As part of the conceptual and preliminary design processes of the Thirty Meter Telescope (TMT), the TMT site testing team has spent the last five years measuring the atmospheric properties of five candidate mountains in North and South America with an unprecedented array of instrumentation. The site testing period was preceded by several years of analyses selecting the five candidates, Cerros Tolar, Armazones and Tolonchar in northern Chile; San Pedro Martir in Baja California, Mexico and the 13 North (13N) site on Mauna Kea, Hawaii. Site testing was concluded by the selection of two remaining sites for further consideration, Armazones and Mauna Kea 13N. It showed that all five candidates are excellent sites for an extremely large astronomical observatory and that none of the sites stands out as the obvious and only logical choice based on its combined properties. This is the first article in a series discussing the TMT site testing project.Comment: Accepted for publication in PASP, April 2009 issu

ABSTRACT. In this article we present an analysis of the statistical and temporal properties of seeing and isoplanatic angle measurements obtained with combined Differential Image Motion Monitor (DIMM) and MultiAperture Scintillation Sensor (MASS) units at the Thirty Meter Telescope (TMT) candidate sites. For each of the five candidate sites we obtained multiyear, high-cadence, high-quality seeing measurements. These data allow for a broad and detailed analysis, giving us a good understanding of the characteristics of each of the sites. The overall seeing statistics for the five candidate sites are presented, broken into total seeing (measured by the DIMM), freeatmosphere seeing and isoplanatic angle (measured by the MASS), and ground-layer seeing (difference between the total and free-atmosphere seeing). We examine the statistical distributions of seeing measurements and investigate annual and nightly behavior. The properties of the seeing measurements are discussed in terms of the geography and meteorological conditions at each site. The temporal variability of the seeing measurements over timescales of minutes to hours is derived for each site. We find that each of the TMT candidate sites has its own strengths and weaknesses when compared against the other candidate sites. The results presented in this article form part of the full set of results that are used for the TMT site-selection process. This is the fifth article in a series discussing the TMT site-testing project.

The results on the vertical distribution of optical turbulence above the five mountains which were investigated by the site testing for the Thirty Meter Telescope (TMT) are reported. On San Pedro Martir in Mexico, the 13 North site on Mauna Kea and three mountains in northern Chile Cerro Tolar, Cerro Armazones and Cerro Tolonchar, MASS-DIMM turbulence profilers have been operated over at least two years. Acoustic turbulence profilers - SODARs - were also operated at these sites. The obtained turbulence profiles indicate that at all sites the lowest 200m are the main source of the total seeing observed, with the Chilean sites showing a weaker ground layer than the other two sites. The two northern hemisphere sites have weaker turbulence at altitudes above 500m, with 13N showing the weakest 16km turbulence, being responsible for the large isoplanatic angle at this site. The influence of the jetstream and wind speeds close to the ground on the clear sky turbulence strength throughout the atmosphere are discussed, as well as seasonal and nocturnal variations. This is the sixth article in a series discussing the TMT site testing project.Comment: 36 pages, 14 figures, accepted for publication in PAS

A simple and inexpensive six‐channel array of photodiodes, LuSci, can measure weak moonlight scintillation produced by optical turbulence within few hundred metres above the ground. We describe the instrument, its operation and data reduction. Measured covariances of lunar scintillation are fitted to a smooth turbulence profile model with few parameters. Complete recipe for calculating the instrument response (including the effects of Moon's phases) is given. The robustness of the results relative to various experimental factors and model assumptions is investigated. We give examples of the data and compare LuSci with other turbulence profilers. LuSci finds numerous applications in nighttime site testing and monitoring.

Differential image motion monitors (DIMMs) have become the industry standard for astronomical site characterization. The calibration of DIMMs is generally considered to be routine, but we show that particular care must be paid to this issue if high-accuracy measurements are to be achieved. In a side by side comparison of several DIMMs, we demonstrate that with proper care we can achieve an agreement between the seeing measurements of two DIMMS operating under the same conditions to better than Ϯ0.02 arc sec.