Meteorological Site Evaluation and Forecasting needs for the Southern African Large Telescope SALT

1
Meteorological Site Evaluationand Forecasting
needs for the Southern African Large Telescope
(SALT)

• D. A. Erasmus
• Certified Consulting Meteorologist
• and
• C. A. van Staden
• South African Astronomical Observatory
• Correspondence erasmus_at_saao.ac.za

2
What is SALT?

• SALT will be the Southern Hemisphere twin of
  The Hobby-Eberly Telescope (HET) in Texas, USA
• SALT will be the largest single telescope in
  the Southern Hemisphere
• SALT is being built by an international
  consortium at Sutherland Observatory
• SALT will use a cost-effective and innovative
  mirror design with 91 hexagonal segments forming
  an array 11 meters across
• SALT will observe in the wavelength range 340
  nm to 2500 nm (ultraviolet to near infrared)

3
How is SALT progressing?
SALT Webcam picture on 4th Sept. 2001 at 119
p.m.
4
Atmospheric Conditions Relevant to SALT Design,
Site and Operations

• Transparency Cloud cover ? optical, water vapour
  ? IR
• (Observing quality photometric, spectroscopic,
  unusable)
• Surface winds
• (structural considerations, operating
  thresholds)
• Temperature
• (thermal controls design, operations)
• Turbulence
• (Image quality or seeing, adaptive optics,
  building height)

5
Atmospheric Transparency and SALT

• Design
• SALT observing wavelengths (UV to near IR) were
  predetermined
• Site
• Predetermined that SALT would be located at
  Sutherland Observatory
• Operations
• SALT will employ a queue scheduling modus
  operandi
• Forecasts of observing conditions based on
  cloud cover and water
• vapour forecasts will be needed to optimise
  scheduling

6
ESO PWV and Cloud Cover Forecasts

7
Sample image forecast product

8

9

Forecast accuracy at Paranal
10
Surface wind speed and SALT

• Design
• SALT will operate in wind speeds up to 16.8 m/s
• Natural ventilation will be used at night to
  keep the telescope in thermal equilibrium with
  the environment
• Site
• Frequency of occurrence of winds above this
  threshold at Sutherland Observatory is unknown
  (only one year of automated weather station data)
• Operations
• Wind speed forecasts will be important for SALT
  operations

11
Surface temperature and SALT

• Design
• Mirror alignment is critically dependent on
  temperature
• Thermal imbalances inside the dome degrade
  seeing
• SALT dome will be air conditioned during day to
  match the expected temperature at start of
  observations
• Site
• Temperature change during the night at
  Sutherland Observatory is relatively large for a
  telescope site
• Operations
• Temperature forecasts will be essential to SALT
  operations

12

• Forecast methodology uses ECMWF model output in
  combination with in situ observations
• 14 days of in situ data are used to train a
  Kalman filter
• Kalman filter corrections are applied to the
  ECMWF forecasts

13
Mean absolute 12-hour forecast errors
14
Atmospheric Turbulence and Seeing (Image
quality)
?T ? ?? ? ?N

• Turbulence creates small air pockets of different
  temperatures, hence densities
• Rapid small scale fluctuations in the refractive
  index of light occur and an aberrant light path
  through the atmosphere results

Turbulence
Turbulence
Image motion and blurring occurs
Telescope Mirror
15
Atmospheric Turbulence and Seeing (Image
quality)

• The apparent angular size of the object is a
  measure of the seeing quality
• Strong turbulence implies greater image motion
  and blurring, hence a large seeing angle

Turbulence
Turbulence
Image motion and blurring occurs
Telescope Mirror
16
Atmospheric Turbulence Effects

• Note the fast drift pattern from left to right
  and a slower drift pattern from top-left to
  lower-right.
• These are produced by turbulent layers at
  different altitudes being transported by winds
  from different directions

17
Atmospheric Turbulence and Seeing (Image
quality)
Theory The long-exposure image size (? ), FWHM,
depends on the integral along the light path
through the atmosphere of the refractive index
structure parameter (CN2), ? 5.35 ? -1/5
? CN2 (z) dz 3/5 (Radians) 1 where ? is the
optical wavelength(m) and z is height(m). CN2 is
a function of CT2 as follows CN2
(7.9x10-5P)/T22 CT2 (m-2/3) 2 where P is
pressure (mb) and T is ambient temperature (K)
and, CT2 lt (T(x) - T(x ?x)2gt / ?x
2/3 (oC2m-2/3) 3 where x is a position vector,
lt gt indicates a time average and the - 2/3
exponent is an artifact of Kolmogorov turbulence
theory.

18
Atmospheric Turbulence and SALT

• Design
• Does the SALT image quality error budget match
  what the atmosphere
• will allow at Sutherland Observatory and vice
  versa?
• Will adaptive optics improve SALT image quality
• What is the optimal construction height for
  SALT?
• Site
• Are there local variations that would make one
  site better than others?
• Operations
• Forecasts of the CN2 - height profile, of wind
  speed at the level of
• turbulent layers and total seeing will help
  to optimise telescope
• scheduling and use of adaptive optics systems

19
Atmospheric Turbulence and SALT Design
SALT image quality error budget 0.6 arcsecond for
50 enclosed energy (FWHM)
20
Atmospheric Turbulence and SALT Design

• Adaptive optics makes good seeing better it does
  not make bad seeing good
• The successful application of adaptive optics
  depends on which turbulent layers dominate the
  seeing

Boundary layer
Tradewind/Westerly boundary (3km)
Jet stream (15-20km)
SCIDAR profiles of turbulent layers at the
Sutherland Observatory
Height (km) ?
21
Atmospheric Turbulence and SALT Design
When seeing is good, most of the time, the free
atmosphere turbulence dominates
22
Atmospheric Turbulence and SALT Design
Turbulence near the ground and SALT construction
height
The shape of the thermal turbulence profile
indicates that little benefit is gained by
locating SALT more than 15m above the ground
23
Atmospheric Turbulence and SALT Site
Sutherland Observatory
Wind direction weighted DIMM seeing at SALT
candidate sites (Median in arcsecond)
R
24
Atmospheric Turbulence and SALT Operations
Forecasting the CN2 - height profile and total
seeing
Mauna Kea Observatory Hawaii

• This is a challenging undertaking
• Forecasts are based on simulation schemes that
  use height profiles of temperature and wind speed
  to model the CN2 height profile
• Valid simulations require high vertical
  resolution
• Synoptic scale forecast models (ECMWF, MRF) are
  inadequate
• MM5 run in high vertical resolution mode shows
  promise (Rick Knabb, see figure)

25
Thank You