site selection, a list of sites was compiled based on precipitation
climatology and topography (primarily suitable site elevation). It is
important to note that the preliminary list was intended as a point of
departure in the site selection process. Other sites were added as these became
known or were later identified following the area analysis.
Figures 5 and
6 show the mean annual precipitation distribution maps for the SW USA and
Mexico. The minimum precipitation area extends from about 37oS at
the California-Nevada border, more-or-less southwards to the Mexico border. In
northern Mexico the dry zone extends into Baja and Sonora. Precipitation
enhancement occurs along the Sierra Madre Occidental but dry pockets are
observed between the west coast and the cordillera even as far south as 20oS.
Another dry area is observed east of the Sierra Madre Occidental due to a rain
shadow effect created by the mountains. This area is flat and low lying.
In terms of
topography, telescope sites need to be located on suitably elevated mountain
tops. Elevation is needed to ensure that the site lies above the temperature
inversion layer that exists at the interface between air that is sinking from
the tropopause level (~ 12km) and air that is near the surface. As explained in
section 2, locations above the height of the inversion layer tend to remain free
of moisture and/or dust from the surface. The height of the inversion layer
varies diurnally, seasonally and geographically but, in order that a site is
above the inversion layer as frequently as possible, sites should be as high as
possible. For the purposes of compiling the preliminary list of sites, a
minimum altitude of 2000m was specified in most instances. Exceptions were made
in the case of some existing sites and sites close to the ocean where the
inversion layer height is considerably lower. Where two or more sites are found
in close proximity, preference was given to the higher site.
The
considerations noted above have led to the preliminary list of sites shown in
Tables 2 - 5 and Figures 7 - 10. In selecting these sites, lack of road access
did not rule out a site. Other factors that may eventually rule out a site such
as geologic stability, mining activity or light pollution were not taken into
account in the compilation of the preliminary list. These factors were given
consideration before the final list of sites was decided upon (see section
6.4). On completion of the aerial analysis of cloud cover and water vapor
(section 6.3), preferred areas with the clearest skies and lowest moisture
conditions were identified. These areas were examined and other sites, not
necessarily on the preliminary list, were added to the pool from which the
final list of potential sites was compiled.
Figure 5. Mean
annual precipitation for the southwestern USA for the period 1961-1990. 
(National Atlas of the United States® ,
www-atlas.usgs.gov/)
Figure 6. Mean
annual precipitation for Mexico.
(Perry-Castañeda Library Map
Collection, www.lib.utexas.edu/maps)